51吃瓜

welcome to the next wave podcast today we're doing a special episode here live in san francisco for hubspot inbound and we have a special guest host carly baker from hubspot she's gonna be having a great conversation with var ana from clay and if you haven't heard of clay they've been really blowing up recently they recently raised around a three point one billion dollar valuation from google's growth fund and a lot of other great investors including sequoia and the simplest way to describe them as ai for go to market i think you're gonna love this conversation so let's just jump right in hey guys how's it going good morning good morning good morning thank you everyone for being here with us today super excited to be able to kick off inbound right before yam keynote yeah to sit down and and chat all things clay yeah at the talk of the town excited to hang this is talk of this town you are oh wow clay play okay great lot thanks for timing we did just launch these billboards are really excited about timed for this particular talk like this whole lot of the open campaign so i was very excited about that but i think we can just kinda keep it conversational that could be fun yeah i i love to kick off exactly what you had shared which was when i was in my uber you were from the airport yeah i saw your billboards everywhere your billboards and our billboards okay great yeah guess we have competing out of home campaigns going on i think out of homes a big like under underrated thing like i don't think people do this remotely as much as they should be doing it's really cheap actually like if you raise venture capital and have money it's actually surprisingly cheap for the mind share that you get from it and i also think most of them are quite boring yeah like most of these ads are really sad i know i mean i i also went down the highway last night airport and i was like very disappointed by the lack of creativity yeah on most of these so tell me how you think about creativity specifically out of home and with billboards right i know maybe we've all seen sort of some of the behind the scenes that you posted i know you had drones out there capturing yeah the the setups and all those things which i love sort of the behind the scenes of of out of home and i could be wrong i'm gonna try to quote exactly what the billboard say but i think it was like picasso had art and go to market has clay or something you have a that i don't like big playful in this at that yeah well basically it's every artist has has their medium and and go to market as clay and then there's like these children ones which are like michael had marble and go to market as clay or monet paint go market clay by the way this wasn't the question we even talked about so we're already off script and we're already big doing good things so just so you know that this is r here which is good well first of all i can't actually take any of credit because it's actually this woman jessica got our team who is maybe the most creative human being i know and bruno leads our marketing and like sarah tanner and hudson on our brand team we by the way as as like a separate aside i think we over invest in brand like i don't know we have a five person brand team or something and we hired like our head of brand when we were like twenty people or eighteen people or something so we probably over invest in it you know it's one of those investments that makes you look good if the company works but if it doesn't work that was a really bad idea because we you just wasted all this money on something that doesn't really matter but i will say the reason i'm starting with this brand thing is because you can't just like come up with something creatively do it like there's a lot of steps that go from you know four years ago to now of like okay actually having a very creative out of home campaign because first of all it's very core to who we are it's like corridor identity yeah like the logo for example inspired by this extraordinary artist of solar and creativity and art like runs through the company's dna in a lot of different ways and then we obviously back it up with like the money we invest to do it like we have a full time clay nation artist on staff you know like a full time and i think he takes a lot of vacation too it's like he and we've actually been trying a recruit of like four years notion got has someone like this too like not cremation but like the person who does all their doodle you know yeah and so you know a lot of companies are not going to make that kind of investment and that's okay like brand doesn't have to be everything for every company it is for us because it's authentic to who we are like for example jessica we can just talk about her for a second right so just goes a longtime time friend of mine we actually started a pickle ball business together like five years ago and she like gained a lot of notoriety like a decade ago when she was in like ut austin and basically there was a law texas that said you couldn't have like guns public but there was also another law like concealed carry laws and there was another law that said that you couldn't have like sex toys public and so she like started this like campaign and protest where people would like march would d those like down the street you know to protest these like gun lost mh and so that's kinda of how she became like famous viral but she does these kinds of stunt all the time mh and so she's kinda the most creative person i know and then like three years ago i think when the team was just like ten fifteen people i was like okay well can we find a place for you a play and even for me i was like okay this is too extreme like this feels irresponsible even for me we only have one marketer and then have another marketer who knows what they're gonna do you know like yeah that felt too extreme but then three years later when the marketing team is fifteen people okay okay now we can have someone who knows with and so her title is like scheme residents but i think like for example like her north star of just coming up a creative stunt by the way here's a plug she really asked me to plug this so it's not even a plug because it feels very salient into the conversation this week we're doing all these events that are very creative so like we hired snoop dogg like a joint ruler person and he's doing a class for people on how you can roll your own joint like snoop dogg does so like that's a thing then that's on friday i think somehow people are not showing up this like i don't fully understand why but people are not signing up but it seems amazing okay i i'm actually flying out tonight if i wasn't flying outside i would go but if you are remotely our ic you should go that sounds very fun number two there's like a kung fu fighting scene that's being filmed okay so if you've ever wanted to be in like a kung fu fighting scene situation this is your moment right and so i'm not kidding like it's series there's it's a professional who's gonna come in and he's gonna coach you about how you can like be in this like battle scene and they film this whole battle scene and and it's like pretty sick so you should do that that's this week as well and there's like a perfume thing where you like go deep into the art of smelling and stuff out know those all sorts of wacky things but anyways like who comes up with that you know jessica does yeah and so this what i mean by like the creativity and like we invest in that because again it feels authentic to us yeah and i definitely wanna dig into that a little bit more too because i think when we think about brand often we think so much about sort of what the online experiences i think specifically in the text spaces you really should sign up for those things but we really are world these sign up yeah you know you think so much about the online experience and to be able to sort of bring those core brand values and be able to do something really against a grain ir i think it's really important yeah i mean even when we're here and even when we're at conferences right it's like how many of you have four happy hours to go to tonight and what are you gonna remember of those four from those four happy hours so like leaning into sort of connecting your ir connecting your out of home connecting like sort of the core values of your brand and being able to do the experiential stuff is super important i do think the more and more and more saturated that the internet gets them more important in those bespoke events are gonna be and to be able to have someone it's very hard to find someone who can be the steward of having those ideas and making them actually feasible and sort of tying them back to your ic right i think it's very easy to be an ideas person and not i have an idea sort of how to take those and connect the dots and be able to sort of make sense in the whole ecosystem and so it sounds like jessica's as your person yeah well i'm not totally sure how much like you know joint ruling really connects to ic but i think it's cool and i think it's fun and i think it's interesting which is more that can be said for most programming at go to market events not of course inbound obviously and i think look i think people are people and so people are people and like you just resonate with them in different ways and like like dream force last year we had like this big spa and everyone was like naked in the spot together and i mean they were technically naked it but you know but but it was when people were connecting because they were at dream force was exhausting or you lose your voice you're doing all this stuff and then you know what you want you wanna massage you know what you wanna hang out and so i think it's just finding ways for people to connect that feel like human a nice and fun and refreshing and like do things people actually wanna do and yeah maybe talked about clay in the middle maybe hang of it i know one thing that you talk about very openly you know as you've shared about jessica here is that you're really a champion for your people and a lot of folks that come to play yourself included sort of come from non traditional backgrounds and sort of you and clay really take bets on sort of more non traditional hires and really champion that and you know use that as a part of your strategy i would just love to hear more sort of about the intentional about that and sort of how maybe that's shifted and and sort of what other people can learn about maybe like taking some of those bets in terms of hiring yeah i actually don't think it's that risky or that much of a bet i just think it's the philosophy how we think about it the art of interviewing and recruiting is just understanding who a human is and i think people too often try to make value judgments about who people are people like when you're interviewing someone that oftentimes feels like hey close this person's smarts or this person's dumb or this person's good at their job and they're competent or they're not and it's like easy to do that by way i used to do that all the time just to to call myself out on that but i think the real point is just understand who they are as a human and they're kind of like a blank canvas when they come in front of you right and your job is to fill out the blank canvas you use pastels styles and watercolors and the references in the interviews they all kind pieced together into filling out the story of who this human is and then once you understand who they are then you can kind of try to understand okay what is the fit right and so like let's just say you thought someone wasn't competent or wasn't smart or whatever something like that it's not actually those things it's actually just that hey they need more mentorship and guidance and help and support than you were able to provide in your current structure right maybe they should work in google because they have a lot more of that than like some fast growing startup up play clay because that's not gonna be able to give them that structure and support right and where someone's gonna be successful so i think it's just thinking about people in this like terms of fit and like where are they in their journey of life and what are they looking for and what do they need to be really successful like what does it take for you karl to be unleashed you know like like what does that take and i think there's actually so easy right over there is video taping this for some reason video taping it by the right next to a very professional video taping setup up i think it's because we want like a fast follow on content so but i why i am grateful but anyways i is a perfect example of this by the way and billy right next to her and we're working on a role so we're still working on that but but an example nonetheless so i she was a product manager at sam sarah we have this like rotational program called the wheel where basically everyone's not hiring new grads and we're going all in on new grads and so with this retention program where we take new grads or people just out of college and then they do product support they learn the product really well they understand their customers really well and they rotate other customer facing functions and then within like a year or something they go join that customer function full time that's usually faster because everyone chases them because they know the product and customers better than anyone mh anyways easy he applied for this and i was like the final interview and it was like a fifteen minute conversation and i was like why do you wanna do this again because like you were a product manager at some sara and now you're trying to do product support and it didn't totally make sense to me and and she explained by the way but i had some other context about i because she sent in this video and then she has is a hamster there's a hamster it's a bunny okay sorry it's hard for computer i think it's i thought it was hamster okay sorry it's a bunny so she has his bunny and i remember this video was bunny and it was like very thorough video about like a clay use case in his bunny and then there was also some other details of how like proactive she was that cr mentioned that she came up to him at some conference it was having you indian food cream it was like hey did you meet this person named and i was nobody damn i'm talking to her tomorrow there was just like some details about this that i had and then we meet for fifteen minutes and then like immediately was clear that like this didn't feel like the right fit you know i was like okay what could be the right fit and anyways i asked her if education could be a good fit and i don't know is it a good fit yeah okay so it's a good fit similar by the way not just as the other person on education team who started in the same day was this a very funny similar story which is she was applying for like a cx role she did growth before that all these things and we have this interview and by the way you there's a lot more serendipity in person interviews but it's harder to do that usually on on virtual interviews than it is in person so anyways sarah this other person were like in this conversation and she keeps referring she has this like notebook right and she keeps referring to it and i'm like i see that notebook and so i just like look at this notebook and i like read through all the notes by the way it was very prepared and so she had lots of notes and then i asked if i can read her apple notes too so i read her apple notes i did by the way before checking into different notes i if they were ambiguous titled i did ask her for permission to read that so i read that too like during the interview and basically i got a lot of insight to who this person is mh and then i asked her hey do you wanna consider education and instead of cx and i didn't feel like this was the right fit and we were kinda shoe warning it because i think she was excited about clay and and so i like something instinctively that doesn't feel right and then you know what she said she was like actually i can't believe you asked that because my whole life i wanted to be a teacher and i've never been able to do it because of the pay and the economics of it but i'm like really excited about it again that was like completely i don't know nowhere but i think it comes from like really being so naturally curious about whose the person is okay i have another tangent sorry i just met this guy brandon who is the founder of humans of new york mh and i don't if you guys familiar with humans of okay sounds like people are and yeah i've been a fan of this for a long time so was pretty cool for me to like meet the sky and i talk to him and i was like how do you get these people to so basically open up their whole life stories to you in like seconds right and then like get deep and honestly the answer he gave which is so similar to what happened with i and sarah is you just have to be naturally curious and i find in interviews a lot of times that even myself by the way you ask for like what is the best question you can ask like how you do this honestly none of that matters if you're actually naturally just really curious about understanding who whose person is the questions kinda come out naturally and the same thing happened with brandon in the humans of new york and people feel that natural curiosity and they lean in yeah because like you're a stranger you sit down you're hey can't take photos of you and posted on the internet right like that seems off but if you're naturally like eager about understanding who they are people respond to that and then the same applies for interviews and so i think that principle is kind of how we're able to like understand who they are and like find the fits for them in the organization alright you've heard the ways var is using ai to do some next level things well he just dropped the database of all databases and we're not talking about those typical use cases make your emails sound better our database has over one hundred unconventional ai use cases across thirty industries each use case includes how you can implement it difficulty level required tools and metrics to measure for business impact this database was made to give you that unfair advantage get it right now click the link in the description now let's get back to the show yeah so i wanna take sort of the philosophy and and thought that you have around hiring and talk about sort of how that's related to just the pivot and sort of the changes of clay as a product over time i think there's actually a lot of parallels and your sort of personal and the team's personal curiosity really played into that so can you talk about sort of just the journey of going you know just sort of from fancy spreadsheets to going all in to being a full sort of go to market ai powered tool and sort of how that story maps with how you think about hiring yeah basically in like late twenty twenty one early twenty twenty two we called ourselves like a spreadsheet that would fill itself and it did but no particular audience that well and i think that really reflected a lack of commitment and a lack of focus from all of us and so basically there was a moment where we were like okay let's commit and where should we commit to well i think there's pull here in this go to market space and then we kinda need to understand where to start and so for me i was like okay well this is a abroad audience how do we like narrow it down to something that we can start with mh i talked to a lot of different people and the only audience that i was feeling any type of pull from where are like agencies mh like these cold email agencies marketing agencies that's where i felt immediate pull and so we just kinda focused on them now in retrospect it might seem a little silly like if we went to go raise venture that and we were like we're building software cold email agencies i don't think that like pitch would have landed very well i don't think would have won the hubspot pitch contest about like like ar dead plug for ar because it's not like a big market or whatever right but i think these things are very unpredictable there's also a good book on this by the way the greatness cannot be planned by kent stanley in which i really like is i actually almost successfully hired to clay again for an non traditional role kinda of study curiosity and stuff the reason i resonate with the the premise of that book which you can kinda guess for the title is that greatness can't be planned and greatness could be defined as product market fit it could be defined as an iconic company whatever you you kinda have to follow your instinct in your curiosity and each step and each stone unlocks the next one right and all you're trying to do is just get to a place like a base cap place right and so that's what happened here where we followed our curiosity we started with these cold email agency agencies we didn't know where that was gonna lead but we knew we were helping them and we knew we were helping them because they felt the pain so acutely because they had lots of clients and then we were kinda able to par that into them posting about us on internet we're were able to probably par that into working with their clients take that and part and meet with of work with other startups ups and smb and eventually look our first enterprise customers rip and then anton open ai and so we were able to take that and like get momentum with like these enterprise customers and and like the way it works is the bigger companies look at the company right below them right and so the onion kind of unfolds in that way yeah and then that's kind of been the journey of how we've grown and then from the product perspective we've like gone from really just a spreadsheet to a data richmond tool and to now like a go platform that we call like an id maybe that lets you kinda come up with any idea really quickly execute those ideas and kinda execute any growth play and we're still on that evolution and still because i still don't think we're achieving that vision remotely as much as we want to i wanna talk a little bit about sort of the intersection of curiosity and the fact that we're sort of increasingly more and more in an ai world where think curiosity is being sort of put on the sidelines i'd love to hear from you personally as someone who really champions curiosity sort of any exercises or anything that you do or encourage others to do to sort of stay in touch with that side of yourself or just to build that up sort of as you know being curious or finding information or getting ideas sort of easier than it's ever been in the age of ai i guess a simple way to put it is can you help people become more curious can you help people become more curious you're right that it's it's more difficult because people basically just default chat for everything right now mh and i think there's a lot of like trade offs with that that are not so obvious because yeah you have these ai tools and people are like okay immediately you just go use them and then like you don't think about what happens because of that so like for example a lot of people use chat as like a therapist right and like what are the trade offs of that are people then not going to talk to their friends and other problems what does that do to friendships you know what does that do to our social circles so i think that's like this challenge and then also by the way that's exacerbated like in in san francisco where we are now there's this whole push to like nine nine six and like those types of companies which play by the way definitively not and anything i think those companies also kind of that philosophy kind of exacerbate as problem actually because people are like they like they like they kind of like success is like a treadmill right where a lot of things come from curiosity from openness from creativity for having time and space to think of about problems actually just ran the sort of a last night about this guy jun so who is an economist position at princeton and he just won the fields medal mh which is like the best prize in mathematics and he was a high school dropout and he was a high school dropout to study poetry and then didn't actually get into math until like a sixth year at like seoul university or something and you look at his days like you wouldn't as describe any of those things to actually like doing good thing like getting her work done like his wife almost like didn't get married to him or divorced him because like he couldn't like do the basics of like household items and stuff right but not creativity in that magic and that space that he creates is what enabled him to like come up with these breakthroughs that warranted into them a field battle right and i'm not saying you need that type of thing for everything you do yeah right but i think you just need that space and that time and it's hard to do that when you're in this like treadmill you know this grind i feel like i've still been evading the core question of like how do people actually do this practically mh and so i don't have a great answer of how people can be more curious i think they just have to like want it and they have to create space for it and they apply it in their day to day but that's kind of the best answer to can come up with right now yeah well that's a great answer i wanna hear from you about sort of creative use cases or things that you're seeing customers doing that you just never sort of had the idea of sort of your customers using play in that way just sort of to follow the thread of curiosity and creativity yeah well not to call him i he again but i think i he had a very creative use case we have a single clean boot camp for all new hires of company and they basically have to use clay in some creative use case in their lives easy do you wanna tell them about your like creative use case yeah in my table i was building out an offer negotiator using clay so it'll pull similar job data as a similar companies for roles that you wanna apply for or have a applied for and you can also input using like lovable as almost like a front end and clay has a back end where we're processing the data that you put in if you put in details about your offer letter it'll go ahead and compare that to like the industry standard that clay has pulled in from like the recent linkedin job postings and then give you back on lovable a negotiation letter that you can now send back to a recruiter over email so var actually wanted to use this for all our candidates yeah did we do that by the way i don't think we actually did it but we're posting about it on our g tammy blog okay let's post about it but also let's post about it in the recruiting channel or send it to depot decode me and like i think we should actually include it in the materials because people should have access to the latest things that could get them more money and if it actually helps them get more money then either they're right or our process is wrong and so either outcome is fine so we actually should do that okay that was pretty creative and then if i remember correctly and there's a lot of like my ea just started last week and she had a creative use case in her boot camp where she like found all the gm of all the fancy hotels and all these places and got their contact information and sent them all like a very personalized letter that makes it so that they're more likely to give me an upgrade when i go she calls it nice hacking i have encouraged her to make did post about that iv from our team like created a builder in play for like finding tennis courts and hiring tennis coaches and stuff which is pretty creative like there's like a use case of like tracking all the new content on hbo of max and amazon prime and apple tv and then like having like okay what what what scores are they gonna run to than i d and like what should you watch this week right so these are all kinda like very fun creative use kids in your personal life i can give like a real work one too but okay so there's a company a logistic staffing company for example and they serve warehouses of a very particular size because that's how a chat account they have but the traditional data sources for warehouses kind of suck actually and are very good so what to do well they're were like okay well maybe ai can help and it does actually so they actually use clay to automate finding all the warehouses maps and then they use ai to analyze the satellite images and then in the satellite images they count how many parking spots are around the warehouse and then they can't how many cars are in the parking spots and it turns out that that's the best predictor for head count size warehouse size and anything else very interesting that's creative yeah another one i'll hit you with canvas you guys don't know canada big company they one of their products helps you keep your brand consistent across lots different social channels so they use play to automate social listening on linkedin and twitter then they like use ai to analyze those posts and be like are they on brand or off brand and if they're off brand it's like hey like the font is off the coloring is off spacing off then they're like okay well let's find the head of design let's message them and be like y you just post on linkedin yesterday it's off brand here's why doesn't it need your brand guidelines that's not good for you here's how we can help right so it's like timely comes with a solution it's very effective right those are some like in the wild created use cases yeah i love those all very different as tools sort of become more natively integrated in workers lives we're gonna see so many more of those sort of personal use cases where they're building using a tool at work and they're expanding that into how they're using it in their real life and just like tools and software being more natively i think as a part of just people's lives outside of work which is yeah something we've never haven't really seen before at least at this type of scale yeah think very interesting i think we have time for one more question i do wanna go back more towards the product side and just hear sort of what you're building for twenty twenty six and beyond as a guys sort of continuing to to grow and shape technology as a whole like what do you really sing on and and where you're looking to grow with the product yeah i think there's a lot and i think basically our ambition is to build a space where it's very easy to come up with ideas really quickly execute them and then the ambition of those ideas can be really really high and i think we are meeting certain elements of those problems of that ambition right now to certain audiences mh but definitely not to to everyone and to all these use cases and like the bar to access product to super high right now how do we make that more accessible so we're working on like lovable s like interfaces and clay to make that really much more accessible how do we like help you be way more creative of those and be a priority how do we help you like be way more ambitious and so like we're considering like agent based node based architectures to help you do way more all the way the tables in some ways and be more ambitious so there's a lot kinda coming up down the bike that we're really excited about and then also like how do we expand into new audiences right like how do we not only just help ops teams but how do we help like marketers do ab and sophisticated ways how do we help reps be more efficient at their day to day right so i think there's both surface area and scale ambition that we're excited about and kinda of a lot coming down the pipe yeah are you more excited about sort of surface or scale like how are you sort of bouncing your priorities it's tough tough balance i was just talking to our friend cheek bond like i think he was asking me about our ambition and it's like it was like use referencing this other company and their ambition is like to become a hundred billion dollar company and it's like i hear about that like not trying like that's not the point the point is that a lot of people in this room are starting companies running companies and when you have an idea or when you start doing your new company we want your first instinct to be to use clay the same way your first instinct right now is to open up your phone and call it uber or way or or search something on like that should be the default mindset and then not only is that the instinct but we meet that promise right and we actually help you do something that genuinely grows your business in a way that's differentiated and that's ambition we're we're trying to meet that but that's kinda where we're at it awesome well thank you so much everyone for joining us this breakfast and this conversation was presented in partnership with clay of course with our friends at hubspot for startups which is your go to place where tools and resources and exclusive access to the starter community at hubspot with mind which is hubspot media from your ai newsletter and then of course the next wave which is our ai podcast and newsletter we're really excited to share this conversation we'll be available in just a couple weeks both on youtube and on audio for anyone who wasn't able to join us or if you wanna revisit the conversation so we'll send a follow with that but thank you so much for var for the time this morning thanks everyone for joining us and we'll see you over at nine thirty for yam keynote out next door yeah thanks guys

welcome to the next wave podcast i'm matt wolf and today i'm joined by kevin scott the cto of microsoft kevin has been at the center of of the most important technology shifts of the past decade leading microsoft's work on ai cloud computing and the partnerships that brought tools like open models to millions of people around the world in this conversation we get into the big questions everyone's asking about ai are we already at a point where these systems are too complex for humans to fully understand how do we make sure that as they get smarter they stay aligned with our goals we'll talk about the massive challenge of energy efficiency in ai what microsoft is doing to make these systems more sustainable and accessible and even when we might see autonomous ai agents that can go off and complete entire projects for us while we sleep kevin also shares where he thinks ai is headed in the next year or two what kind of jobs and opportunities that might create and why he believes the tools we're billing today could unlock an explosion of creativity and entrepreneurship it's a fascinating wide ranging discussion with one of the most influential people in tech right now so stick around because you're not gonna wanna miss this one my first question do you think ai will get to a point where ai is so smart that humans truly don't understand what's going on under the hood well look i think in certain ways we are although i don't think that that's any different from some other like really complicated systems that we built or like honestly a bunch of complicated phenomena that we don't quite understand so in my mind technology g is always about like is there a path to being able to debug it when it doesn't behave the way that you want it to behave and i think you know there are an increasingly good set of tools that we're developing to try to you know be able to characterize the performance of really complicated ai systems and to debug them when they you know aren't doing what you intended them to do right you know that we've been talking over for years and years and years about full stack developers or developers who can understand systems from top to bottom and i think you still need that a lot you will be more successful as a developer in the age of agent ai like in this era where you're using ai to do significant parts of your software development job if you understand that full system so that when it mis behavior you're like okay like i'm gonna punch down a level of abstraction and like go you know try to investigate what's going on here and then like you know if that doesn't do the trick like another one and another one and like eventually getting all the way down to a bare metal right right my follow question that would be like if ai gets to a point where it's like you know super intelligent right we get to that super intelligence phase where maybe it's writing code that you know humans have even figured out how to write this kind of code yet how do we make sure that it sort of continues to stay aligned with our our goals yeah i mean there there is a tremendous amount of active research and development and engineering on alignment both at microsoft and like a whole bunch of the companies that we partner with and you know a whole bunch of the companies that we're big fans of and like honestly it like much companies that we compete with like it's one of the things that i think all of us building this technology really want to make sure functions well the same way that all of us have a right you know universally high level of focus on security and classical software yeah no one wishes for anybody's software to be less secure or ai systems to be you know less safe or less responsible or less aligned or than what we need them to be so yeah i do think that there's just a ton of activity there and i've even seen it over the past you three plus years as we've made really powerful ai available in apis as we've plumb it through to products like i run microsoft's deployment safety board which is this thing that really make sure that we rigorously test and investigate everything that we're doing to make sure that the things that we're launching it here to microsoft publish responsible ai standards and like the sophistication of the tools that we use to do the job of the deployment safety board has just increased exponentially over the past couple of years right right yep yeah no i love hearing that a lot of the sort of big tech companies are all sort of aligned on the same goal and they're all kind of working together even though they might be competitors i wanted to ask about the whole energy efficiency kind of thing i know there's some concern that with you know data centers and ai and all of that using up so much energy what sort of things is microsoft doing to sort of a bring the energy usage down but also be kind of democrat ai so it can be more accessible to more people yeah well look i think the core thing that all of us are doing and like it's been a particular focus at microsoft over the past handful of years is ai demand has surged is like you've got a bunch of goals that are aligned so we just spend an extraordinary amount of energy trying to optimize the energy consuming parts of the system because energy cost a lot and like we've made very specific hard sustainable energy commitments to the entire world that we intend to meet and so inside of those constraints like you have to make sure that is the ai demand is growing things are getting cheaper like that's the way to make them more accessible so that like you're just lowering the barrier of entry to people doing increasingly complicated things with the system and like you have to make them efficient so that you can run more ai computations on a fixed power budget or a fixed hardware budget or a fixed you know floor plan budget that you have inside of a data center you know and then i think outside of that we're doing a couple of other things that are interesting so one is like we've got this constant background threat running trying to figure out whether or not they are these sort of disrupt efficiency breakthroughs so like the transformer for instance which is the foundational technology that all of these modern large language model ais are built on now one way to think about it is like it was a disruptive increase in the capability of the systems but it was also a disruptive increase in the efficiency of systems and so we're doing a bunch of basic computer science research on trying to find you know what those next disrupt active things might be gross research so you can't predict when the disruption might happen but like we feel good about the amount of energy we're investing in that and again it's one of those things where incentives are all aligned like we just desperately for all sorts of reasons like need those disruptions to happen whether you know they're our research or you know some research right it emerges you know elsewhere that's publicly available and then the other thing too that we're doing is like we're trying to help the electric power industry find new sustainable scalable sources of production so that you don't have to live in a world where you've act as if there's energy scarcity because they're like ai aside like you actually do want a world where there's energy abundance so and and sustainable energy abundance right right lots of good thing that they just think about like some of these problems that we have in the world right now so you know there are major social unrest happening in parts of the world right now because we have water scarcity right and there all sorts of ways that you could solve water scarcity problems if for instance des desalination we're cheaper to do right right but it is an energy intensive process right now and so it's too expensive and like too unsustainable to do if all of a sudden you had a energy breakthrough where you know energy became an order of magnitude cheaper and you know orders of magnitude more abundant then like you could go solve problems like that right right yeah i got an opportunity to go see the applied science lab yesterday and talk to some of them and one of the things they were saying is that the new n mp units are actually quite a bit less expensive to develop than gpu units yep making it so that you know maybe you don't have the best graphic cards to play cyberpunk twenty seventy seven yep but you'll be able to have ai right yeah ai agents are transforming marketing they're changing it as we know it and the old ways of marketing they're gone the new ways are really agent first ways and kieran just wrote this amazing blog posts about how to think about marketing and do marketing in an agent first world in a world where agents might be buying from agents or agents are facilitating ways people buy and there are these three very specific changes that he outlined and then if you're in marketing today and you are not clear what these changes are you're missing the boat you're gonna get left behind and we don't want that we wanna help you stay ahead of the pack and so you wanna read kirin post right now you can click the link of the description below that is gonna give you the blueprint you need to do marketing in an ai first world it's an important thing to understand that every year over the past five or six years as we've been building these transformer based ai systems you know every couple of years or so you get a new hardware generation that's given you about a two x price performance improvements right which is enormous like it's just really really extraordinary but on top of that like you're getting a set of even greater efficiency improvements that are happening in every single year in the software layer that's improving you know energy efficiency and capital efficiency in these systems and it looks you know plus or minus like ten x a year right and so yeah i i was you know at this event last night where i said you know like when i the young developer like because i'm a old fart now like all i had was this crappy old moore's law which would you know again was like this extraordinary exponential but like it it's not the same exponential that we're on right now and so like you you just gotta remember like the hardware is getting better and the software is getting better at an incredible clip which means like all the ai is gonna get cheaper faster like more capable and more energy efficient at the same time right i wanna shift over to agents because it seems like agents was a big piece of yep you know what you've been talking about how far off do you think we are to being able to sort of tell an agent before i go to bed redesign my website coded up and had it live for me and then i wake up the next morning and i've got everything completed how far off is that yeah i don't think that that is actually as far off as it might it may kinda sounds like a science fictional scenario especially like if you rewind a last year's build but like i i think it's closer than you might imagine like i've got a dad at my kid's school like every school event to like where friends and he comes up to me and he's like oh my god you can't believe the crap that i'm doing and like he runs his own business and like he's you know one man shop you know with contractors that he's hired over the years to go help him build apps and like he still has you know these contractors but like now he has this other thing called the you know software development agents that are helping him do all of this stuff and he's using everything like just a super sophisticated user uses our stuff he uses open ai stuff he uses you know and philanthropic stuff he uses a google stuff and the rate at which he is able to do his work is really incredible right i mean like really you know so you just described a thing that's gonna sort of asynchronously and autonomously go off and like do a bunch of work while i'm sleeping great he's almost there right now i have my kids end of year review of some social entrepreneurship work that they were doing and like my daughter was in this group with two other girls like building this app that was trying to get kids to take calcium supplements right and completely unbeknownst to me like she was off using these tools to go build an app and like the app like if i remember the first mobile apps that i wrote in two thousand and eight it's probably better than the first one that i wrote by by a mile and like my daughter has never taken a programming course you know she's had a little bit of programming exposure and like a multi design class that she took she's in tenth grade sixteen years old and yeah she did all of this it also didn't never occurred her to like tell me what she was doing our asked for dad's help which is rich so yeah i'd look i i think we're probably closer than most people think to that scenario you just described right right one of the things that you mentioned both last night and again today on stage was memory sort of being a habit bit of a bottleneck why is it a bottleneck and what needs to be accomplished to sort of get past the memory issues first let me describe why it's a problem so you know like you and i are interacting right now and you know even though both you and i are busy like we will have a recollection of this interaction that we've had so that if we ever are interacting again like we can recall you know this conversation and what we were talking about and like we've just sort of got a foundation you know between the two of us for you know future interactions if you're thinking about agents is things they you can delegate things too and that you can like think of as collaborators that memory is just gonna be a really foundational important part of the user experience of these things and it's also pretty important from again an efficiency point of view so right now a lot of the times when you're using an agent because memory isn't as good as it needs to be like it just hasn't remembered anything about your previous transaction and it doesn't even remember much about what it has done itself in taking a sequence of actions to go solve a problem which is kinda crazy and so you spend a bunch of time like rebuilding state inside of these agents that you shouldn't have to if memory was functioning well and so part of that is just you know it's kind of a constrained thing so the way that these systems work is you have context windows so like a prompt you can only put so many tokens or words of instruction or information or context like into this window that you then feed to the inference system to get a response back and the response also consumes space inside of the context window and you know if you're iterating over the course of a session like that consume space and a context window and these context windows are sort of bound because like there's some ways that you can implement context and inference where you the inference is quadratic in the length of the context windows which means like it gets not just expensive you know with one little increment per additional token like it it gets a lot more expensive per additional token of processing that you're doing so one of the things that we've had to invent over the past handful of years are more efficient ways to use that context more efficient right attention algorithms so that you like don't necessarily for each part of inference calculation you have to look at the entire context of information that's there and like building things that are you know honestly a little bit more efficient than you know retrieval augmented generation which was sort of the way that we had before to take you know contextual or semantic relevant things to the task at hand and like pulling them into those context windows for processing right so it's just been like one of those things it was too expensive and like we needed to go do some efficiency work to make it better and like also some you know like really if you think about how your memory and my memory work we have pretty good precision and recall but it's not pretty good one shot precision of recall so if you asked me to go remember something that happened five years ago i am not going to be able to just like that tell you exactly what that recollection is but like i have a way to go get to it it's like okay like i kinda remember that like i can put it in context that context helps me like go out to my sources of information this kinda of search around and like i have a way to like have really broad recall and a way to take imp precise and make them precise and so like that's also what we need to do inside of these memory systems so like memory in a agent isn't a database lookup right it's like a iterative process that the agent you know may need to do to get the right piece of information very precisely right right i'm curious if we look ahead let's say like a year or two from now yeah and what sort of things do you think an agent will be able to do that like we all do on like a daily basis right now is there anything you think agents are just gonna take off our plate and people don't even realize agents are gonna take that off our plate yeah like i think there are things that i hope for that probably are hard not for technical reasons but for other reasons and then there are you know certainly a whole bunch of knowledge worker toil that i hope you gets resolved and then then there are things where i think they're just sort of hard mh so like the hard things are like the embodied ai things so like there's just a bunch stuff that all of us have to do in the physical world like you know do your laundry and you like i can't get my kids to like put the dishes you know take them from the sink and put them in the dishwasher like it just really irritates my wish out a robot to do that unfortunately i think those are like really hard technological problems where we're probably not on as fast path to getting those embodied ai problems solved as we are with some of these cognitive problems you know i think a whole bunch of stuff like the scenario you described is somewhat likely to happen in the next year like i have a whole bunch of things that i wish an agent could go do for me while i slept where i could wake up in the morning and you know it's already started writing responses to you know emails that i need to like get right to first thing in the morning where you know maybe someone else's agent has like talk to my agent you know try to get quick responses to things that you know the other person needs where you know they no longer have to wait for me because like i'm blocking them so like i i think a bunch of this asynchronous stuff is actually gonna start happening pretty quickly and like the the reason it hasn't happened so far is because you just need agents when they're taking action like they have to be really precise and like so far we've had kinda imp emphasized memory and you know action taking actually gotten dramatically better over the past year with these reasoning models but like we need you know all of this mc plumbing that i'm really talking about to happen so that you know agents can do this communication to other agents and systems that needs to be done and then they've like stuff that i you know hope for and they're just choices that we make as society whether or not we get them they're not technical barriers so i wish that medical diagnostics were more available to people everywhere over the next year that like folks like my mom who lives in rural central virginia and she may not have access to the same diagnostic medicine that i have access to living in silicon valley right like i wish these ai systems which are already like right now plenty good enough to help give a real boost to people living in rural central virginia where my mom lives like i i wish like we would get to more adoption of that but again like that's not a technology problem that's a set of choices that we're making right right so i have two last questions for you so with sort of every tech right new jobs that are created some jobs disappear i'm curious what sort of new jobs do you think will be created as a result of ai well i have a what i think unfortunately is a contra an opinion about software development i think people sometimes think that all of this leverage that programmers are getting with their software development tools powered by ai means that there's gonna be fewer programming jobs i think they're actually gonna be more programming jobs i look at that app that my that my daughter did like you know the ai tools basically turned her into a programmer right and like she she doesn't want to be a computer scientist and she doesn't have time to like make herself a good programmer or she wants to be a biologist and so like i think we're gonna have like a ton of people who are effectively doing software development in places where software development couldn't happen before at all because there just aren't enough developers in the world and i think that you know we're going to so lower the barrier to entry the application creation that you're just gonna have a lot more folks wanting to create a lot more apps and this is by the way what has happened every time in the history of software right that we have made a big leap forward in leverage like you know building tools for programmers that make the more of has always resulted in us needing more programmers so like that's certainly a thing that i think we're gonna need more of and then you know it's sort of hard to predict what the other jobs are going to be i do think that the shape of the jobs that we're going to need the really really blinding important thing is going to be people who are really sensitive to the needs of their fellow human beings who are sort of thinking about like okay what are long term needs versus short term opportunities yeah like how do i do things that are like legitimately and seriously in service to you know what society needs and like what the people around me need and like the problems they're are struggling with and the solutions that you know they ought to be looking for that they might not be because they don't understand what the art of the possible is with you know this new disruptive technology and so i think we're gonna just need all sorts of new product makers yeah who have problems and they don't even realize yet what capability they have available to them to go solve those problems and like that's just gonna hopefully unlock a ton of entrepreneurship and creativity that is gonna benefit a ton of people i think yeah yeah amazing so my last question is it's kinda a two part what excites you most about what you could do with ai today what excites you most about what we'll be able to do with ai in the near future yeah i'm just really personally like this is me kevin not microsoft like i am such a curious person by nature like i'm in i i got just a near sc array of hobbies that i'm working on like right now like i literally i was sitting on stage waiting to go on for my bill keynote and i was texting back and forth with a bunch of friends about a hi guru japanese ceramics kiln that we're designing the second version of and you know one of the interesting things about that process is like we're using ai to help with it to do a whole bunch of things so like a lot of like i hi guru is like a particular type of firing process that was very common in japan in the sixteenth century and earlier and like it just isn't a lot of documentation about like how you design one of these things and like you know what all the other considerations are in like firing ceramic objects and them and so yeah these deep research agents can like look at documents that aren't in your language they can like help filling gaps that are there in the published literature and then like you can use them for doing sort of weird scientific thing like we're trying to figure out like where to place the burner port on this kiln where what you want is like a vortex a stab forming vortex so fire and it's angular and say like you gotta figure out like where the you know eddy currents are gonna be with this hot gas inside of the kiln and like we're asking the system to help us figure out how to do a simulation of this combustion process and like there's just no way i would have time to go into this depth on this thing given i've got a deja of without these deep research and so like that's just wild it's wild that in twenty twenty five we're already here right and it's just sort of a great and awesome thing for a curious person to have these tools so over the next year like the thing that i'm really hoping for is that we will have systems that are like reliably taking action on our behalf that like we sort of have transition from this mode of you know you are having to sit down with these tools synchronous and like have a session with them to like you really do say go sort this out for me and like you could have a week to go do it and like i'm not asking you to burn a week worth the gpu time but like you can make request into systems and like wait for the response to come back you can you know sort of interact with other people and other agents on my behalf and like wait for those responses and then come back to me asynchronously like i get a signal when like he got something to show me and like you know we can collaborate and iterate a little bit like that like i think just getting into that mode where that's what the ux of these things look like i think it is just gonna be super interesting and fantastic and we're gonna discover a bunch of problems that we'll then have to go solve and like that's what my talk will probably be like next year build is like it will be immediately obvious like we've got a half a dozen like complicated things about what i just described that we're gonna have to go sort out and like complicated problems to go sort out is my favorite thing yeah amazing well thank you so much for spending the time with me you today i really really appreciate yeah thank you it was a pleasure chatting with you well that's it for today's episode of the next wave podcast i really hope you enjoyed this conversation with kevin scott as much as i did a couple of things really stood out to me first how much progress is being made around alignment and safety and how seriously microsoft is treating those challenges and second kevin's point that ai isn't just about efficiency it's also opening the door for more people to build create and solve problems even if they've never written a line of code before the big takeaway here is that ai isn't just a technical shift it's a societal one and that the people and companies who lean into curiosity efficiency and responsibility are the ones who will thrive in this next wave so a huge thank you again to kevin from microsoft for joining me and sharing his perspective if you enjoyed this conversation make sure to subscribe to the next wave over on youtube spotify apple podcast or wherever you listen to podcast and if you found this episode valuable please share it with a friend maybe someone who's trying to wrap their head around the future of ai thanks so much for listening and hopefully i'll see you in the next one bye bye

welcome to the next wave podcast i'm matt wolf today i'm joined by deepak singh the vice president at aws where he leads the teams building amazon's gen ai infrastructure deepak has been with amazon for over seventeen years and has been working with neural networks since long before most people had even heard of ai he's played a key role in shaping how developers use generative ai at scale and now he's focused on building the next generation of ai agents so in this episode we dig into what actually makes something a true ai agent hears just another buzz word how agents are already transforming industries from formula one racing to drug discovery and where this technology is all headed next from multi agent systems to new standards like mc and a to a we'll also explore how businesses of any size can start using agents today and what it all means for jobs developers and the future of work if you want clear insiders look at how ai agents are going to reshape business in technology you're going to love this conversation so let's just jump right in listen up the old playbook is slowing you down ai broke the funnel loop marketing fixes it it's a new era for mark ai is capturing search traffic channels are fragmented and generating leads feels less predictable than ever before hubspot loop marketing playbook will guide you through this unprecedented disruption it's the system your marketing teams should use to move faster connect deeper and gross smarter the modern growth playbook for the ai era built by hubspot designed for today's marketer get the loop marketing playbook at hubspot dot com slash loop dash marketing to find out more and leave your competition behind alright back to the pod hey deepak thanks so much for joining us on the show today i'm really excited to chat with you about ai agents and amazon's role in this whole agent world and and all of that kind of stuff so how are you doing today i'm doing great there's a lot of interesting stuff happening at work and in the space of ai agents so excited to have this conversation awesome well let's get into a little bit of backstory i don't wanna go too deep on the stories because i wanna get into like the meat of talking about agents but i would love to learn a little bit more about how you got involved in amazon how you got involved in ai and the agent world what's the sort of backstory there yeah so i've been at amazon aws for over seventeen years it's still the only tech company i've ever worked at prior to that my life was all in mostly biotech and scientific software have a phd in theoretical chemistry it's quite interesting my first job after my phd was actually writing ai algorithms for predicting protein structures many of those techniques are completely ancient by modern standards but the first time i wrote neural network was over twenty five years ago wow over the years i've become more and more interested in two things one is develop experience and the other one is how can we use the power of computing to make the lives of developers much easier mh so as generative ai became popular and became a thing know in the last three years it was natural move to start trying to see how generative ai could impact how developers workday today so that sort of combination of their interest is what brought me into the jenny iv and agents will so i'd like to say my team is a team that cares mostly about developers and developer experience and gender ai is the most powerful tool we've ever had to radically change that so that's what i'm doing what i do it's so interesting because most people have seen how ai has exploded over the last you know call it three years maybe twenty twenty one twenty twenty two is when it really really started to bubble up yep but really i feel like before that nobody really had much interest was talking about ai but obviously there was a lot of progress behind the scenes a lot of things happening from your perspective what do you think it was like the biggest like pivotal moment that shift that got the world on board yeah so machine learning i'll call it machine learning has been around for a very long time mh you know along the way people have called it ai and depending on how you think about it people have said oh it's overs sold or doesn't work but machine learning has always worked it worked thirty years ago the ability to learn and then predict has been there from as long as i've been working with computers and those techniques got more and more powerful several years ago there are a lot of excitement about deep learning part of the reason we don't talk about deep learning is it's there everywhere and it's almost accepted that it's there everywhere in everything that we do whether it's how we shop or with how we use the internet but i think what made the gender ai such a big deal and why people saw it and we're able to get excited by it in a much more general way is because you did not have to have phd in computer science to really appreciate it because you could go and type a query into a chatbot and immediately start getting answers you know you we at this point in life now where when i'm planning a vacation i generally tend to use some ai tool to figure out where i want to go and right i want to set up my itinerary so the fact that it has an impact on your day to day life is what's me i think made it so impactful and i think that started off in the last three years with the grid customer of powers complexity with the kind of search engines start building or you know back in the david chart and of consumer side of it and of course the fun part is that one of the first use cases for gen ai outside of the pure consumer space was developer tools was programming part of meetings thinks it's because the people building these tools were programmed themselves but it's also an a problem that's particularly well suited for generative ai so you know combined the excitement that the world has with these tools and the fact that developer tooling was the first area of interest makes it even more exciting for folks like me yeah no totally i feel like the world really started shake notice around chat gp right end of twenty twenty one chat gp came out that was sort of a very easy consumer entry point into ai and i feel like that was kind of when when everything really kicked into high gear you mentioned machine learning and you mentioned deep learning and i know it's it's a very sort of nuance complex topic but can you sort of get me like the lay definition of each like what's the difference between machine learning and deep learning yeah i i'll keep it very simple i actually differentiate ml with generative ai more than anything else and sort of what differentiates between the two mh historically even when we've talked about machine learning we've talked about predictive modeling where your system learns and it's able to predict based on statistical analysis like it these are statistical systems and they're using neural networks and other techniques to predict what the next step might be the most probably well known examples of these are things like you know when you start typing an email and the next few lines get written up right or in the areas of computer vision like your self driving cars and things like that all use those kinds of techniques they're using deep learning techniques to figure out how to react in a particular condition they're making a statistical analysis of the situation the great part about generative ai and why people get so excited about is it is generating information not based on some common match but they're prob in nature they try and be for the lack of a better word they're creative which the traditional ml techniques links were not they were strictly mathematical and statistical generative ai allows you to create you know my favorite example is write a poem with these topics and write it like words virtual shakespeare and whoever i those kind like the the fun apps that you show your kids so the ability to generate from all the learning and training that has happened but try and come up with something novel and that's what made janet we are so exciting of course he found other ways of applying it well that's the big difference between the two at least from a lay standpoint yeah and it totally makes sense because you know we also get what the industry calls hallucinations right which kind of comes from that thing of trying to predict right it doesn't necessarily always know the answer which could be a feature or a bug right if it doesn't know the answer and it just makes up an answer we call that a nation it doesn't know the answer but it comes up with something novel that nobody else has thought of before but something not creative yeah now it's creative right yeah exactly but to me that's been really really fascinating to watch but so when it comes to deep learning versus machine learning my kind of understanding and i i don't know very deep i i don't have like a computer science degree or anything but my understanding is all deep learning is machine learning but not all machine learning is deep learning that is correct ml is a very broad space you can just do simple statistical approaches to things like cnn and cnn and advanced neural networks which is all lot of what deep learning is right right and i don't know if this is like an accurate description but the way i've sort of heard it is like deep learning is when you sort of give an input and it passes through a whole bunch of different neural networks before you get the output that's what makes it deep is the fact that there's the a depth to the amount of neural networks that it passes through correct i mean there's a lot of unlabeled data which and you you go through layers and layers and in many ways what the biggest difference between the neural networks that like the ones i worked down twenty five years ago there was like a single layer of feed forward neural networks because their compute power was limited and now your compute power is quite significant and that allowed you to do deep learning over gpus and start building these multi layered systems and that's actually what's allowed you to train the large language models i mean language models have been around for a long time as well right it's large language models these frontier models that are new that are able to then give you these prob responses that make them feel so magical totally makes sense everyone everywhere is talking about ai agents right now but here's the thing most companies are going about it all wrong this guy cuts through the hype and shows you what's actually working right now hubspot has gathered insights from top industry leaders or implementing ai agents the right way you'll discover which agent setups actually deliver roi and how businesses are automating their marketing sales and operations without replacing their teams get it right now by scanning the code or clicking the link in the description now let's get back to the show let's shift gears into agents a little bit and when it comes to like ai agents i feel like the definition has been kind of fuzzy right like i feel like you can ask ten different people in the ai space and their definition of an agent might be slightly different than other people's how do you or aws how do you guys sort of define agents yeah and agent in the end we see it as a combination of a set of things one or more frontier of models underneath the hood you have to have an l that's the heart of an agent mh an agent is connected to tools a set of tools that can use and that's actually been one of the biggest advances in agent systems there's the ability for these l to tools to data often like a ride system or through mc p to other things so an agent is a system with an l at its hard that then you interact with to do anything from right software to create images to check your email to do customer service but the key is an agent is given an outcome mh it has access to tools and given a goal and the tools and data is access to it's able to autonomously figure out what it needs to do to accomplish it task a great example of i think the current state of the art or in some way of systems is an example that we have from somebody inside amazon where they posted a slack of they were working with about a hundred files they needed to make some changes to all those files and they gave it a prompt make this change to he has all the files the agent started making those changes they made changes to three files he had recognized that had done the same thing to three files and it had ninety seven other files to make that change to it wrote itself a script to go and do the other ninety seven in an automated way that ability to autonomously make that decision right is what makes agents interesting by the way we build them now it actually asked you user is hey i want to do this are you okay with it and that autonomy is what makes an agent an agent gotcha yeah because there's a lot of tools out there right you've got like you know make dot com and zapier or zapier these tools that can use ai within a workflow but they won't necessarily realize if they made a mistake and then sort of go back and kind of double check themselves yeah but i feel like maybe when we're talking about a like a real ai agent a real ai agent is gonna do the work double check did i do this right are there any issues and if it finds anything sort of double back and try again yeah we call the reflection it's able to reflect on its own work go back and then maybe correct it if it sees a bug it then goes back and tries to figure out on different path to solve that bug and it's not uncommon in software development for example to just post your error log into the agent and then the agent goes okay i think i know how to fix it and you'll keep trying until it does or if it can't it'll ask for help and again this level of autonomy is what makes agents agents i think part of the challenge out there and why it gets so confusing is people will label a lot of things agents but the autonomy part the ability to track go towards a goal for us though the fundamental capability of an agent right right yeah i feel like agent's kind of a buzz word so you get a lot of software companies out there wanting to toss agent in as a selling point when maybe it doesn't totally meet the sort of true definition of an agent yeah so how have you seen agents evolve over the last few years in my mind an agent ideally we get to a point where you give it a task and you just walk away from your computer for a couple hours come back and the task has been completed i feel like that's where we want to kind of get to but most agents i feel like aren't there yet like but how have you seen that evolution of agents evolve to where we are now yeah i'll actually go you one step further and go back to the sort of early days of l mh where there were single shot you sent a query and you got a response now somewhat very quickly you started getting these multi turn agents will call them the early agents were basically multi turn conversations these chatbot that you started seeing appear everywhere you gave it a query get back a response you might say yeah but i wanted you to think about it this way not that way and they came back with the response and they had memory they kept context of what was going on but a got more powerful they got better reasoning capabilities they got the ability to use tools and i think starting somewhere in the middle of last year at at least in my mind realistically agent ai started about a year ago what this you know and what made agent ai possible was l that you could just drive and they had this ability to go use tools to reflect look at different parts and you had to give them an outcome give them the right context the context could come from here's the data you should care about here are some of the apis and tools you can use mh and they were able to then make decisions based on that i gave you one example already mh but i think that's what's allowed agent ai to evolve from just chat bots to meaningful systems where they're able to to work they could be doing software development there could be agents that help a business analysts get work done where they can you know it can create dashboard or something like that you know a great example is formula one used to have this problem where you know they're collecting data from their machines from all their cars right and let's say there's a technical issue they had to create a root cause analysis like to figure out what was happening because i don't know how familiar with you have with formula one but one of the beauty the formula one is between one race and the other the car changes right because they're look into it and they change their dynamics they change to engines etcetera so if you have an issue in the past it used to often take three weeks to analyze all the data and figure out what to do three weeks of a long time there's three races in formula one time so they created this root cause analysis assistant using amazon bedrock agents and that was able to produce resolution time significantly about eighty six percent so their triage got cut down and overall they were able to resolve the issue in three days now you can make changes for the next race which in formula one could be the difference between drivers championship or not right so that's an example of the kind of things you can have an agent do and it's that level of impact i think that makes agents are excited to all of us yeah absolutely now you mentioned that you feel like agents kind of bubble up onto the scene about a year ago is there like a a moment or a tool or something that happened that you feel like was sort of that initial catalyst yeah i would say the key catalysts is the rise of what you'll call thinking reasoning l because prior to last year elements reason and then you got the generation reasoning l right and those made a huge impact so i think that at least to me that's like the moment in time that took agents from just being multi turn assistance to being truly agent ai that was able to do work autonomously some of that i it could do before but that just took it to another level yeah yeah i mean you mentioned per complexity a a moment ago and per complexity is one of the first times that i i feel like i started to get like a little bit of a taste of agent ai because you would ask it a question and then you would see it actually sort of make a checklist and then work through and like check off the things on its checklist before actually giving you their response and for me that was when the concept started to click of like okay when you're using a chat gp or claude or something like that right you yeah give it a prompt you get your output back and you just kinda go back and forth but now when when i first saw complexity doing this it was like you give it a prompt and it sort of plans out a set of actions and then it follows through on all those actions it may not have that what did you call it where it sort of recur checks back yeah it didn't have that yet but you kinda started to get that glimpse of like okay it's actually going through and doing a whole bunch of tasks for me instead of just the one task of responding to my prompts yeah so i'll i'll actually use example of k which is an agent integrated development environment that we released in july mh and k as an agent ide that's how it works right so typically if you go back even to the early days of a developer assistance using unity ai what they did was auto completion you started typing and they would try and finish your function or whatever then he got to the next generation of assistance it's a more chat based where you know you're able to say here's my code base i want to write a function that does actually to look at a code base and say this is a here's the code for that function now with agents you can actually drive that they're pretty much write all year code mh but here's where and what we decided to do was a ux that made it much more powerful and i think it gives you a good idea of agents behave so using qui as a developer i might say hey i want to build an app that does x like you know i want an app that build a shopping cart or visualize proteins or something like that pick your problem right and you are talking in natural language and this agent collaborate with you to create a set of user requirements and a design document and dependencies and then what it does is it creates a list of tasks that it needs to do to complete its work it's gonna say it's i'm gonna to use these libraries i'm gonna use these tools this is how i'm gonna do it and as a user you can just keep hitting yes yes yes or you can say you know now i wanna make some changes i think there's a better way and use your judgment and then you hit go and then it goes ahead and does it but with this chain of thought reasoning that these modern l have and modern agents have you can see it getting stuck mh if for example in clearer the first thing the agent does it's right a test for itself because you want to write a test doing test driven development and if it fails the test it goes back and tries to solve it because it learned something from there that is i think true agent behavior right and i think over the last few months you started to see people doing such meaningful work with it and and i think it's only gonna evolve from here yeah absolutely looking forward a little bit you know mentioning evolving ai where do you think it's all headed like where do you think agents are gonna be a year from now maybe five years i have a hard time seeing five years for the future i have a hard time actually guessing what we're gonna see next year i'm usually pretty wrong i'll say this is coming next year and then the thing i predicted is coming next year comes three months later but i'm just curious if you have a vision of where you see agents like a year from now or maybe five years from now where is this all headed yeah five years is really hard in this world which evolved so quickly i have this joke that anytime i present a document it's obsolete by the time i get back to my desk because of how quickly the world of ai and who these days i think there's two or three areas that it's going to move one is even the agent systems we have right now are somewhat limited they tend to be a single agent not multi agent applications at most they have two or three agents in them mh you're going to have applications that are connections of agents like lots of agents talking to each other asking each other for help in an autonomous world as well as we drive more and more autonomy into the systems you'll get these more complex applications and you're tooling to do those needs to evolve as well you know couple of months ago we launched something called amazon bedrock agent core and an agent core is a set of primitives but the idea is somewhat simple you know at least conceptually which is you need a core set of primitives to run agents at scale you need secure safe front times where you know agents can get isolated from each other and run in isolation you need a way to identify an agent an agent isn't a human but it's not a traditional service either it has a difference you know how do you give an agent and identity how does that agent communicate with the rest of the world and by the way that may not even be a digital world it may be running on a robot where it needs to communicate with the physical world and then how do you go and audit what's happening to these agents so and the protocols and mc and k are all improving underneath it this is the boring part of running agents the operational infrastructure and capability the running agent application of scale is going to explode into the maturity over the next year right yeah agent imports two three months old and that shows you it's just getting going but by next year it'll be very much mature it'll be generally available and people will be running large scale agent applications on something like agent core right that's one what do you think the bottlenecks are right now to you know being there i think the big ones are just understanding six months ago there's twenty ways of i'm doing these things now there's a set of sort of standard protocols that are developing so at least two agents know how to talk to each other it's instead of you having to write bespoke code most of us are hand crossing run times to run these agents now the standard ways of running these run runtime so the barrier to entry gets lower right you don't have to be a team that's an expert at building all of these components you can use managed services from somebody like aws to go run them so the barrier entry for a customer or any business that wants to build these agents goes down significantly so that's one the second one and i think this is the one that most people talk about when they talk about agents is the level of autonomy mh right now for the most part you still have humans in the loop a lot like in software development you have people driving the behavior of the agent you're are prompting the agent and even if it goes in dust dust in the background you still have a human they don't run for two days right they'll run for ten minutes fifteen minutes twenty minutes maybe a couple of hours but you're going to have systems that run much more autonomously mh when you're doing that a few things happen one your build trust and the quality has to be there and so you need systems that can verify and validate along the way that becomes even more important because you don't have eyes on the problem so at aws we believe in neuro symbolic ai techniques which help you verify the behavior of an ai system make sure it's giving you the results that you want to that are safe and so on there's many techniques that's one area that we are heavily invested in and then how does this all interact with each other so the level of autonomy i hate to say this can almost like self driving car world right in some ways the level of autonomy is gonna go up right and then that happens how businesses companies are integrating ai into their day to day work changes so i think that's gonna be the other big one beyond that your right guess is as good as mine and what do you see is amazon and aws is role in all of this obviously you guys have bedrock you know you've got your cloud services and things like that what is aws doing to sort of progress agents forward yeah and i'll go back to agent core and bedrock in the end to run an agent applications you need great agent infrastructure mh so at the most core level we want to give you the best agent infrastructure to run your agents that includes something like bedrock that gives you l choice so you can pick your then you need a way place to run your agents which is agent core run time so this is it's server as it's secure it's isolated we've made a lot of innovation on the micro vm side like these innovations that we've done over the years like the nit system and micro vm they're taking advantage of them mh to run agents in these isolated sandbox and scale them we support any ll any agent framework you can write to agent how you want it we want you to run them at scale we are investing in agent core identity so be how do you identify an agent what identity does that have in agent core gateway which allows you to connect agents to the rest of the world agent code memory agents and memory short term and long term they to remember what they did to reflect and make decisions and at end of it you want audit well you want ob for two reasons you want to respond in real time if you have access to all the metrics and all the logs you can do that well later you can also go and see what are all the decisions has this agent taken what did it interact with at what time did it do so because those become important so at a base level we are building the most high performance scalable and lowest cost infrastructure to run your agents it's what aws has been doing for applications for a very long time the second area that we heavily invested in is building agent take applications ourselves mh so for me that will mean agents to help you build software we are increasingly using agents in other areas for example we have a something called connect for customer service which increasingly has agent ai capabilities inside it to help customer service agents and customer you know virtual call centers be more effective and those are good examples we are building these ai agents into the fabric of how aws operates that our customers can benefit from it they don't need to know anything about agents they're using them how do you see amazon and aws approaching responsible ai like is there specific things that we feel like the human should always be in the loop on or do you think we're gonna get to a point where we're totally just trusting this to go and handle the task we give it yeah very specifically today for example if you do software deployment we don't have any tools out there to help you do too well we have tools to help you but we have no agents that it deploy software for you we want the agent in the loop there are other areas where we'll give you the tools and recommendations but you as a company can decide whether you want to apply those or not and i'm thinking at a fairly high level right now for example if you're writing software with an agent our recommendation is and we give you tools to do that is to do code reviews but you may be a two percent start that goes yolo mh i'm not going to do court review i we'll trust the agent and just merge this code in and ship it right but that's the decision you make we'll give you our guidelines but deploying software we don't give you any agent that helps you do that today as an example the area that we are really really invested in is so bedrock has bedrock guard mh that if you are building an agent is actually something that a lot of folks don't always realize based on my conversations with them you have to make sure that the prompt that the end user is submitting goes through a set of guard before it goes into the agent right so that you make sure that there's nothing malicious is gonna happen and then on the other end you need to apply guard to make sure that the results that are coming out meet your company standards don't have anything malicious etcetera or toxic in there so the set of guard rails we publish them and with better guard rails we allow you to tune it so that you can do it based on your policies etcetera for example we use for our own applications we will use under l that we believe are a high standard themselves that have built in guard rails but then you can apply layers on top of it that's one area the other area is your favorite one hallucinations mh how do you make sure like i'm a company that needs to provide factual verifiable information and like you as a consumer asking even the price of something mh i do i make sure that i give you a the real price and not a made up price right unprecedented price so we have a new feature as of two months ago in bedrock called automated reasoning checks these automated reasoning checks use mathematical formal models that you can create to verify that something is factual actually correct so for example if you're asking me a pricing question i'm the agent we can build these checks into it which make sure that when i'm giving a pricing response it's a factual correct response but this goes back into these automated reasoning and neuro symbolic techniques that we invest in very heavily so in addition to the guard around how do you make sure your agent is not using bad language or using style that is not appropriate for your company you can also do things around lines are making sure that the accuracy of the responses is very high with using these some of these other techniques so those are the areas that we invest in ourselves quite heavily both from a research perspective we're also a product perspective gotcha yeah so basically there's guard rail so when somebody enters the prompt it kind of passes through the guard rail before letting the ll even see what you prompted and then when the ll gives a response there's another guard rail that that response passes through before the end user sees it and you can kind of tailor what does or doesn't make it through some that's built into the l some a lot of l are for example some of the products that from my team we son it four son four has built in guard that are pretty good but we put additional guard on both sides to make sure that the quality and responses are appropriate yeah makes sense there's a couple terms that have been brought up a a couple times throughout this interview and i'd love to help sort of define them for people listing that may not understand them you mentioned mc model context protocol and you mentioned a two which is i believe agent to agent agent to agent can you help sort of define those and and what they're used for within the agent process yeah mc is the world's best hammer everything looks like a nail i'm only half joking even nine months ago when sort of these reasoning came one of the first things that we observed was that l were now capable of using tools they have been for about a year and a half mh and the tool could be ability to read and write files for example about a year ago maybe a little more we published blog post about something called text editor mh and text that it was a but just a text editor like an editor that we had made available to the l to read and write files to look at file look at text to delete it to change it and that's a tool so people are writing tools for their relevance and their agents in a very bespoke way we have an agent framework called strands which before we released it publicly had its own tool system like how to make tools available to it mh but then mc came around and what mc is just a standard protocol to make tools available to an era mh and what's happened is it organically became very popular to the point that pretty much everybody supports it now so if you want to give your l access to anything whether it's an api another agent an sdk a set of tools that you've written you can wrap them in an mc server and make them available to the and the great thing on mc is those tools can be any as long as somebody making it available as an mc server mh it is now available to l as long as you're giving you permission to access it so right mc in the end is a standard protocol a call for two access that's alright about it right so you you have all these software products out there right all sorts of saas products that you might be able to develop something and tie into it via like an api right but the problem is everybody's api is sort of developed different they all have their own documentation there's no real standardized api everybody kind of holds their api differently and so when you want an ll to go and talk to some of these tools telling it to go talk to the api there's no sort of universal way of doing that and i think mc my understanding solves that right it kind of creates this version drops yeah yeah it drops it in a way that an can understand so anyone can now write a tool the reason i say it's a world that hammer is it's like if you talk to anybody but i want you to do this that's like oh just put it behind an mc server so it becomes sort of it's it's it's almost a half joke in the industry now right but it's a very powerful tool it's really powerful hammer and it's changed the way we think about how agents are written yeah because it's a standard and then what about agent to agent how is that different from mc you know how does that fit into the workflow mc is about dual access now mc is going beyond dual access this capabilities coming into it but that's where it started mh agent to agent is a protocol it's like a wire protocol for one agent to talk to another agent so how does agent a know how to talk to agent b and if you have a standard protocol to do that all agents can talk to each other so i think of it more like a networking or sort of buyer protocol standard like you've got things like protocol buffers and thrift back in the day they're allowed two services to talk to each other mh it's just a communication protocol between two agents that's the best way to think about it i don't know agent to agent communication is so early mh so i think that's how a and standards like that will have always less clear to me mc is pretty much ubiquitous at this point of time right right so i wanna talk a little bit about like real world applications of all this stuff so this show actually has a fairly large b to b audience there's a lot of people listening that they wanna know how do i leverage ai how do i leverage agents how do leverage some of this technology into my business to make things easier so there's a couple things that i wanna sort of go down with you real quick one of them being can you share some real world examples of agents you've seen then after that i would love to sort of dive into if somebody wanted to go build their agent can we sort of walk them through here's what you need to go do to do it but first start with some like real world examples yeah so my favorite one maybe because of my past life is one from gene gene worked with us to build set of agents to improve their drug research project they use amazon bedrock agents amazon bedrock agents is a great way to build simple agents you know you can't mix and match too much stuff but it's gives you very simple path to building an agent mh and what they wanted to do was they have to all these therapeutic areas and they collect millions of biomarkers and they need to validate them it's actually very manual process having been in the industry so they built an agent that automate biomarker validation you know as scientists generate these biomarkers but actually they're coming off instruments before they decide which ones should be sent on for more experimentation or to build a new drug they validate them as this even real but this even make sense and so they built an agent along with us to improve this process and they safe automated about five years worth of manual work by doing that and so this helps them do drug target identification improve their research agency and so on and so forth and a more sort of consumer application is door has the ai agents and their contact centers you know so they do for voice and chat to answer frequently i ask questions i mean i don't use to as that often and when i do if there's a question i suspect there's hundreds of people or similar questions right have agents have become very good at answering those the key part about a jenny agent it's not a set of canned answers they're responsive as said they can react right because they've access to my order they have access to my data they know what i've done before they can give me an answer that's relevant to me right so they're able to feel hundreds of thousands of calls per day they can actually respond to dashes like guys that's the main one when a dash has a question about what they need to do about order or they need to clarify they're able to get responses back in like two and a half seconds they don't have to go look somebody who might know the answer right because they have access to their own system and the dashes the what the dash is doing to the order and so on and so forth so these examples are the kinds of agents people are building and there's obviously from my perspective the best examples of the agents that we have built ourselves for software development right which are allowing people to build powerful systems and change the video of engineering teams you know right you're using agents write software yeah that makes a lot of sense and i imagine with like the door dash example it's either fine tuned with all of the past sort of customer support queries so it's seen how all those have been handled or they're all you know dumped into a document and retrieve via rag or something like that right like they're able to actually see all of the past histories see how everything was responded to in the past and there's probably enough data there that the ai could respond pretty dang well based on everything it's seen throughout his yeah and with modern than sort of contemporary agents it's not just the fact that they've access to the data it's what they do with the data so the data as part of the context you know i i often say context the king the better context the better the agent does but because they as you said they have good context their agents able to be very useful to the end user because it is able to interpret natural language questions so much better right take action right right yeah that totally makes sense now when it comes to let's just pick a random example let's say there's an online commerce business and they wanna set up some sort of agent to help them out what's their first step like what what do they go do to sell their first agent so i'll give you my recommendation because it is kinda amazing how easy it is to use in some way especially for developer agents need a framework to build the agent frameworks are you know lack a better word it's their libraries or some kind of diff development framework that makes starting an agent easy and allows you to ship it it's been frameworks coming mean around for a couple of years the early frameworks were quite complex because they had to work around the limitations off the l at the time they to do a lot of the heavy lifting themselves but as l have become more powerful and a significant you had to build chain of thought outside the l but now l themselves are capable of reasoning right so what becomes more important is how do you drive the behavior of l take how do you make help it reason better so we built something which we've seen so concern was called the strands sd sdk so library i think it may be more than one language now the original one was just in python mh the python library that you use to essentially drive agent behavior to what we call an the agent loop or model driven development and you put their code you describe your agent behavior and trans has access to tools it has access to mc to access to way to away it's able to help you build that application very very quickly just the way you would build any application as a software developer if you want more the drag drop i won't be just pull things out of it you something like bedrock agents it's a drag and drop system and then when it comes to how you want to run your agents that's where agent core comes in it feels just like any other aws service it has apis it has a console you've built your agent in strands for example you don't have to use strengths but let's say you have you can deploy it on agent core run time you can use agent core memory for its reflection if you want to do reflections you can use agent core identity to give it an identity you want to connect it to fifty other things use agent core gateway because that gives you the connectivity and also the ability to manage traffic all the things that you want to do so from an aws perspective what you would say is use bedrock to host your you know pick an l you can use any framework or stranded it's just open source you can use strive to build your agent and you run the agent on agent core but the way we've built agent core you can make and match let's say you want to use some other gateway mh you can use that gateway with the agent core run time and the tooling around this is just gonna get better and better and better so does agent core is that like the gpus where the inference is happening or it's one level higher that's bedrock agent core is one level higher than that okay the way to think about agent core run runtime is it's lambda for agents that's the simplest definition i can give you it's a server less run runtime but you're not putting functions on it you're running an agent gotcha okay cool that makes sense now i'm sort of wrapping up here that but there's a few questions i'd love your take on when it comes to like the whole you know ai agents might take our jobs right that's obviously a concern that people have what's your take on it what sort of advice do you give people that have those concerns i'm gonna speak from a software development perspective because that's the world i live in mh the better and more you know you are as a software engineer the more effective you get it using these agents mh as a software developer you get stuck all the time and when you're using agents you can get yourself on stock much faster there are problems that you've not wanted to solve or you haven't had time to solve that you're able to get unblock on so that you can solve really hard ones mh the skill that i tell software developers to think about and you can see this because you can see the ones who are good at that be very effective at amazon we have a principle engineering tenant called illuminate and clarify our most senior engineers are really good at it mh they're able to illuminate a problem and clarify simplified break it up into smaller things and any software developer and they supplies to any domain as you look at a problem how are you able to explain it how are you able to express it in natural language mh potentially break it up into smaller chunks and give that context to the l to the agent the more effective we are going to be how do you interact with agents is a little bit different than how you may have interacted with an for writing software in the past but once you figure that out the outcome is it much much better that's what it gets people excited so you have to change the way you think your type thinking in natural language you're thinking in terms of breaking up problems in fact the whole goal of hero was how can we give you a user experience that actually makes it easy for you to do that right how can we help you think like that and then express them by that's how your interface works but the idea is that the person at the other end of it the more you know if you're a really good developer who can think like that you're gonna be even more effective yeah that's very much the approach we took what you're not doing is typing code yourself but that's okay that's busy work that's typing right right yeah and i mean that sort of answer the next question i had lined up which was going to be what advice would you have for like code and engineers do you think it's still a valuable path to study absolutely i mean we see people being so effective and over eighty percent of the software developers at amazon using yeah tools for their work and the thing that i tell them or anybody asked me is it's less about how well you know javascript or ruby or rust stop python it's more about how you think about breaking a problem down giving a goal to an l understanding how to express that goal and then having the skill and judgment to understand to help it move along like if it's going in the wrong path how do you tell it hang on and a lot of what my team is trying to do and what aws is trying to do is give you the tooling that helps you do that more effectively yeah that makes sense i i actually had a conversation recently with kevin scott who's the cto over at microsoft and during that conversation he actually mentioned that he believes that the rise of ai is creating more code not less and that people starting to use like the these various coding tools they're actually starting to make progress which is getting them excited about code and when they run into bugs they go and want to learn more code so that they can better troubleshoot their vibe coding right and he actually believes that we're gonna see a rise in people wanting to become code as a result of ai not the other way around at happier quote i'll give a very concrete example you've probably heard this one before andy talked about it we built you transformation agent that takes java code from one version to the next version right and in the history of any software company when you have to do version upgrade a pin and pain because you have to go packaged by package every team has to do it and there's been your millions of package that takes a long time like years sometimes because it's worth that nobody wants to do right so what we did was we took our transformation agent and plugged it into our build systems and how we do work at amazon and a small team was able to transform very last hundreds of thousands of packages if i remember correctly and what they did was the ship code reviews through the teams that owned the package the teams didn't themselves didn't need to do anything other than code reviews mh and the outcome was it saved us forty five hundred developer years and two hundred and sixty million dollars in arr from the infrastructure savings the interesting point is now suddenly you had the people who originally would have been spending their time doing the upgrades actually writing software and building creating and inner innovating which showed much rather do right and i think that's what speaks to why you as a developer you're gonna get more excited and they'll be more because you can accomplish more absolutely so you know i think this is probably the deepest we've ever gone on the show into ai agents and so like i think this really sort of gives people a pretty clear picture of what ai agents are right now and and where they're all headed so i really really appreciate you hanging out and taking the time with me is there some sort of resource or url or somewhere we should send people after listening to this episode to go learn more about what you guys are up to yeah so for my team i'll recommend people go to keto dot dev that's a website for k it be a good eye yeah how we think about software development in an agent world mh and if you want to learn about agent core another agent stuff that we're doing at aws go to the aws website and search for agent ai you'll find all the stuff that we do there and you know we love feedback for hero for example your discord channel join us there user it tell us how with what you'll like to see our goal is to make it easier for you to build and run your agents and use amazing well deepak this has been a absolutely fascinating conversation i've had a blast chatting with you so thank you so much for hanging out with me today and going in deep on agents with me yeah thanks for having me and that wraps up today's episode of the next wave podcast a couple of big takeaways from this conversation with deepak first ai agents aren't just chatbot bots with a fancy name their systems with autonomy reflection and the ability to actually solve problems on their own we heard real world examples from formula one teams cutting analysis time by eighty six percent to gene tech saving five years of manual research work to amazon saving forty five hundred developer years and over two hundred and sixty million dollars by using agents for code transformation so the big lesson here whether you're running a startup or a fortune five hundred ai agents are about to fundamentally change how work gets done and now is the time to start experimenting with them i wanna thank deepak again for joining me and for sharing such a clear inside look at how aws and amazon are building the agent future and to everyone listening if you enjoyed this episode make sure to subscribe to the next wave podcast on youtube spotify apple podcasts or wherever you like to listen that way you'll never miss the conversations that help you stay ahead of what coming next thanks again for tuning in i really appreciate you hopefully we'll see you in the next one bye bye

what i told you build a fully functional web app in under sixty seconds without writing a single line of code and not just a landing page a fully functional web app with a database user authentication and everything in live on the web today we have on eric simon to poke under bolt when the hottest ai startups and eric's is gonna show us how to use bolt to create an application that before the cost hundreds of thousands of dollars to make and now it costs next to nothing so if you've ever had any different app before and got stuck because you couldn't code you couldn't find a technical c cofounder this episode is gonna unlock so many new opportunities for you let's just jump right in listen up the old playbook is slowing you down ai broke the funnel loop marketing fixes it it's a new era for marketers ai is capturing search traffic channels are fragmented and generating leads feels less predictable than ever before hubspot loop marketing playbook will guide you through this unprecedented disruption it's the system your marketing teams should use to move faster connect deeper and grow smarter the modern growth playbook for the ai era built by hubspot designed for today's marketer get the loop marketing playbook at hubspot dot com slash loop dash marketing to find out more and leave your competition behind alright back to the pod eric thanks for coming on the next wave awesome yeah thanks for be excited to be here you know so i've been following bolt for a long time we have kind have a weird connection because the first time i sold bolt was your episode with greg you know greg was one of our very first guests on the show and actually when i was thinking about starting the show with my friend matt wolf in hubspot i went for a long walk with greg here in kyoto when he was visiting i was like should i do this podcast thing he's was like yeah he'd be great at you should do it and i've been doing it for a long time it's hard but it's worth it and so that's kinda like our connected through greg iso i love it that's awesome yeah it was funny yeah that was like one of the first podcast podcasts i really went on after we had launched bull and it ended up being huge like we got just you know massive number of people then coming and trying out bolt as a result of be being on greg so he's like it's an angel in the company and that sort of thing now tell that's awesome you stoked talk about bold is like you guys are raising you know i'd love to put in a chance yeah that's great yeah greg is also doesn't even invest in that when he started up that's impressive maybe you should demo a quick demo of like what bolt is like what you know what you can use it for and that sort of thing that sounds good yeah for sure let's just do that just jump right into there's almost a year ago exactly we launched pull and the idea is like you know what if chat g could actually make web apps you know and and mobile apps you know now there's like a lot of these things i can do that but both was really the first one that could actually build full stack web applications just by like typing text in the box and hitting enter right so i go and share my screen and we can like go ahead and use this thing and both improved a on you know over the past year since we originally launched it but we've like stayed true to the roots of the thing of just having it be this really really simple interface where you can literally it's just a text box and you could type in what you want to build with it you know for the purposes of this demo let's just say you know make me a dashboard for viewing company metrics or something right and so you can imagine like a very common use case folks use for both or like pms designers at companies that are using this instead of fig to prototype out right their applications because it's like way faster to just type in words into a box yeah and you can show something real that actually works and like it's easier to talk about that versus like some design file exactly so if you kinda come back to like you know marty ka wrote this book you know it's kinda like the foundational like how to build products sort of book i think it's called like inspired if i mistaken but yeah had this like a concept of high fidelity prototypes mh and really the quality is you you you're gonna make a prototype you want it to be as true to form as possible to get the most highly accurate feedback from the end user on whether or not it's like something they would use right right and prior to this type of tool existing the fastest way to create a prototype was dragging stuff around in fig then basically showing people jpeg mean like does this look good which is like better than nothing and then like if it if if it passed that test then maybe you get some engineers like spend a couple weeks prototyping something that is a real application to put in front of them but it it just took too long it was too expensive to just yeah it feel like there could be so much loss in there too like you think you understand the idea but when you actually eat your hands on it's like oh i don't really like that exactly right bolt is in like kinda flipped that around where it's like you can just like come in and start prototyping an idea that's like a real application you can try out right like i just type some random thing in and it's kinda ahead and you know i've gotten and created a a dashboard that we can like look out with our different business metrics like whatever have you and you can actually pull in like your company's design system to this thing so we'll actually build applications and use the libraries your company does so it's like on brands right with with your company's ops how do you pull that in it's that like like fig or something like that it's like yep a library of like your fonts and colors and all that kind of stuff absolutely yeah so you can actually pull it in from your fig mask so if you go to the bolt home homepage page it's like a fig button you can paste it in there so kinda cool like in front of anything my url you can just put bolt dot new in front of it and then it'll automatically suck it in you know if you have a coded version of your design system like an mp package or something you can actually configure that in bold and say okay our design system lives in these packages in our mp registry and it'll actually install from there and use those things now that's all we're like the one the i think the only company that actually allows that at this point we're you know where we've designed to be able to pull in from arbitrary registry is pulling from fig etcetera so yeah a lot of companies use us for you know rapid prototype prototyping etcetera in collaboration you know amongst folks on the team and what's cool is like as you get feedback you can just go to bolt and you can you know prototype and you know give it directions say let's like make this optimized you know for mobile devices you know and use a blue color theme or something yeah i know you can keep interacting with it and chatting with it right and building this interactive so it's like very simple you don't want to be technical to use the product and so it's been a game changer you know across the spectrum for businesses like pms especially and even for entrepreneurs that are like going and building their first application ever and like trying to launch a startup on there cool and i assume you get all the code for this too that you could hand off the developer once you prototype it yep absolutely yeah yeah and to that's like one of the most magic things about this entire transition happening where you know you can have ai just spin out actual code this is using again the same code your company does or you know if you know if you're starting a company it's using like react which is like typically the default choice or whatever whatever have you you can instruct it to use anything for that matter but yeah you can take this code we you we actually have an integration with github ups so you can like commit this to a repo and like have it be committing there we have integrations for deployment and that sort of stuff built in so this is not like some proprietary thing like you know maybe wix or squarespace right sort of as mh this is like something that's actually really focused on real workflows and use cases i was kinda thinking about like who is your competition i know there's like lovable and and versa v zero and then obviously there ide ids like you know cursor and win and then there's agents like devon and factory i i am kinda wondering like where this all goes and like they all converge in some way or are they all serving different markets i kinda see devon factory a little bit more targeting enterprises and enterprises that also have engineers it feels like with what you guys are doing is great for like replacing the website builders and things like that but maybe you're thinking about it totally different that i am i think that it's it's somewhat the website builders i think for us we're very focused on having to be a tool that people are using for rapid product development so again i think like the fig use case of like we're making prototypes giving feedback and iterating that strikes me as really our lane and where we are having really really great results mh we have two ic that we see on our pros consumer side it's a largely filled with people that entrepreneurs building their startups that sort of thing on the b b side it's entirely like pms at companies doing rapid prototyping and they're kinda largely the same use case both sides they're are trying to iterate integrate towards great products the pros consumer side of it though they really really care about having production grade hosting and things of that nature which like the business also care about in regards to like creating real prototypes they can share internally and that sort of thing there's just a little bit of weight emphasis difference between what these two different types of users are doing but it's like kind of a classic product led you know growth sort of emotion for us and do you guys provide kind of hosting your like there right now or or or no yeah so by the time this is live like so tomorrow we're actually announcing this but basically we have actually integrated hosting directly into the product i'll go and hit publish so yeah yeah this is okay case this is actually like a time production on feature fly right now so we get to see it actually in action this breaks by the way it's because i'm feature flight and this is like going yeah but what's really cool about what we're up to here with hosting a lot of the tools today for vibe coating they're very fun they're kind of like they're fun toys but they don't actually set you up to build and scale your product two millions of people even thousands of people and there's been a lot of stuff in the news around databases being deleted by other folks and like security issues etcetera and a lot of issue is that the vibe coding tools like the web based ones like lovable and wrap it whatever the way that they've approached posting your applications that you make is that they've kinda tried to home their own hosting infrastructure mh and the problem with that is that all these companies are around the same size like we're like forty five people so it's liable so is replicate if you're gonna go and scale databases and servers and dns to do it well and reliably for actual businesses it's a whole business in itself and so what we ended up doing is we have partnered with null and super base which you know are the best companies in the world yep for databases and hosting and we've actually taken their offerings and for the first time ever like brought it underneath one interface like one seamless experience inside a bolt where you're not even like you're not signing up for an ne suit base account for it's actually just part of both like every bolt project has its own production hosting and database associated with that so when you create a project on both it automatically has those things and it's being scaled and supported by ne and super base with us so you can actually go and scale things to millions of concurrent users right with these different tools yeah so both is actually like really the first vibe coding tool that's gonna actually have this amazing type of you know idea to million sort of experience right yeah that's amazing so it feels like you guys could really in theory dominate like zero to one basically right like when you're when you're creating something new and i can see why greg invest in this you know he's got the whole idea of marketplace and all that right so you got an idea you wanna create something whether you're an individual or the company and it sounds like you guys could be the best at that like just from idea in your head like it's an it's now a real thing and it's actually published on the web which is is awesome yeah exactly and your spot on which is like the people that are primarily using bolt are not developers right when you look at cursor the inverse is true like nearly everyone using cursor is a developer yeah the people are using bowl are people that have great taste in products and and business right and they have good intuitions around those things but have never had the technical knowledge on how to actually build you know a product themselves from the ground up as far as like writing code and that sort of thing and so both for them is really their tool to actually go and take direct control of the building their products and experiences and to that end it's been a very important part of that and there's like a lot of the growth of the other guys too in the space have been seeing from this type of archetype but the issue is it's not just about building a prototype or building an app you have to have this thing wired app to be able to actually be secure and actually be able to choose scale and like not have these crazy obvious holes right so then that just strikes us as it's important to have as a first grade thing in the experience we are actually hosting bolt on bolt hosting like we use the exact same stuff yeah for this and you know and even like lovable also uses the same stuff as what bolt hosting is right the tagline is you can go from idea to production or whatever in a prompt if the five tool itself is not using their own hosting that's like a it's a big red plot right right right alright and there's been a lot of this where like everyone's saying the stuff is secure and stuff but it's like maybe if you really believe that like are you kinda selling people one thing and not actually using it yourself right it kinda goes to show though like where what we think is important from the user experience help people actually get their ideas online in a secure way that will actually scale right i think that's awesome yeah it's on valley there's huge bias towards engineers which i mean i love engineers i code myself a bit people who were like great at business or marketing or sales they were such a disadvantage if you do have an idea okay kid now you gotta find developers and pay them at least ten am maybe ten twenty whatever to build some kind of prototype of something really simple and you have no idea what they're doing you have no idea if their like how much they're charging you if if you even make sense if their time make any sense and then maybe it doesn't works like now instead that you're kind have a world where or someone who's good at talking to people or who has good ideas or is a good good at doing social media marketing or whatever they can have idea and just start talking to bolt and then the website's live and then they can just go test it and go share it if it doesn't work you try something else i've told people that you know people don't realize like you can launch a lot of things now people used be worried about like launching something and it failed and like oh my god people are gonna think you're a failure people forget things like a week later now like literally we just tried new ideas and and seeing worked with something like well people absolutely look vibe coating is the fastest way to go from idea to shift product and most marketers are completely sleeping on it and here's what's wild hubspot just dropped ten vibe coating prompts that'll take you from i have an idea to an actual working product in hours you get the complete zero to ship framework with prompts at every stage from idea creation technical optimization and scaling it's the entire package get it right now click the link in the description now let's get back to the show for like people who are entrepreneurial i think this is the best thing really in the past decade that's come out as far like the capabilities to be able to iterate this fast because like to me i look at early stage entrepreneurship like finding product it's purely a numbers thing right like because said how many shots on goal can you take like how many iterations can you take and how fast too you right like the that speed matters a lot bingo right your speed dictates how many shots on goal you get yep right and this increases that by like an order of magnitude that to me is like why every entrepreneur should be using these things because the it's never been faster to go and invalidate or invalidate the ideas that you have yeah and heck you don't even necessarily if you're not a technical person you don't even need to to necessarily be technical like go and actually do this and we've had like just you know from what we hear from our own users we've had just dozens and dozens of startups that have used to build their their mvp go improve it out and have gone and raised their seed rounds to continue scaling their company out and i love hearing that because i mean what like a great enabler you know right and all the integrations we're building in it as well like like what we're doing with like our bulk cloud offering stuff again it's like these are things that if it wasn't just built in you'd have to go ahead and figure out how to set these things up and get them all configured and data dah and then like and if you're not tactical it's like right the agent's not gonna know how to use these other things right so having you all built in it's like now you can come to bull and be like hey buy my domain like deploy that thing to production like go and add this feature and just knows how to do all this stuff across all of the infrastructure that you don't even know what it is and how it works but it does right and and you can build a business like amazing yeah yeah it is amazing yeah you used be i mean like even like learning how to up a domain and properly and dealing with dns with like a skill you know like at least me me back of the i was like okay have to learn how to do this like could just mess with dns and everything was a pain i think it's like a good moment to like talk about the hack if fits really willing to talk about all the new opportunities for people be able to create things so maybe like tell me about the hack and show you some examples as well like what's you know what people built yeah so back in i think was like march or april this guy k runs the start stuff over at paddle he tweeted out this idea is like if i was the ceo of bowl i would throw the world's largest hack fund you know and have a you know giant cash prize have guinness world record actually like validate the world record title and just as a lightning rod to like get entrepreneurs and builders to come and actually try out you know these new tools like both a lot of people are not aware that they can actually go and build startups and and and you know tech products themselves with old and other things i mean so i i tweeted it back i was like cool it's i didn't know who he was i was like let's do it and a couple hours later we had like over a million dollars of commitments and the thing went like viral on twitter and whatever and so we got rope updated running the world's largest attack on which i will tell you retrospect i mean it was heck a lot more work than i would have ever expected but it was it was it was incredible and we had i think we had like over a hundred thirty thousand people that participated in this thing wow and the top folks that one the thing it's you know it was amazing like i think the first prize is like a hundred grand the second prize is like seventy five or fifty did you say a hundred and thirty thousand people a hundred and thirty thousand wow huge massive massive it was just it was just crazy yeah our cloud and inference bills were through the roof in bottom but bam it was it was such a cool initiative that we ended up doing because there's so many i mean so many people that came from all different walks of life and backgrounds some people there were engineers most people were not they had never and they were just like hey i'll give this a shot like i'll go and you know build something cool and so this heck of them was all in person they all had to be there or it it was all global actually i was oh okay okay and we you had an in person components kick the thing off the chains smokers through a concert for this that was how this thing kicked off believed or not they're like investors in our company and so fancy san francisco there was a concert to kick off this hack farm which is pretty cool other than that it was like pretty much online though best i felt the one first play super cool guy he made a ai app that actually lets you like edit videos with ai so like i honestly done a lot of video editing it's kind of an anonymous a little bit yeah so it's like the idea of having an ai integration or agent where you can actually just tell it hey like go and take these parts clip them together and dot it just like does it pretty magic right right do ended making you know a actual app for this that's cool you know i i own lord dot com there's been so many people who told me nathan you should be doing that with lore that's what laura should and it's a really good idea you know was really sophisticated application actually because it he had a full timeline scrubber and everything and i mean it was like a full on you know project and he used both for all of bolt was the tool that you used to premiere with this so you can kinda see like this is not yeah you know it's super trivial stuff to make here right especially in the course of just like a month like pretty wild so you can you know chat with that and actually tell it to edit and whatever have you but was like this is like a reasonably sophisticated application right mh this like a real start mean you about like couple years ago the idea of a one person building something like this is this would taken like a year to build solo right right but you know made this you know entirely through prompting this is the guy that one first place and then on second place was this guy as named search so he made this app called weight coach and basically the idea is it allows you to take control of what you're eating and it'll help create recipes for you based on the ingredients you have and help you like you know actually get down to like get the target way that you're looking for yeah it actually uses ai to go and like create recipes and create different attic aspects of the experience which is like really cool because there's like these weight loss apps you know that you can download but his is like i thought was very interesting because there's something where it actually got to know you and like your preferences and it wasn't like hey here's like five different aspects that are like hard coded into this it's like it can actually go and dynamically adjust to like oh i'm traveling okay well here's something you could do in that area blah blah right so right but it's like very very cool app that he made and is actually a native mobile app so it's not just like a web app it's actually a native app which is something unique that bolt is the only slash one of the only web based tools you can like make a native ios or android app with the thing yeah so there's just a couple of cool collect projects that came out of that oh very cool that's awesome i was still thinking when you were talk about this i was remembering back that one of my cousins in alabama was asking me like hey i've got this idea for an app you know you to have tons of people in my family he like you know because i knew i was you know having some success in in san francisco and they're like i have an different app you know how would i make it or how would i get started and i was like you gotta go learn this programming language for like a year are you gotta learn whatever and yeah then maybe make a prototype and it's probably not gonna work i just told him that too was like it's probably not gonna work you got this idea it sounds great you probably should do it though because you're gonna learn a lot you know and i do wonder like if they would change things that they could've have just like went to something like this and just tried it in in a day saw if their idea had any merit or not you know that'd been awesome absolutely yeah i mean i i think and that's like you know to my c founder and he and actually group down the street from each other into summer chicago and he and not learned out a code together and we're were like thirteen and and we've been building stuff together ever since and awesome it's like we look at this where like we're building the thing we wish we have and we're were thirteen years old because right it's so painful to have to learn out a program from scratch and the drop off is just huge like a lot of people have interest in this stuff but just to do anything it's like you there's all this random kinda archaic things almost you have to become good at that's really ancillary to the point of act actually building something yeah and then you're just getting started and that that doesn't mean that you're gonna have any success at all hundred percent a hundred percent right and so i think that it seems and this is like greg says this vein end of thinking a lot but like this is just worst you know already started see an explosion here of ideas and products that just would have never existed before yeah because you know the sheer drop off that inevitably happens if you have to know how to set up dev environments how to code and like and then getting good at coding to the actually make something that's beautiful and like secure can scale and stuff i mean that's like it it takes years to even just get to that point and then you get to start taking shots on going in an actual way right right right right right so yeah i mean this it's it's just transformative right to the dynamics of startups ups in general you know yeah i guess my mind was like going like should they be doing something like almost like competing with yc in some way where we're like they're like subsidize people aim teach them how to do business and then use bolt i don't know they're just like random idea off my head it's a great point right like we've been kinda kicking around the ideas around it because the and the hack was was like really our first foray into this where it's like how do we go in like foster entrepreneurship on our platform right because if we think this is one of the most exciting unlocks that this thing has right and the hack was kind of the first way and we've been exploring what would it look like to have something that's like more ongoing like maybe it's like an accelerator built in or i don't know yeah know things of that nature right yeah i know you've talked about this before but maybe if we could briefly go into like how that happened like from like stack blitz to bolt because i feel like i heard of stack blitz i think i did on twitter or something at some point i'm not entirely sure i think you know and then definitely like i saw bolt in the early days like when you guys first launch and stuff yeah totally i'll do the quick zero to sixty but yeah back in twenty seventeen albert pie my founder and i he and i launched which is web based ide e and the idea was really we want to make it as easy to build like full stack web of applications as it was to use fig or canvas the idea i guess kinda started because back when we first came the valley in twenty eleven twenty twelve we'd had the good fortune of bumping and dylan field and eleven wallace when they were first starting fig and you know back then like they didn't have a design tool like the initial pitch was like this demo of a three d ball dropping into water inside of a browser tab and the pitch basically along the lines of like you know browsers have become very powerful you can have to do complex graphics rendering in them and so like if you can create a rendering engine that runs in a browser you can bring design to the browser in the same way that google brought docs and email etcetera and they'll be hugely valuable but obviously we've seen how that story has played out now and so back in twenty seventeen we saw the for the first time like oh it might be possible to do that same play but for development environments so for the first time you can actually mount that environments in a browser using the end users you know cpu memory etcetera and that'd would be huge for the same reason was for fig and google docs and whatever half you we started the company in twenty seventeen we launched that technology in twenty twenty one a lot of developers thought was very cool but over a couple of years after that like we had millions developers using the product for free every month we just should not get anyone to pay us money and that was like by and large like if you think about you know what's really the difference between stack splits and bold stock which was an ide where you still had to know how to code like you didn't have to set environment but you still had to know how to code which is huge yeah you were saying was like to me that would been great for like maybe competing with like codec academy kind of thing where we're like here actually teaching people to code or something like you have exactly exactly now and it's funny because we started established because we were competing with code academy on our previous company yeah i in a sense like we were teaching people how to code online needed an online environment to do that and that's really where stacked what's kinda popped out but you know there wasn't an obvious way for us to make money at least the things that seemed obvious we tried all the other competitors tried no one figured out how to make money on cloud to use any meaningful waves it was just a mirage and so that came to last year we've been around for seven years as a company oh know if we can't figure how to make a venture scale business like it just doesn't and we don't have a clear path to doing so like what are we doing here right and so the decision was made like if we don't have something by the end of the year you know that's like changes the trajectory we're gonna start spinning down the company or getting acquired like whatever right and so what we ended up doing is we launched a couple of different products last year all of them were very cool but you know the the first couple we launched you know didn't get the commercial traction that we were you know hoping to see and you know when we launched both it was the last experiment that we were gonna run we were launched in october our board meeting was in november where we were gonna you know begin dis assembling the company and and when it happened the surprised guys we get this this bolt yeah yeah totally right i mean it was yeah unexpected for literally everyone at that point we've been around for seven years and we got like seven hundred k of arr and then we launched bolt in the first thirty days we added four million of error and then in the month after that we went to twenty million was it was crazy yeah so it was like and there's a lot of these stories now these companies that are that can they're seeing these like revenue ramps like we did but we were really the first one that actually saw that type of hockey stick and we're like what the hell is going on that was how both came to be is it was an idea had earlier in the year and then once we saw the sonic three five model from ant ent i was like okay this is actually the coding models have really caught up and this this would be an interesting product right we had no idea it would be as well loved and receipts it has been and grow as fast as at then like we thought we would add like the best case scenario in our view was if we add a hundred k and arr r by the end of the year off this that would be amazing and we cleared that i think on like the first or second day or something you know well yeah i was that's kind of the story of of how bolt came to be but dude that you congrats on that i mean the fact that you did it for seven years and it sound like you were just like hustling and like you know barely holding on and you know and then had this huge success that's amazing it actually relied on the technology we had made the in browser dev environment and stuff like when you go to bowl all that technology we made over the past seven years like that is what is actually making and that possible we built bolt in ninety days because we had spent seven years building this technology to run dev environment so was it was a very natural product progression actually it was really a effectively like a ui change and a little like every brand because it it's instead of an ide you have a a chat box you know right yeah so it was it was very natural actually to cool i'm curious you know which model are you guys using right now like primarily like with boulder or are have you switched over to gp five or are you still using claude or yeah we primarily use claude i mean we found claude to have generally for when you when you're you know doing cog generation stuff like claude has the best models you we we've also we were you know one of the launch partners on gp five and and g five is a great model i think the certain areas where we're using it but and t has just done a really good job of really laser in on software cog gen and just being the best in the world at that and they're doing a lot of interesting stuff there where i i i just i suspect that they're gonna be the preferred tool for for some time maybe other things will come but you know it just it seems like that's really the niche that they're focusing on and focus matters a lot yeah yeah i went a vacation for a few days and i came back and like twitter was full of like all this all the people are arguing over gp five and some people were like you know it's it's amazing and other ones like it's horrible and this is like the end of ai was all a big bubble we're all done for and there was like what the hell is gonna when i come back to what i come back to what the you know i tried gb five it seemed really good like you know not like a gigantic step forward like a nice step forward i thought it was really smart and how they had like switch to like an automatic model because like for the average person the ui was horrible and then now everyone's basically like forced them back into having the same kind of ui again or like similar and and i'm curious what are your thoughts on ai progress it depends on like kind of the dimension of it coding and like building applications yeah i think on coding i think there's very meaningful improvements that that we've witnessed to cross model the model especially across the end of the ent family just over the past year like it seems like the the rl stuff that's going on there has legs and and has a lot of room to run look at the more ag api stuff i think that all these things are kind of reliant on new approaches being invented and proven out right so i think on the current if you kinda take the current state right the whole idea like scaling things up like that has topped out but you know the adding like the reasoning and and you know rl stuff has kind of moved the needle a bit in a way that you qui it forecast so i think we might see more stuff there but it's like you know it it kind of a a tb i guess it's kinda hard to know if that's also gonna like cap out too like the you know the the the thinking the reasoning side is that gonna continue to scale up or it's also gonna have diminishing returns it's kinda hard to so it might right i mean i think there's just gonna be a lot of new research and invention done but i think even with the current state of stuff we're looking at a lot of the games like moving away from like do we have the best model and kinda moving more towards do we have the best model that's designed to work well with agents right because if you think about like how human brain works it's kinda rare or there's like off the top of your head on any given sufficiently complicated question you will just zero shot an answer that is completely perfect right it it makes sense that maybe we can get machines to be at that point but but it seems like a a pragmatic way to approach this is like you know that's maybe the north star or whatever but like the way in which people think is by bouncing things around and going back forth and that's like what agents do like cloud code is going and looking at their files and like kinda digging it and sometimes we'll hit like a a dead path and like what it's trying to figure out and it'll revert back and like you know blah blah and so to me it's like the area where their innovation is ripe and there's like early days is really making that sort of critical thinking really really good and making good at tool calling and like building agents and they and that sort of thing and and and so i i think it actually kinda matters less about the specific model capabilities because even with like some of the frontier models like i just i i i don't think we're really maxing out what they can do if you actually have a really good agent that's orchestrating them but that also relies on the models themselves being trained and be need to work better as agents i think there's some number of years we're just focusing on that is going to yield some incredible results i think yeah it seems like that plus more compute and we'll get like a amazing progress at at least like three to five years for sure you could argue that ag is like twenty years away or like we've already reached it like depending on your definition i mean yeah yep and i think too it just it's gonna take time for the application link to actually fully utilize even the current state of the art right so i i think that that there's not actually to me an urgent need for these things to go like ten x better tomorrow there's real revenue coming into us and other companies and there's a lot of new innovation happening that's not even related to the models get getting better so from that end i think it's gonna be transformative on just how the world operates without really needing some giant step function in any short period of time i think the average person doesn't even realize how good the the current models are they're using is like a basic chatbot or or google for sure and anyway then it's just not a a proliferation of products that have are really leveraging these things to the max and and integrating the people's workflows like it's taking time and that's like the same thing we've seen with every other wave of technology where it's like the adoption of cloud services you know adoption of the internet it's like it kinda takes time for there to be proliferation of these of of the builders who are delivering the services and the consumers that are purchasing right right and there is it is faster than ever has been in previous cycles but even still it's again coming back to like for our products like most people don't know that they can write software no right and so like the tam is like ninety eight percent of the planet right there's not even a hundred million people that are using these there's not even like twenty million people before yeah they wonder how you reach those people like even if you just told them they like they they wouldn't be able to process looked that was actually we're going that you could just go type in and it would make a website for you or an application or whatever like people have a hard time you know grasping that yeah exactly so mean i i to again a lot of this is like it's it's really people builders and and doctors need to go and solve the last mile problems of these things at the application and distribution side of this problem side and so for those reasons i feel like everyone and maybe in large part because of some companies have been just like hammering about ag for the past five years or whatever but you know it's like i think there's is some expectation it's like ag or bust right right this isn't zero sum right like this is not zero sum like maybe all pass leads to that but like that's not needed like if you look at software or development specifically what is the value of going and like creating products that whose purpose is to like go and basically rewrite all of the world's software i mean that's like that's what we're talking about here like with bolts and other products you can go and just build new software that replaces previous human written software the market size of that is like trillions of dollars you could rewrite how the world works with just basically the current state of technology it gets better interest in happened the password right yeah me and matt did an episode early on in the podcast where were talking about the implications for the software industry and i was talking about like the ref ref of software or like rewriting of software you didn't with like the government like the some of the software they're using saying it's like that's been there for like thirty years because it's just like well we didn't know how it was written in the first place so we don't know how to change it we're scared to touch things there's so many complex things like that ai definitely like be help us do i think that's the kind of stuff the next five years all look become realistic totally so that's kinda of my take on like you know what's progress or not or like wait but i think it's just like pragmatic it's like we haven't even scratched the surface of the current level of capabilities as far as applying it into the world yeah and then it's still improving and that's crazy because again even with the current state of the art there's so much more to be done you know know and i'm convinced also like you if you ask people like ten years ago if if you show them what we have today and they be like you got ag like like i like i think a lot of people actually would be calling what we have today ag but anyways there's two questions that i'd like to ask people just like kind of for fun as we end the podcast and one is i have a eleven year old son i'm constantly thinking of you know what should i be teaching him and i know i i apparently said to become a pretty common question of people in silicon valley ask each other if you had a son what would you be teaching your son and would you still be teaching him to code would you be teaching me him like hey use things like bolt and just go create your ideas yeah good question i i've got a one year old daughter so she's not at the point where i can be like teaching her and use bolt yeah yeah me exactly but i think like in a couple of years i mean i will apps literally have her going and using bowl if she likes to build things at least right it's this is amazing i mean you know the reason that we we give kids like legos and like ton trucks instead of actually putting them behind a forklift is that it's just it's not safe to like give kids access to the real tools that professional use right that's not true anymore right like and there's also of course it's it's not safe and there's a huge learning curve right even if it is safe right and when you look at things like both in neither of those things are a problem right but let we have people of all ages coming and using both of all skill types of all backgrounds and to me it's like often people don't end up becoming engineers or looking into building things being entrepreneurs because they don't get exposed to what it actually is about because the drop off is so big and i think that's to me why it's important for my own child like to use bowl or things like it is to go and experience when it's like to build stuff for yourself right yeah so that would certainly be one of the main things but she's one so should she should she learn to code or not or would let be totally irrelevant in seventeen years so to me it's like it's all like i think learning how to code is is still a great thing to learn right because it's like yeah that your ability to really understand what's happening at a at a lower level with these tools that's always always been helpful in this profession right and i think that's gonna continue to be the case but if assuming you wanna like this is what you wanna work on right for for as a career right other things cheers is i think arts are cool i think going and learning how to sing and dance answer whatever is fun or sports and that's sort of thing those are the main things i think all of my own kid i would encourage them to go dig into nah i that's a good answer okay you get into a time machine here in san francisco you step out it's twenty fifty what do you see like is it dramatically different or is it like you know robots everywhere you know or is it just like pretty much the same is it i think i would expect it to be the same in some ways but i i i think we tend to get like the frontier technology way faster than the rest of the world like i i remember when i moved here from chicago back in early yeah like twenty ten twenty eleven uber was just a speaking would just becoming a thing that yep yep me too i was there yeah i saw travis at a party you like when it was first starting yep yep and so i i was using uber all the time in the bay and i went back chicago to visit my family in like twenty twelve and like it it was like there's there's no uber in the entire state right right and i and so it's stuff like that where i think i think the the the bay area is gonna have they're gonna have i i twenty fifty it's gonna be robo taxis or whatever in like robot delivering food i i think it's gonna be probably the epicenter of what that stuff looks like i would imagine mh and you know i'd would imagine this be probably a lot less cars on the road in that sense or at least you know non robotic ones but yeah that be being my base expectation of the thing because we already see that now it's like we're the first have way mo and so i suspect by that time it's probably gonna be the d facto way of getting around or whatever you know yeah what eric this has been awesome where should people find you online and you know like that you what sure yeah for sure so yeah like bolts is like bolt dot new that's the domain you can find me on twitter just at eric simon and on linkedin you can search eric simon i guess hope that works but it's like forward slash in or you crazy yeah yeah yeah i'd like this like you a ut tracking code you need in all that yeah so yes you've been awesome thank you awesome thanks for having me yep

welcome to the next wave podcast the show where we sit down with the people shaping the future of business and technology today's guest is kit bad the chief marketing officer over at hubspot one of the most influential marketing leaders in the world kip us on the front lines of how ai is transforming marketing from boosting productivity and reshaping roles to rethinking how brands actually stand out when everyone has the same tools we dive into hubspot brand new loop marketing playbook the bold idea that taste beats tactics and even how they drove a four hundred percent increase in email engagement using ai personalization so if you're a creator entrepreneur or just someone curious about the future of marketing and the ai era this episode is packed with insights you don't wanna miss so let's jump right in with kip wagner hey kip thanks so much for joining us on the show today really appreciate you of being here matt thanks so much for having me i'm honored as a lover of all things ai to be on the next wave of big deal yeah man we're gonna definitely nerd out on this one so what i wanna dive into to begin with is i wanna learn a little bit about what the day to day looks like of a cmo over at hubspot know i just kinda wanna start by exploring my own curiosity a little bit here it's a great question when you're like running a marketing org for a skilled company what you're really doing is figuring out what did the react of problems i have to go and solve immediately that are super urgent and how do i not let those distract myself from like the big priorities that we have and so that's at a high level but what that really means is i actually just like doing a lot of real marketing making content writing copywriting approving web pages talking about campaigns with a big given event this weekend inbound obviously so prepping and planning and making sure you know everybody attendance is having a great time you're trying to basically tell a story across a big team people it's awesome it's a lot of fun and i could use ai all the time which makes my life very happy yeah yeah and you've got the marketing against the grain podcast with you and kieran which i almost feel like that podcast probably keeps you like finger on the pulse you gotta know what's going on in the world of ai and tech because you're jumping on a podcast this week to chat about it yeah i what's interesting right is i grew up in a household where both my parents were teachers and so i learned the most important lesson which is like teaching something is how you learn something right right and so that's basically what doing a podcast is is like it's a forcing function where like i gotta record two days a week which means i gotta know what i'm gonna talk about two different times a week which means i gotta constantly be reading using learning and then i record and explain which makes it like actually really set in and like make sure that i've actually grasped the concepts that i'm trying to figure out yeah a hundred i i mean i've experienced the same thing turning around and making videos about it is why i know as much about ai as i do is i don't wanna get stuff wrong so i really wanna dive deep and understand it that's my honest question if you didn't do videos how much less do you think you would know i think it's like for me like it at least know half is less yeah yeah you know i would i would be way short of where i am like making stuff in public is like a cheat code for learning totally i'm in the same exact boat if i wasn't making videos about it you know i'd probably still be nerd out about ai i'm still kind of like a tech guy but i would probably be blind to like half of what's going on right i'd probably only be touching the tools and using the stuff that really really benefits me and then anything where i'm like i don't see a benefit for my own purposes i probably wouldn't be paying attention to but because i make content i'm sort of watching the entire landscape that's how i feel to yeah yeah i'm curious when it comes to the actual cmo role how has it evolved with the emergence of ai are there things that you used to do as a cmo that you don't have to anymore how has that role changed well i think there's a lot that's changed more that's changed that stayed the same i mean i think a massive amount has change on like the personal productivity side for example right like we have this amazing prompt that we put in ant philanthropic that goes and gets all of our ass asana data and bake me a custom dashboard right so like i can get this amazing super dialed in summary of how we're doing at kinda anytime live in a really great format so like that's one example that's be pretty great one of the biggest things that have changed my life is we have this cloud project all about our target customers so everything we know about them what they like what they don't like interviews transcripts all of that and i can just go ask them what they think about things and what's interesting matt is like i've gone and done that and i've also gone and run like big surveys the results are almost always very similar and it's way cheaper way faster so it's just like i feel like i have the pulse of feedback i'm like are we doing something right are we doing something that's good much better now in the ai era than like in the pre ai era and then i just personally feel just like ten x more productive right like i just feel like i can get so much more done in terms of research packaging i love complexity comment i don't talk to you since it came out so i don't know very big comment person but like i have all these shortcuts made a complexity comment to just basically shortcut like just how i consume package and process information and that's like that's been a game changer for me do you have any other like little productivity hacks that you've learned with ai that have really sped up other areas i think my big productivity hacks are like make a custom gp or your cloud project that's a prompt engineer so that you can always generate really good prompts and so you just get way better first or second shot output of anything you're doing with the frontier models like that is for sure the second thing is ai taking over just like repetitive tasks like the fact that comment can take over my browser and fill out forms and do all of that and anything i know i have to do repeated i could just make a shortcut for and give it instructions and the context and then just bang it out and it'll have it go off and do it a tab while i'm doing something else and then the biggest thing is that comment can do things across tabs so i can like do multiple documents or things at once and that has proven to be like a game changer as well and just like the workflow of how i get stuff done i love it yeah i mean one thing i've just gotten into and this is very very recent for here is i just started using whisper flow where you pressed down a key on your keyboard and you just talking to your microphone and it's almost like the old dragon naturally speaking kind of thing right where you can just sort of like dictate and it will type for you but what whisper flow does is it actually listens to you and then cleans up what you just said so if you repeat yourself it cuts out the repeats if you say or or have pauses or any of that kind of stuff it cuts it all out it sort of listens understands the intention and then writes up the text in a sort of cleaned up format and so i've actually started using that when i'm like writing content sometimes i'll use it inside of chat gp and just sort of brain dump my thoughts and then have it respond to me i've gotten into like the whole vibe coding thing right when i'm using cloud code i'm and there just speaking what i want the app to do and it's interpreting it so that to me has been a a pretty big sort of change for me just in the last like week or so that i found really really beneficial i've been using willow voice which is a similar product i'm an investor in willow so i'm biased when to use that versus whisper but like the use case they and and the other tools out there solve to your point is freaking magic yeah first of all it's so much faster you save a ton of time and you can do it in whatever window you are in because they have access to like your system controls on your mac at least yeah i'm a mac user so chat slack gmail all of those things it has been a massive productivity changer for me i think voice ai in terms of like dictation is probably one of the most slept on categories of tools overall in terms of just like instant productivity game yeah totally agree now you've mentioned claude a few times yeah do you have like a preference between claude or chad gp do you use one over the other for different things what's your approach to using the different models so it's a good question i use complexity max a bunch when i'm using comment and then i'll i'll pick between the different models and everything in there depending on what i'm doing but when you talk about the core frontier models gemini claude and open a google i'm using v and now nano banana like their image and video stuff is is really good claude is the most like enterprise of all of the models so if you need to collaborate with other people like i think cloud projects are just so superior for collaboration i still think claude is just really great at like spinning up an artifact and visualizing something or or coding something really quickly so i use claude basically for those use cases and then chat at gp i tend to think it's better for search i like it's deep research a little bit better yeah and for some reason i just really like the canvas feature in chat gp like i'm writing sometimes i will even take a draft from clog put it in chat gp and then like edit with the ai within the canvas feature there's something about that kind of word processor but like highlight a part of it and right edit and get feedback or research with ai that i find very useful yeah yeah i i would agree with that i would say with claude i really think there's no better model for coding still even though g is out and the benchmark show it should be better at coding my own experience i'm still getting better results from claude when i'm coding so you know the the benchmarks to me mean less and less these days yeah and and i think artifacts even if you're not like a hardcore coat if you're just a lightweight code like me artifacts is just an elegant way to like bring that code to life right right real quick like you just heard the loop marketing playbook has just dropped and honestly it's a game changer growth isn't what it used to be ai has flipped the script and loop marketing is going to be your competitive advantage kip is giving you over a hundred prompts to start your own loop marketing journey so you can differentiate your brand voice evolve in real time and future proof your strategy get it right now scan the qr code or click the link in the description now let's get back to the show so i wanna ask you about the whole jobs question of ai right there's obviously a lot of fear about losing jobs and that's sort of some of the biggest pushback that the like general public has around ai right is that it's gonna take our jobs as a cmo of a you know a mega company like hubspot how are you approaching you know getting the team on board with ai so i think ai adoption there are some fear around it right it's like oh well am i gonna like automate myself out of a job right and look the best thing i have read probably in the last three years is the article by packing mccormick called super intelligence abundance and the whole thesis of that article which it seems like by your smile might have read is that like as we can do more humans just consume more like we have an endless ability to consume i'll give you like a very practical example ai video is getting really good yeah right you've used it it's getting really good however i can go and make the same brand creative ad that i might run on youtube for example for a few hundred dollars versus hundreds of thousands of dollars because i don't have to hire actors a director do a photo shoot and everything right does that mean that i don't need those people anymore know what it actually means is i actually now wanna go make three hundred three thousand different really hyper customized targeted brand ads for different markets for different segments of my business because people are gonna respond better to them i'll still spend the same amount of money if not maybe even more money because i'm getting better results but i'm gonna spend in a different way so i i think the challenge with this is there are gonna be some roles where you just need less humans in those roles but overall you're still gonna need just as many people as my apologies and all this because you're gonna want wanna do more and you're gonna get better results and those are and that's what we're seeing is like we're getting a lot better results and when we get better results you don't go into cost cutting mode we're don't say like hey i i wanna spend less doing this thing you're like no i wanna get as much of that value where that revenue as i possibly can right yeah i think that's what's happening i think some people are just gonna have to do some skill transformation and if you're just dead set in doing one very specific job i think the next few years will be hard like for example like translation we know has a bunch of product market fit with ai so like if you're somebody who's you used to be a translation person you probably need to move into more of like a native contact creation with ai role versus just manual translation for example right yeah that makes a lot of sense i almost see it as like every individual inside of an organization becomes much more scalable than they used to be right like every individual is going to be so much more capable of outputting and doing a lot more with the same amount of effort they're currently putting in right you know nobody's asking you to do a ton more work but you're able to get a lot more done with the sort of existing amount of effort that you're putting in exactly that and that's where we're at and i find the best individuals the best teams right now have the right balance of like i've really embraced ai but i also know that ai needs a lot of context support help and guidance for me as a a human and when those things come together we can create a new better business result than either one of us could individually right yeah yeah absolutely let's talk about marketing how have you seen like marketing costs shift with ai now in the picture obviously we're seeing like a a huge overhaul in the way people do marketing seo is becoming less of an an effective marketing strategy like where is like the marketing spent going where is the marketing spend sort of disappearing yep if you wanna have a nice cry go to like the marketing subreddit the seo subreddit there's like a bunch of you just go take a pull through the online communities in the marketing world it's kinda bleak times and there's good reason for that matt it's because ai has disrupted a lot of how traditional marketing worked because now answers are free everywhere thanks to and so companies used to get a lot of traffic to their websites by providing answers on their website google would index them people would come so you're having people search traffic kinda go down and so you're seeing a big drop and click the rates from google and traffic from google so that's a challenge that a lot of people are facing because people are living more in these ll they're also just like visiting your site less like if i can ask chat gp all my questions i don't have to go to your website and look through ten different pages of your website so you're getting less visits there so that's that's another challenge because people are getting less visits they're spending more money on ads brexit they wanna make up the difference like go i lost these vents over here i wanna get more visits well all these online ad models are auction models so like the more competition the more expensive they're gonna be so ad prices are going up right and then so then people are sending more email and so emails becoming flooded there's this like basically chain of effect that has happened because ai has changed how people work learn and interact online and what i'm here to tell everybody is i've been through the trough of despair on this over the last i would say a year and a half to two years and we've lost tons of visits tons of traffic all of i've seen all of it the beauty is with this change comes a lot of opportunity and we we can talk a little bit about what that opportunity looks like but that i actually think that there has never been a more exciting time to do marketing to be a business owner who cares about marketing who wants to grow their business through marketing because there's a bunch of new stuff that is on the horizon that creates a bunch of new opportunity right right there's almost like a new term that's popped up and i've heard it with different acronyms right i've heard like an ai oh and ai e o and ae basically optimizing for the artificial intelligence and the large language models instead of the search engines so that's gonna be a a really interesting field that i think is gonna pop up as people just in in my opinion seo in the past for lack of a better explanation was kind of like figuring out how to gain the search engines right you're trying to figure out how do i do the thing that the search engines wanna see to get my rankings to the top and i feel like we're starting to see the same sort of thing in ai now like what sort of things do i need to do to get the ai to reference me when those questions are asked whether that leads to more traffic or not that'll still be kind of to be determined like a few things i'm in the ae camp for what it's worth ai engine optimization a yep but and the metrics there look different because you're looking at like how many times are you mentioned for the kind of questions your buyers would look at and it's not just about visits it's definitely very different but right the foundations of traditional search optimization still matter like if you don't have public information on your website or on youtube or wherever the l can access it they don't know that you would be relevant for that person on that topic right so you still have to tell great public stories have a great website and matt one of the things i would tell you is that it's a lot more instant than it used to be like i remember like ten years ago i be like oh do this inbound marketing thing create a blog rights some articles and like in like six months google finally index you and you'll get ranking and you'll get a bunch of traffic and now it's like if you and i were to make a video on this like very trending ai topic today and publish it we would be cited and be getting views of that video almost instantaneously if it was like a big important topic right yeah yeah yeah hundred percent i've noticed that myself i've put out youtube videos and then a week later been asking complexity a question and it's referencing the video that i made a week ago stay same yeah yeah it's crazy i i do feel like a lot of the sort of philosophies behind seo will probably still apply to ae right so i've been doing digital marketing for sixteen years now this is has been like my full time gig and the thing i used to tell people when it comes to seo is quality content like just making stuff people want to read and want to share is the trick to seo right you know you get people to land on the page if they stay on the page for a long time and read your article that's a signal to google that they liked it well a lot of these engines right when you're using per complexity when you're using chat ep and claude and all these tools that have access to search the web they're actually still tapping the apis of you know google or bing or whatever the search engines are so the thing that got you to rank with seo is still gonna be the thing that probably gets you to mentioned in the artificial intelligence models as well well when you have a period of transformation of anything right whether it's ai technology whether it's marketing whether it's it's anything mh you're sometimes tempted to throw the old out and i think what you and i are saying is what people have been doing in like traditional search engine optimization for example it's still relevant there's just some new stuff you have to do on top of it you're not throwing it out but you have to build it upon it and i think that's what the exciting opportunity is you built these really good skills over the last five to ten years and now you just need to build some new skills to take them to the next level yeah yeah so i i wanna actually get into the marketing let's get into like the meat of it i know you guys have the loop marketing playbook which i believe is dropping the day this episode is dropping so you should be able to access it now as this episode goes live but one of the things you mentioned in the playbook is that taste beats tactics i'm curious if you could sort of explain what that means okay so i've been spending the last eighteen months trying to get out of the trough of despair and get into this new happy land and and then figuring out can i help all the other marketers can come with me to this new happier land and a bunch of us put our heads together came up with a new framework to do marketing we call it loop marketing and part of that framework is the last ten years of marketing were very mechanical matt like it was like oh can i tweak my luke copy on my facebook ad can i change my link mapping to get more search visits it's very tactical no ai helps you automate so much of the tactics that like tactics don't really differentiate you much more right taste actually differentiates you and tastes always feels like to i think most people this like thing of like yeah that person has good taste or that person has bad taste but all taste really is to me and i think it's true is somebody or some group of people who have looked at and consumed such a high volume of information on its specific topic that they know what's good and bad and can articulate why it's good and bad right and that is what your customers need now so if you're a company out there it's enough to just like put out boring stuff you have to say hey i know what my customers actually care about i know what they think is cool i know what channels they work and consume information in and because of that i'm gonna make sure i live and play and tell my stories in a way that's going to really like work for them and that i'm gonna get away from the knob and dial turning and more into helping my customers know like i deeply care about them and know what they think is cool and interesting yeah yeah that's really interesting because i remember several years ago when i was doing a lot of you know building funnels and landing pages and all of that kind of stuff you had the russell b of the world going out there and saying hey if you change your button from yellow to greed you're gonna boost your conversions by like three percent and you know everybody was trying to do these little like micro optimizations but now in a world of ai we can split test that stuff at scale and pretty much anybody can do that kind of thing like right of the gate exactly and what wild though is like for you right like i could go run a deep research project and know more about you than i could have ever known before right so then subsequently i can know that about a group of customers which means i can intimately know what they want and i can give that to them instead of guessing that they might like a color button difference because somewhere in they're subconscious they're like more inclined to to think that's a better option right right right right exactly and i mean you know full disclosure i'll admit it whenever i do a podcast episode i'm interviewing somebody i always run their name through deep research through complexity through chat gp and pull up as much as i can about that person and that at everybody's fingertips now right like even in the podcast podcasting content creation world the ability to get just tons of information on any topic any person anything you want is available to everybody but the difference is really that taste right like yeah figuring out or alright which of these topics do we drill in on which of these things do we think our specific demographic our specific audience is going to be interested in i was talking about right that information is more of a a sort of taste is sort of a learn skill by going through the motions but anybody has access to all of the information that we have access to now yeah one of the big things that you have to understand in this next wave of marketing is that there's information symmetry that everybody has access to everything it used to be that companies have more information than the people buying this stuff everybody kinda of has access to the same thing and there's new opportunities that come with that access to information like you're saying that sort of leads perfectly into my next question it's how do brands actually stand out when anybody can create like really pretty decent content with ai and now almost anybody can create little like software apps i wouldn't say like somebody can go and make a clone of hubspot right but somebody can go and create like a little software app that they can turn around and market and sell it just seems like it's so democrat now how does somebody stand out when anybody can do that stuff so yeah so there's a few ways that i think you can stand out to answer your question very directly even just use the web app example if anybody can vibe code and app and sell it that's great but if i have a brand and if i have a real business can i vibe code an app and give it away for free because i don't need to monetize it because i can use it to provide another layer of value to enrich my core product as service offering right and so like that's what we're talking about we're talking about using ai to do the next order level of tactics and how i kinda think about this is can you use ai to create really good campaign assets that are like really relevant to your audience and like i said can you use things like a cloud project or what have you to basically simulate a customer and get real time feedback from that simulation so that you've actually know it's good and it's gonna work mh the other thing that we don't talk enough about with ai is the biggest challenge in marketing has always been the guessing game oh i'm gonna try to segment my audience guess what they want and send my best guess to these different groups of people right like that's been marketing now i don't have to guess i can say i know specific things about each of these individual people and based on what i know ai will create a custom message for that person that is relevant to them that's our best guess not just for a group of people but for that one person right and it turns out in doing that like matt we've been doing that at house hubspot our conversion rates have gone up a hundred percent two hundred three hundred four hundred percent because it turns out it's a way better experience to have ai guessing just about me than just to like let me into this broad swath of humanity right yeah yeah no i mean i've seen it firsthand right like hubspot is one of my sponsors on my youtube channel i think we've been working together for almost two years now on the youtube channel and i have custom playbook with hubspot that are like matt wolf custom prompt playbook yep and so when i'm doing sponsors on my channel a lot of times i'm shouting out a playbook with my name my face on it that we know is going to be super sad gonna learn to that audience yeah because it's like oh they have a connection with you and so we've already like gotten on that same level of commonality with them and they're gonna be just from the jump more interested yeah yeah and i love what you said so there's a a marketer you're probably familiar frank ke right he's yes he's probably one of the most well known internet marketers right he had a concept called moving the free line right used to teach his concept of moving the free line where you know at one point the free thing you gave out was like a blog post and then you would sell them off the blog post and then it became like ebooks then people were giving away ebooks and that's what you were selling and the value of the thing you're giving away for free you wanna constantly keep on moving that up so that the free thing gets more and more and more valuable so that people see well if this is what you're giving away for free well then the paid thing must be really really pretty incredible right so like i feel like ai really really increases people's ability to move the free line right create a little software app and give that a away for free before selling them on the bigger product they're trying to sell yeah like look even if you're selling consultant services you can give away a lightweight app that just shows how your unique perspective and insights on the problem right like there's a lot of leverage that comes with just demonstrating value in a very clear easy way yeah absolutely so with this playbook you guys are putting out there's what i think you call the dynamic loop and it's got four steps express yep taylor amplify evolve i'm wondering if maybe we can kinda go through and break down each of those four steps real quick yeah of course so everything will be on loop marketing dot com so before i talk fast or if you get confused go check out loop marketing dot com but yeah so we call this loop marketing and it's basically a four stage playbook that matt i was out to solve this problem if you do marketing right now at a business your ceo comes to you and says hey i need you to to like ai marketing and make our marketing ai first and get my better results go and do that and everybody's like i don't know how the hell do i do that like what does that was does that actually look like and so i wanna give people a very easy set of things that they can go to so you kinda went through the four stages the first step is express and that is literally just like hey i got this idea and now thanks to ai i can get that idea out in the world really quickly at a really good quality because i can use nano banana v three hubspot you can use a bunch of different things to basically go from idea to assets or campaigns really quickly and that's really powerful and you can then subsequently run more campaigns and and learn faster the next thing we have is tailoring which is just what we were talking about a couple of minutes ago which is like how do you use ai to create personal message so it's like the difference between you know go into a department store and buying a suit in the men's section versus like going into a tailor and the taylor makes your suit like man taylor makes your suit that's fancy that fits in a whole different way and that's kinda what getting these very personalized messages is about and that might be in in the form of ads email sms whatsapp but those types of tailored messages perform much better you need to send less of them to get the same level or better level of engagements and that's like something anybody can go and do tomorrow we just rolled out an really awesome email ai personalization product today at inbound when the show's coming out so i think that's a way to do it for email but you can can do it across all those channels then the next step here all about amplification which is like how do you take that message and get it out to people get it out to the world and the biggest thing that's changed there as we talked about earlier is a and all of the crazy things that are happening with how people are searching and interacting but there's also influencers like you said like we work with you at hubspot you work with other brands and everything and there's never been a better more cost effective way to reach your target audience than partnering with content creators and influencers it's an incredible opportunity and ai actually just makes that identification like the workflow of creating the content partnering working so much easier now that it's actually available to all these businesses and then the last step we have is evolved which is basically it used to be matt you've done marketing you run a marketing campaign and then you do your campaign retro perspective if you remember those were like maybe a week two weeks three weeks later you'd look back and be like alright how'd this thing do well now you don't have to wait anymore because these ai tools make reporting and insights so easy you can kinda see how it's going in real time make adjustments in real time and even once it's done with the hubspot and cloud connector i can just like get information into my cloud i can have artifacts like build custom charts and everything for me based on the questions i have about this specific campaign and i can make decisions instantly around how i make the next one better so that's why i think of it as a loop because you kinda close off with like oh i'm learning these things then i'll start all over again but this time i've got like more insights and those insights kinda compound over time now my marketing is just gonna get better and better every week and every month because of that compounding right you mentioned the sort of email thing that you guys are coming out with i saw on a another podcast that you were on recently as part of my complexity research or that you guys recently increased your email engagement by upwards of four hundred percent is that using what you're talking about like what does that look like what is yeah yeah so it's basically we product ties what we were doing matt and the initial like beta of this was we created prompts we enriched the people we're mailing with data about hey this is their job this is their business these are the things that we think they would be interested in then we pass that data and a prompt through to open ai through the api to generate an email you know the prompt is based like this is the thing we wanna communicate to them this is the data this is what we want this email to look like an open will write that email just on a one to one basis push it back in a hubspot and we'll send it and so now we've just done that so you don't have to do the api all the crazy work that we had to do when we were first like validating the idea right so does it in some information from like the hubspot crm so you guys crm data to know to make that email way more personal gotcha so so everybody was getting a slightly different email than everybody else right and even a little bit more than slightly different like if you look at some of the examples like upped at one for like a coffee shop owner and it had like coffee puns throughout you know like it was like it very contextual where you would been like oh no like somebody wrote this email just to me it wasn't like personalization tokens it felt like oh somebody really put some thought and wrote this just to me but you were able to do that because you had good data and then the ai was able to take it from there and make it really compelling yeah yeah so what advice would you have for you know brand new creators brand new entrepreneurs that are just now sort of building something from scratch like how would you approach it if you're giving advice in somebody just starting to build something now because i do think there's a lot of people that are like well what's the point right because you know ai can do a lot of this stuff if i build a business it's hard to build the moat right so what advice would you give to somebody starting today well i'm gonna answer that the first thing though sam alt has a statement that i very much agree with it the more you use ai the dumb you realize it is that like he said that different flavors that's a para phrase of over the years and i wonder if you agree with me that the more you use ai you understand its limitations definitely right and i'm here to tell everybody there massive limitations still and some of those limitations are actually even just around like the software the data connections things that are outside of the core models that still just take a lot of time and effort to do it's not that we're not gonna get there but it's not like we're gonna get there tomorrow it's gonna take years to get there first of i don't think anyone should have the attitude of like oh my gosh i should just give up ai is gonna do all of this and i don't have a chance and so if you let's say you agree that that hey alright i wanna start this business because ai is not gonna eat my lunch then it's like what do i actually do and the thing you actually do is can you know your prospective customer base better than anybody else this kinda goes back to that topic of taste can you just understand their problems they're pain what they need more deeply than anybody else because i don't care how good a model gets they're never going to know the deep pain of the five thousand people you're out trying to help or whatever your market is right right and if you do that and then you just ob obsessed about building a product to solve that pain as well as marketing so it's like content storytelling around that pain traditional search engines but that now our ae optimization maybe you build a little lightweight web app on you know cloud artifacts or with cloud code that you're using to help demonstrate that you understand what they need like that's what i would do and i think you will find that you get deep deep traction quickly if you're just like very focused on the market you're serving and understanding their pain yeah yeah there's two things i wanna say about what you just said there one going back to like finding the limitations it's funny because i just saw a tweet this morning from a friend of mine named riley brown who's building an app called the vibe code app i believe and he posted a a tweet today saying that he's hiring he's looking for video editors he's looking for storyteller and writers and all of these people to help him with his content marketing strategy for his new app and like half the replies are people going well you're an ai guy why do you hire anybody why you just hire ai to do this stuff for you and it's like i mean there's the limitation right like i wouldn't trust ai to go edit my videos for me yet i wouldn't trust ai to go and actually do a lot of the actual shopping in the motion graphics there's still a lot of limitations in that area yet but i almost feel like from the general public the thirty thousand foot view of people that just kind of see ai as a concept actually they believe like ai i could just do anything for you why are you trying to hire for this but that was interesting i think it's super interesting to build on that the one thing i would say to all of those people is that humans are social creatures but i'm fascinated by two studies one is the harvard longitudinal study that showed of the people who lived the longest and had the happiest lives it was because they had the best social connections mh and then the next the most recently there's been a different study i've was just listening to the plain english podcast by derek thompson about this where people who have superior memories who are less likely to get alzheimer's dementia all of those things there's fin findings that the part of their brain that is developed by intense frequency of social action which is far more developed than everybody else just had way more social interaction and so if you're going to build a great business and you're going to sell a product or a service and you think you can do it without that social connection without the humanity without the need for those people to tell the story be the face of those stories then i don't think you understand people that well right right that's totally true i mean i feel like when you're building a business it it would be really tough to just do it on your own right even with my old youtube video i have a producer who i bounce ideas off with i've got three different editors who helped me edit my videos you know we've got darren here in the background who's helping us make sure this podcast comes out really good right like i feel like i would probably fail if i wasn't working with other people i need that help i want the ideas to bounce i want that interact and i think that is so powerful to grow business and i don't really feel like i get the same feeling with ai no definitely not yet that only think it's close yeah the other point i wanted to make about we were just saying too is when it comes to like the advice of somebody just getting started the the advice that i find myself giving is that i don't really feel like there's a better time in history to rapidly sort of iterate and testing yeah right you could come up with an mvp of a any software you want run some facebook or google ads to it and find out pretty damn quickly if it's it's gonna work or not and if it doesn't turn around and try something else right this is the best time in history to make an mvp quickly test it and then when you find that thing the thing that okay this is clicking with people this is getting people opting in for my free tool or whatever okay now maybe you've found your niche let's go in a little bit deeper on that build it into something bigger turn it into a business but i feel like this is the best time in history to like test stuff at scale couldn't agree more i'd even make the argument that if you are a curious person who likes to learn and you're willing to like apply and take action on what you learn that it's almost impossible to not be financially successful now like if you just do something every day and learn it about it every day within like a year or two you will probably have more money than you need yeah absolutely absolutely there's one sort of last quick rabbit while i wanna go down with you before we wrap up here and it's just following my own curiosity again are there any ai tools or software platforms or or things that just have you really excited that you've just been really impressed by lately you know what i think what eric and the folks over jin spark do is very remarkable i don't heard gen spark i've heard that term phone around but i actually haven't tried it yeah so gen spark what they're doing that is pretty remarkable they're just building these very focused use cases i discovered them because they have a really good slide builder that like you can do a slide outline and it'll create great presentations with it they just rolled out this ai designer that's kind of like a true like ai first version of can that can like make t shirts soar posters or kinda whatever you wanna do but they have all these vertical ai apps i think there's like twelve of them now and they're just really good and like they've built a good user experience they're still really early but like that is the one to me that i go back and use that nobody really talks a lot about that i think are doing like like a real nice job gotcha yeah yeah i've been hearing about gen spark i feel like for a little while now and i i remember going to the website and playing with it like i don't know maybe a year ago or something but i'm i'm assuming it's probably evolved quite a bit since then so i think it's about time for reading try the designer and and some of the new the designer and the slide thing like if you just have gp five make you a slide outline and then have it go build it in jens as long as you have like a creative style guide to tell help the colors and the fonts and all that comes up it does a really nice job yeah yeah absolutely i'll off to go check that out and then so to sort of wrap this thing up let's wrap it up with the loop marketing playbook i mean maybe give us the quick elevator pitch of why somebody would want it and then how do go get it okay so you're gonna wanna go to loop marketing dot com to get this playbook i think it'll be link down in the description below here but you're gonna want it because you're frustrated with your marketing results like if you're anybody you're not happy with where your marketing results are going and what i'm trying to do is give everybody a playbook of how to do and think about marketing in this post era and it'll give you something easy to communicate with your coworkers your boss everything about this is what we should be doing here are the prompts here the apps here the tactics we didn't even scratch the surface of what is available out there it's a very robust playbook go check it out and that's what i recommend doing next yeah i can attest as well i mean i've seen the playbook it's like tactics and steps and there's things that you can do and there's sometimes sometimes there prompts right that you should just go and use that prompt to go take advantage of but it's trying to d oh what do i do with this ai stuff it's like go to these specific things and this is how you do these specific things and they're not they won't be all the stuff you do but man is it a great place to start and is a great way to shift how you're doing your marketing in this post ai world very cool so loop marketing again i've seen it it's in dev it's here super awesome well awesome kip thank you so much for hanging out and i mean nerd out about this stuff with me it's always a blast thank you for having me it's been a blast it's an honor thank you so much if you remember just one thing from this episode it's that in an ai world where ai automate the tactics what truly sets you apart is taste your ability to deeply understand your audience and discover what actually resonates so here's your action item take one piece of your current marketing maybe an email a landing page or even a social post and run it through an ai tool to tailor it for a specific audience segment see how personalization changes the response and don't forget you can grab hubspot brand new loop marketing playbook right now at the link in the description also if you enjoyed this conversation make sure to like and subscribe on youtube follow us on spotify and share this episode with someone who wants to help build the future of marketing we've got more on the next wave in just a minute but first i wanna tell you about hubspot for startups state of ai in start up go to market report there are some really eye opening findings here get this vcs are now making ai adoption a funding requirement nearly half of venture backed startups dedicate over twenty five percent of their go to market stack to ai tools and it's paying off big time seventy two percent see improved sell rates thirty seven percent have lower customer acquisition costs and some startups ups are scaling to sixty million dollars in arr with just thirty employees so whether it's customer service sales or marketing ai is driving measurable roi in hours not months so don't get left behind download hubspot free state of ai and start up go to market report now seriously stop what you're doing right now look at your podcast player and click the link in this episodes show notes see how ai powered startups are growing faster while staying lean

welcome back to the next wave podcast where we dive into the ideas shaping the future of technology today's guest is someone who's been at the forefront of ai for over fifteen years he c founded deep mind he built inflection ai and now he's leading microsoft's massive c pilot shift today we're talking to sole in this conversation you'll hear why the so called ai training wall is a myth how hall hallucinations might actually be a feature that we want and why the next software mode isn't head count or credentials we also explore the agent era of ai systems that don't just chat but plan click and get things done for you and wait until you hear the wild story about pilot vision predicting the future of a flight that must gustavo was supposed to be on by the end of this episode you'll walk away with a practical playbook you'll know what you can hand off to an ai agent this week as well as how to think about mo in a world where anyone can ship software in minutes so if you're living an ai or building right now this one's for you let's go chat with alpha sole so i'm gonna go ahead and start with your story a little bit i know you've been in ai now for over fifteen years with you know deep mind and inflection in now microsoft i'm curious what was it about microsoft that drew you to them you know it it is quite incredible to think that microsoft is fifty years old yeah like significantly older than me and you know for most of that time has been one of the leading tech companies in the world mh and also adapted to every new tech revolution so each new wave requires the company to pretty much change its culture sometimes change business models learn new programming languages learn new go to market motions and stuff like that so i think i was always inspired by the adaptability of it mh and that's kind of what we have to do now we have to adaptability to this c copilot agent ai companion era mh which is just a totally different mindset to what the company used to yeah absolutely have your beliefs or you know views on ai changed over the last fifteen years since the beginning of deep mind today yeah i think when i originally got involved in ai back in two thousand and ten mh we were very very obsessed with the ag framing artificial general intelligence right and i can still see that happening but i think i've always been a bit more focused on what happens day to day practically with real users you know real consumers of technology right and how that makes people feel you know rather than like the super intelligence that solves all the grand scientific problems i still think that's super cool mh but i can also see it really living life alongside us being an integral part of my day to day workflow as significant as your smartphone or your tablet or having access to the internet right right so i'm curious about the whole like narrative around ai training there was a a narrative not too long ago that the ai training was gonna like hit this wall and that we would sort of run out of things to train on and then we saw the test time compute and now they're sort of moving that in for the amount of compute going to that other end of it do you think there is a wall and do you think the answer to this stuff is just throwing more compute at at all yeah it's interesting because the other side of that equation was people also saying that there wasn't gonna be enough training data remember just before everyone's was like well pre training is done and you know we're gonna hit this wall and i think the the lesson is that whatever obstacle you know comes up over a six to twelve month period mh there are gonna be five different you know ways that people are trying to make the models more efficient use compute in a different way generate data synthetic which is what we're all doing now like we're producing human quality data yeah completely generated by one of these models for reinforcement learning from ai feedback instead of from human feedback so and if you look at like how big the models have got yes they've got super huge but they've also got tiny right you know you can now train a gb three capable model at one hundred times smaller terms of the inference cost you know the cost that it takes you to produce an answer at test time than it was when it was originally trained it was only three years ago so it's getting cheaper easier smaller and it's also getting bigger more powerful using more data and more compute mh and that's what you expect when you know everything is working it's like having like struck an oil well and now every everywhere you look it's just like okay we have another amazing capability that's right coming out so do you think that like throwing more compute at it and just kind of the bigger data centers the more gpus we put at it that's sort of how we get to the next level or is the next level gonna come from something entirely different yeah i think that trend is definitely gonna continue there's gonna be larger models more compute and we will definitely you know continue to get massive gains out of that mh but there's also like an emerging craft now which is in giving feedback to the models from other models right so you get larger and more capable models to really efficiently teach smaller models with very high quality data to perform a specific task really well right and i think that that trend is is also gonna continue and that's where you get is kind of efficient frontier like it's very expensive to build the biggest and best model first mh and then it's like ten x cheaper to build what was frontier twelve months ago today you know kind of amazing right so then you have this entire like swarm of you know second you know phase or second wave right model development capabilities which are coming very cool so like stacking reinforcement learning on top of what we're doing is kind of what's providing this sort of next era of ai yeah simultaneously with just training larger models right getting higher quality data feedback from other ai you know the kind of rl paradigm we'll still use you know feedback from humans as well so you know i think the lesson is that you know all of the capabilities end up cumulative being valuable mh and you sort of chip away at the problem by both exploiting what you know to be rewarding and exploring new techniques and new methods that then get piled onto the staff right right everyone everywhere is talking about ai agents right now but here's the thing most companies are going about it all wrong hubspot has gathered insights from top industry leaders were implementing ai agents the right way you'll discover which agent setups actually deliver roi and how businesses are automating their marketing sales and operations without replacing their teams get it right now by scanning the code or clicking the link in the description now let's get back to the show so i wanna talk about hallucinations for a second so hallucinations i guess can kinda be considered both a feature and a bug depending on how you're using ai right so you know if you need something factual you kinda wanna eliminate hallucinations but if you want something creative or you wanna sort of solve new problems that nobody solved before hallucinations are something you want how do we get to a place or do you see a pathway to getting to a place where we can trust what's coming out of these models one of the things i find really interesting in technology is that sometimes a new trend gets a brand name or a label and then it confuses everybody about what's happened you know in some ways it causes like the entire field to hall about what they're actually seeing it's a really weird word hall because if you think about the weaknesses of relational databases that have basically built pretty much all the value of software for forty or fifty years mh its big weakness is that it can't adapt it can't change it doesn't have any fuzz you only get out precisely what you put in whereas you know this now kinda solves a bit of that problem because it has this like fuzzy adaptive you know sort of abstraction that it can transfer knowledge from one domain to another right that it can inter you know you have two knowledge points and it can join those knowledge points by finding the gap in the middle which is what you see when it's generating a new image that's in the style of something for example so hallucinations is just an unfortunate word and i don't really see that as a long term problem i mean if you look at like three years ago the models were really difficult to steer right stylist factual they couldn't draw on you know citations or reference documents that are more reliable and now like two or three years later with more compute they're extremely controllable right they're very adhering to the behavior policies that we set for them right they're very responsive to the stylist training so they don't actually produce a huge amount of buyers they still get some things wrong but it's very very different and so i i do think that that is one of the most promising signs of the moment we're in is that they're getting easier to control and steer as they get bigger right which i think was counter to a lot of people because a lot of people two three years ago like oh these are always gonna be misinformation machines right they're gonna be super chaotic they're gonna get it all wrong etcetera etcetera and that's turning out not to be true yeah and i feel like more and more now they're sort of grounding themselves with citations right so if they're giving you factual information you could see the citation and actually go and look at the source and sort of double check things if you need to right exactly and that's the definition of trust right if if it knows when it doesn't know or you know when you go and like verify a claim that it makes you feel confident that's a a sensible source you'll start to trust it more and more and i think that's you know also very very encouraging because the more people that use it the better it'll get and then you know again more people will use it because it's better quality right right now i know you mentioned ag is not really the the area that you you tend to focus on but i'm curious do you have a definition for ag because i feel like that term the definition is still fuzzy and i feel like everybody who responds you might have a little different definition so i'm curious what you're definite i mean that's another challenge right like flu hallucinations or i mean words are difficult when you're sort of trying to predict the future like we don't actually have good definitions so you know people talk about super intelligence or general intelligence you know i i like the method that we or the the definition that we use the deep mind which is the ability to perform well across a wide range of environments mh so the emphasis was on general and you know high quality performance mh where the threshold is you know maybe it's human level performance in some tasks maybe it's all tasks maybe it's exceeding human level performance you know i think we'll have to just wait and see that the definition that i proposed in my book was artificial capable intelligence because what we all really care about is not some abstract idea of what intelligence is which is really hard to define right it's storm as hard as consciousness it's really unclear or soul or spirit like they're very un that hard to define but you can define capabilities mh you know like how much power does it draw how many tokens does it produce per second does it solve this specific challenge can it retrieve over documents in the web right you know can it take actions in your browser like can it understand your scene and identify objects when you use copilot vision to interact with the environment they're very measurable and specific capabilities so my mindset is always like this you know month this quarter like what are the specific things that it can do and you know that i think is a kind of practical way of handling that question very cool so i know you're familiar with yan le cocoon right he yeah close he mentioned that he doesn't believe that ll are going to get us to ag he thinks it's gonna be some other type of ai model do you see any other types of ai models after l do what do you think is next to beyond large language models i don't i think that we already have the kernel of very powerful systems because tool use is a meta capability right you know that that's kind of what has made us unique as a species we're able to like pick things up move them around and then use that to manipulate our environment to augment our intelligence right it makes us more smart and capable that we can make fire or that we can use a stone acts or that we can you know create a gpu right invent electricity so we now have this system that obviously is nowhere near human level smart but it can use tools mh and we have a lot of very powerful software tools around right right and a lot of those are not yet hooked up so i think they even if you sort of stopped model development for a couple years you know ego core foundation model development somehow magically but just theoretically then we've already got this kind of technological overhang where the models have this amazing ability to use tools talk to other ai source factual knowledge mh orchestrate things in sequence right and so yeah i don't see it slowing down anytime soon gotcha i wanna just touch on as really quickly was listening to an interview with neville ra con and he said that he believes as artificial super intelligence is a fantasy we'll never actually see it i'm just curious what your take is i don't know i'm i'm kinda one of these people whenever i hear never or fantasy to see i was just like okay well you know it's a bit like the young thing like maybe saying that these models will never scale to an ag gi i just think it's really hard to say for sure either way and i i think that our intuitions are changing every sort of year as we see these like massive wave of new advances coming out mh so you know i definitely think that there are certain capabilities you know with respect to as that we need to be a little bit more careful of right for example naturally if you just give one of these systems full autonomy with no human in the loop or on the loop or in any sense overseeing you know it's pretty powerful system right so i think we should think about these things in terms of like what are the specific capabilities we wanna accelerate and engineer into the models what are the ones where we're not sure you know and and potentially could have very very powerful consequences if we misha them and really think about it from that perspective right right i also think it's one of those areas where we might see those goal posts continually move and no really knows how to define it so it's hard to to go there honestly but what i wanna talk about next is sort of some of the concerns that pop up you know a lot of the things that i see like in the youtube comments and things like that when i make videos about ai and you know one of the things that i see probably most often is ai is gonna take all of our jobs right yeah ai is learning to code ai is doing art ai can be your graphic designer should people be worried about ai taking their job i do think the nature of work is gonna change fundamentally the thing is no question about that you know just as fifty years ago there was no pc and everything was done on paper and we still relied on snail mail you know i mean we're going through a big shift right it just gonna affect how we work what we do where we live you know that is a profound shift and i think your day to day job has changed just as your job as a you know youtuber and a media star now didn't really exist five years ago right right well maybe ten years ago didn't really exist i think we've gotta get ready for that transformation gotcha so i wanna ask really quickly about software companies i'd like maybe some advice from you on software companies because there is this whole narrative that software companies don't have a moat because everybody can just sort of prompt to software into existence or maybe one of the big incumbent tech companies might just build it it is a feature what advice would you have to somebody trying to build a company right now good question i mean it depends on the size of your company but for smaller companies i think this is like an electric time it's pretty unbelievable i mean i i was just playing with this new version of github copilot now that that has this unbelievable prompting ability weight in natural language you just give an instruction or you just say what you want and what kind of app or what kind of game what kind of program you want and then obviously it produces the code you can then read all the code and then ask for specific chunks of it to change so in the process you're actually learning about how code is written right where the bugs are you know it's very very kind of insight four and that iterative back and forth is a great way to prototype and actually build production grade code and like ship the website for example mh so the the barrier to entry has never been lower which means experimentation by everybody is about to completely explode what that means is that it's gonna be a very competitive time for the big companies medium companies small companies everybody which i think is an amazing thing for the consumer mh because we're gonna get a lot of magical products and experiences but i think it is gonna be a very competitive and explosive you know hard time to pre succeed in because it's gonna be difficult to create those long term moat because things are gonna change so so quickly but i think what we've seen over and over with these things is that people figure out ways of creating tremendous value and that value always gets paid out you know and people make you know big returns and have have have great outcomes so i think it's a really positive story yeah i totally agree so last question i'm gonna make it sort of a two part what excites you most about what you can do with ai today and what excites you most about what we'll be able to do with ai in the coming years you know one of the things i really enjoy doing is just having a conversation with copilot on voice on a commute on the way home like at the end of the day just chatting through a problem asking a question rabbit hauling around something i love basically learning i love interactive dialogue because that's kind of the way i learn i mean love reading as well but in some ways my books and my ideas just come alive because i'm able to kind of you know run down any corner of a topic with copilot so i do love that experience i've also been using the c copilot vision experience a lot i mean it sees everything that you see in real time you can talk to it you know about exactly what you're both seeing and that is quite a magical experience i was in lax the other day and i opened c copilot vision and it knew which i was in it looked at the gate number it was like oh it looks like you're heading to seattle because it's four thirty on a sunday so there's only one flight right now by the way did you know that your flight is delayed oh and i was like it's not delayed code copilot you've obviously got it wrong illumination so now i went to the front desk and i was like is the flight delayed and they were like oh yeah just about to make an announcement about it so whitney like how did you know yeah yeah it's like that's a magical experience is is crazy it's incredible so but the things i'm really excited about coming up other actions c copilot actions is mind blowing i mean it will really just operate on your desktop in windows highlighting areas where your stock pointing at things helping you with settings or in your browser like we'll buy things book things plan ahead it's quite magical to see copilot moving the mouse around clicking on things opening new tabs entering stuff forms you're just like we're we're living in the future yeah yeah yeah amazing sounds like problem solving in windows is gonna get automated pretty quickly here exactly yeah amazing well thank you so much for spending the time with me this has been a fascinating interview i really really appreciate you spending the time pleasure man thanks a lot pretty fun cheers awesome if you remember one thing from today it's that agents that do beat apps that wait that's the shift so here's your challenge this week pick one repetitive task it could be booking a meeting filling out a form summarizing notes or cross checking data and hand it off to an ai agent then let us know in the comments how it went all the links and resources are in the description and if this episode helped you make sure to subscribe and share with a builder who needs this playbook thanks so much for tuning in and hopefully we'll see you in the next twenty

what if building an ai agent were as simple as texting a friend this week on the next wave i sit down with flo cr founder and ceo of lin the ai agent company everyone in silicon valley is watching after a year in the lab lin is unveiling will lin three point o and flow is ready to show us why it could change how we work we demo agents that run your crm qa your product book grocery block spam and scout competitors all through natural language you'll want an agent take over a browser click and scroll like a human and fish task in minutes imagine if every employee at your company had an army of ai interns that's actually possible today before we get started now we spend a lot of time working on these episodes and getting on the best guess possible to help make sure that you're prepared for the age of ai but most of you who listen aren't subscribed so if you can please take a moment subscribe you would me a lot to me and the team now let's dive in with flow cutting yourselves like will half sounds pretty impossible but that's exactly what sandler training did with hubspot they used breeze hubspot ai tools to tailor every customer interaction without losing their personal touch and the results were pretty incredible click through rates jumped twenty five percent and qualified leads quadrupled and people spent three times longer on their landing page go to hubspot dot com to see how breeze can help your business grow hey flo thanks for coming all hey nathan thanks for javier me i was thinking about it earlier i was like i you're like the third returning guests that's ever been back on the next wave you other ones are like greg is and logan hill patrick at google so you're in great company okay i am in a really good company yeah it's an else to be here yes so yesterday you showed me two demos of of new things that you're building at lin which i would just blown away by and i'm like okay this is perfect for our audience and everyone's to get a ton of value out of both of these if you're calling them features or the new products but maybe you can just kinda explain and then we can just jump right in yeah it's a revamp of the app freely if we've been working on these for like six or nine months at this point like the entire team we call it like ling three point o so it's basically an entire rewrite of the app the long term vision of what we want to do is build an ai employee and we are driving towards maximal capability and maximal ease of use ease of use is like we won't need to be as easy as like you just talk to it and you tell it what you want and just does it for you right and the capability is like it should be able to do anything that anyone can do on the computer yeah today like in order to build agents like you either have to code or like even like no good tools like lin it's all of a pain in the ass it's a no good tool but you have like drag and drop but like you need to have like a technical mindset and the capabilities is like we've always been bottleneck by integrations like we made like a big announcement like six months ago we released like six thousand integrations at once and we were like that's it like we will never need integrations again and in turns that we are always blocked by integrations and that's why ant topic is doing mc right it's like a like we're gonna build just like uni universal protocol but it's never gonna be enough and so what we're releasing with three point o is we are releasing an agent builder so you can just talk to it it's just a prompt and you have a back and forth conversation and just builds the agent for you it's actually really good because the agent builder is built with in the call so it's like we don't even have an admin tool to build the agent builder we just use india ourselves and that's the yeah and is then we have computer use so the agenda builder takes scale of the usability ease of use part of the house and computer use takes scale of the capability side of the house so it's like the final integration basically like you can just see your computer and use it right and it works it's checking it works like better then result that's crazy so you just talk to lin like if i want agent for whatever check my email or check website or whatever i want to do you just talk to it and it creates the first version or something then you probably can go in there and tweak it i would assume and things like that yeah no one hundred percent that's what it is it's like i have created an agent to manage my bill sold crm and literally just like help me manage my postal crm it's like okay i'm gonna create an agent this is what it does if you wanna add to your crm if you wanna retrieve from your crm and let me just create the agents for you and that's what it created like feel shot wow right yeah and this is why i wanna show video i wanna be real with people like yeah people who have like experience with like lovable or these like vibe coating applications they know that like this actually takes like one or two minutes so it like it makes for like a pretty boring demo because like water's is gonna it for like a minute looking at it right but it creates this agent and then i can be like okay who are ai for i know and like i like this like fake crm like a bunch of like fake names but it's like hey these are the people you know oh i can go like hey add a john to my crm he's the founder of ag cool and it as a role to my google sheets so right now it's mainly with google sheets so like in theory could you guys add you know connections to hubspot or or other serums in the future or how does that or we do have connections to hubspot it's actually like the most used crm on the platform oh great hair no one hundred percent so you you can just use any connection you want that's why like the six thousand connectors that we built also solving to be useful for and when you guys launching this on monday hopefully i think by the time these ai you should be live yeah awesome have you guys had people testing like other people yeah outside of lin and and like we're the main use cases that people are using it for oh yeah we've been in private beta for like weeks at this point like maybe even months is the main use cases oh i think like crm management are really really big one like go to market applications or like a really big one so it's like as the workflows account executive workflows like marketing and like content production workflow like i can give you an example this is the new design of the app i love with it i think it looks gorgeous but i can give you a super simple example like find me ten software engineers working at z and right here it's like i found tens software engineers i just like a previous task but like i did the same like boom these all the names and i can just click and these are like actual people who work at zapier oh that's amazing so like could i go the next step to help you contact those people or gonna not do that it will be able to do that mh right now it cannot connect to your personal small accounts just yet like this is like an an authenticated action like the prospecting action but like the gmail action is going to be authenticated and we're going to release that in a couple of weeks awesome yeah so the computer use could it just like go user your computer eventually and use linkedin and like and how was linkedin if we're gonna deal stuff like that it seems like some point that's gonna be be crazy is that is there's a really big use case so i can go like reach out to i can go flo cri yeah who works at lin on linkedin and it is going to reach out actually think for this one we also have an action and not handed wasn't sure oh it's going to use computer use let's see right here so it has found the linkedin and now it is starting a computer and so again like this is an example of of computer use like while it's booting i can like show this other demo it's getting very meta up because now we're using lin as a qa engineer full lin right so we have created this agent that like you can see it's super simple it's like okay every i will you start a computer you log in into lin we've given the credentials you create an agent you test the agent and then if it fails you send a message to the call channel on slack and right here like this is all completely automated like this is the agent that logs in and does the credential creates an agent and end to end like creates an agent ham scratch it's like a pretty simple agent at first it's like yeah an agent that like sends and receives messages it tested like hello can you hear me so this is an agent tuck an agent and it works wow that's amazing i would be remiss if i didn't mention i feel like at this point it's a tradition every time anyone talks about computer use agents like they talk about grocery and like flight booking you know is of this the i do have an agent that does that for me yeah so you can see you know every week it wakes up it puts maybe this is t like this is my grocery list on door ash i use pretty tame stuff the cucumber brussels pads all of that stuff and then you see it just goes on door ash and it searching for everything so it's like looking for brussels sprouts adding them to the cart like same for steaks salmon and all of that stuff and then in the end it just all and i received the groceries like this is how you do my groceries and can you set look a timer where or it does it automatically and you don't have to do anything or do need to kinda like approve it every time or like what's the steps there is that is how it just set up right now i'd like to be switched to the flow details so yeah every friday at five pm second can end a week thing i've actually changed it so now it sends me an email to me and my girlfriend and so like hey this is the grocery list do you guys wanna add or remove anything and then it weights like ten hour basically in the morning after if we don't reply just assumes that like the grocery list is good and then it does the grocery for us and when it's done it sends another email to say did it that is crazy i mean i'm gonna set this something like this up yeah another one this one is like so simple and it was also created with agent builder i'd actually show you twitter computer use this was like one of the very first things i tried with agent builder people asking me like how did you create this decision agent i'm like with agent builder it's literally just you talk to it and literally it's like i have a bunch of spam mills and dimensions on twitter i will you to create an agent that wakes up every three hour wheels and block them that's it and you can see here wakes up i changed it to every twelve hour wheels because i'm not spam that much like twelve by wheels is good and then goes to my mentions finds like this guy i met a blogger who really is awesome and that is like really his stock information day he is very good and deserves attention rainbow emoji is like boom blocked and so i i know don't longer have spam on twitter it's just like i don't that's great like blog it's amazing i i have this account on twitter i have the name learn on twitter and i'm not really using right now but occasionally i check it and like it gets tagged by i don't know hundreds of people a day who were like just talking about random stuff and then tag at learn for some reason right i would love to use this and just block all of them i mean now it's a tool that you can just prompt to just ask you to do it and do it for you how does it turn who was a scammer like would it possibly like block people they shouldn't who were like just like leaving a comment underneath then he was just kinda like a generic comment like lf g or something like that or you know so it it happened once yeah i prompted it to like the ai grief refills you know like old wall it's completely over open i release the thing like nothing will be the same again yeah and it started blocking them and i'm like hey don't block mentions of ai grief sales hey i i put out one of those like every month or two so i mean hopefully then hey you're not gotta feed the algorithm occasionally you know yeah they looks spam me eb homeless so look i've just i've just prompted it and like hey like those are examples of of spam oils yeah cool so you're thinking you know i've had elon musk like responded to some of my tweets before like a crying emoji or so something like maybe the ai would be like yeah that's definitely a spam fucking block i no you don't wanna you don't want to plug that i mean like i i've been pleasantly surprised because i received a lot of like cryo emojis like i live g and all of that stuff and i was like you it great think it's a spam and i've been like keeping an eye on it and no i mean like so far so good it's awesome the new version of lin you know lin three point out like what's the main way you're using it besides shopping like are you guys using it internally already a lot or like you know we all using it quite a we're using it a lot while the qa stuff is awesome like it's basically replaced the qa engineer for us like this is a kind of thing we would have paid a hundred and eighty thousand dollars a year before you know and now it's just an agent prospecting is another awesome one we have built an agent that goes on the linkedin and finds the post sells another push that are like i have built this awesome agent like comment agent to get it in your dms and all of that stuff and so we we could it like comment forming like when people do that we can see who's commenting agent right and so we can reach out to them by dms as well and we can be like hey you know like reach out to the person and we actually reach out to them in in a personalized fashion so like indie goes on linkedin and clicks and like looks at the profile and then sends them a message just like hey like we saw you go agent you wanna learn more the different agents that we could deploy for you are you doing tactics like that to you know like see who's following your competitors and and and things like that and interacting with them somehow well we do have an agent for sales snoop absolutely and and that's the kind of thing i think that's like the beauty of ai agents i think people under underestimate the you extent to reach like oh my gonna be able to save money because i can replace stuff that humans used to do but actually what you can do is that you can ask these agents to go much deeper than humans did so for example the competitor and that is this agent right it is kind of shocking how open salespeople people are sometimes about n on linkedin because i guess the numbers are like the resumes right but like they can say bragging them yeah i have closed like four million dollars of revenue in one year i have closed like ac three hundred thousand dollars average contract value and whatnot and i'm like this interesting thing and i couldn't ask a human on my team i would never ask a human on my team to like every month go and open the linkedin team profiles of every single of every single one with michael details and see what they posted and all of that stuff i can ask yeah the person he would ask to do that it would probably be like a lower level employee who would do something like that and they maybe they'd make more mistakes and and whatnot if you do to have someone do that yeah that's right so with the ai agents feel no shame i can ask you to do this thing and it's like an analysis that would have taken a human i goes so deep like this this would that have taken a human like two or three full days like i'm never gonna spend two or three full days doing competitor analysis as a startup right i can do it with the agents okay let's be honest your ai prompts aren't giving you the results you deserve but with a little coaching you can transform from basic prompts to engineering conversations that get you exactly what you want from chad gb that's what this playbook delivers not just random prompts but a step by step system with the exact techniques top ai professionals use every day you'll have your own personalized prompt engineering system that delivers consistent results get it right now scan the qr code or click the link in the description below now let's get back to the show so content production we are in the process of rolling out an agent this one is really good it's a so agents can work together so there is one agent that's like a content strategist and there is one agent that's a content producer so the content strategist comes up with a bunch of ideas of videos okay and then it forms them out to a bunch of content producers the content producer produces the content on v three and then post it on instagram and tiktok and youtube and then after it posted it a week later it wakes back up it looks at the videos it posted it and it's like okay how did they do both in terms of views and in terms of comments like what did people like and it reports the findings back to the content strategist which then uses these learnings to plan the next batch of content that's amazing i'm sorry but like we've been working on this for like two and a half years like to me like the fact that i can describe this workflow with this trade phase just blows my mind like it's just it's insane and it's crazy like all this kind of stuff to like i know how amazing it is but have to like try to like you almost like have like a beginner mindset to like remember like how amazing all this is you know you told me like three years ago it's like what the hell are you talking about there's no you know it's crazy especially like five years ago to i would i have no idea that something like this was gonna be possible but now i like every time i a new guest on they they tell me something that's insane and i'm just like i kinda get used to it and a lot of people you talk to they seem to get completely used it like oh yeah course they can do that i agree we we feel this quite a bit in the office like we were at the office last night at like ten pm and we were like white pudding like the plan basically this like okay we're gonna do this we're gonna ship that then like you're gonna work on that lend four and then twelve months from now we're gonna have like basically something that approaches ag api right like approaches an ai employee you can just talk to it and it does it and i'm like guys it's like how insane is that like we just done it's the plan like it's not like a sci f five book it's like no it's the road roadmap right right i mean since the last time we talked this already seems like half the way there compared what you showed me last time you've already made the fact that you can literally just talk to the ai and it helps you kinda create the agent mari you have to manually set everything up or use a template whatever that's that's a huge step forward yes yes i mean and this is us right like this is the application then there is the model layer like imagine what's happening right now inside the labs like i find it kinda strange that the labs basically right now are doing apps like they all saying we've got to like super is coming guys like we we're in the take off right now we've basically got ag gi and now all like on our way to as it's weird yeah and mostly of were just like yeah yeah no biggie you know yeah and which is why like i'm mean these group chats and there was this conversation is the other day i'm like in all of these ai group chats and there was this one with like a bunch of fun thumbnails and they will be mu the frenzy around ai these days they're like i'm so tired out of it and i totally get it is like i'm so tired out of it there's so much like grand just like so much tourism so many people just like promising the moon and all of that stuff and i'm like look those one gonna be sales but i believe that ai is actually under hyped and i'm saying that knowing food well how hyped is and like guys like brace yourselves because if you're tired the it like this is only the beginning like it's gonna get wow that there i agree i i actually have a presentation now that starts with ai is under hack it massively that yeah me it's kind of like you you're like an early two thousands you know and there was like the dot com bubble and you're like i'm so tired of the of the internet the internet like everyone's like who's gonna use the internet can't wait for it to be over all the house still sucking about time it can't wait for you to blow i'm like yeah i knew it was a scam look they crashed crashed it's the internet's done it was all bs exactly exactly the entire wheelchair just didn't yeah and people can't realize it you know ai has in my opinion has the potential to be way more impactful than the internet i mean you're talking an intelligent about powering robots you're talking it's so much bigger than the internet i think the average person right now a lot of them would be comparing this to like crypto or something like they think it's just like okay a you know it's another new thing and it's like they're not getting the paradigm check chip that's happening right now for all of humanity i think some people compare it to the internet some people compare it to like deal to automobile like electricity like some people compare to like the wheel or fire like again i'm i'm gonna be one of this like grand development but okay go for i compare to life i compare to like the bills of life on earth yeah yeah i think it's a new life form basically and i think it's going to be that transformative i agree i i i say like electricity because i feel like people can slightly comprehend what i'm saying but like and they think it's like a exaggeration but i i agree you know we're we're creating a a form of life here yeah it feels weird sometimes out somebody post about this other day where they're talking about i like hell if was like way less than one percent of people get what's happening right now it's just an odd time to be alive and understand what's happening i kinda find it funny now i guess just like all this is happening and most people just think with is like a cool chatbot or something do you ever feel like ver like do you do you yourself like feel overwhelmed and you're like this is wild like oh yeah totally especially because right now with ai i'm like you know juggling four different little businesses they're all cash flowing and getting involved more things in ais it's enabling that we're in the past no way i would have been able to have done that and now you're showing me this i'm like oh maybe that could be doing more yeah yeah and with my personality is like why not do more yeah course do more yeah i feel like that's that's a common journey that people are going through and like i feels they are ignorant and then they they find out about it and like those skeptical then they're really excited then they get so excited that if anything it gives them vertical that like this is insane like this is this is way bigger than than anyone is realizing yeah and then i feel like the final stage is almost one of denial it's like once you get past the point of the vertical like some people like hangouts there for like a while i i saw was there for like a very long time and i feel like at some point you almost it turns almost into a shrug it's like yeah i guess it's coming and like i don't know else do or say about it you know i guess i'm just gonna keep living my life in a normal way until it happens right yeah well mean you know had this conversation last about the end kind of slightly on a dark note you know i am more on the optimistic side of where this all goes i think you're kinda a little bit more worried about where the technology goes and not to today i'm not worried at all but i'm i'm definitely more optimistic so for me i'm definitely on the side of let's accelerate this as fast as possible and for me the main thing is you know making sure that the us and it's allies win at this right but that that's like my overall goal in life right now was like i've been trying like what can i possibly do in my life it's actually important any small little thing i can do to make sure that countries that care about freedoms and civil rights actually win at this and that's what spreads around the world that's what's important to do and look like for the record like i am ninety percent optimistic i think there's a ninety percent chance that everything goes well i'm just saying like even if it's ten percent chance of literal extinction that's very high that's much to high for con but yeah look even if things go well i think like to some extent there is a degree of like instinct mammal reaction where it's like change is scary i think it's just like we currently leave in the middle ages and we're going to be catapult into the three thousand yeah it's gonna be weird and fast overwhelming yeah totally actually i was having this conversation with my mom the other day if i talked to her like twenty years ago there's no way she would imagine that she'd be like all in youtube all the time right and checking all these websites all the time i mean like life for like people all ages has already like dramatically changed and it is amazing that we just adapt so it it feels like yeah we are gonna be catapult into this crazy new future and i think people are just gonna get used to things over and over you're right i think you know you'll you're totally right actually so i was just laugh about it it's like things are gonna get humorous like are gonna be very unusual and that's great just fun and it's so laugh about it i i think yeah as long as it goes well no i guess you're rights that like humans are very deductible right any of these true that like we've guns through many chin like waves of change like if you think of covid it right it was like alright i guess that's life lifestyle you know we just get used to weather is going on like almost alarming quickly right my first million hosted by sam par and sean burr is brought to you by the hubspot podcast network the audio destination for business professionals my first million features famous guest like alex ram sophia ama and hassan hassan min h sharing their secrets for how they made their first million and how to apply their learnings to capitalize on today's business trends and opportunities they recently had a fascinating episode about how you can scale a profitable agency with zero employees using ai agents listen to my first million wherever ever you get your podcast one other thing i've been thinking about too is like to get back more i'm talking about like some material things people who are listening like business owners and things like that if who listening one thing is really crazy and i was like it's getting really hard to predict where anything is going with companies like i've seen like perplexed announced comment so they've got a browser now and i like okay well maybe the browser is gonna touch it's like almost like a new operating system in way with ai it's gonna touch everything you know you guys are like interacting with crm now hubspot connecting to quad mc claude announced that they're going to be you know doing emails and managing your contacts it's like it feels like almost all of the different companies or like stepping on each other's toes in one way or another and i think in the past it feels like oh you know google's doing this and you can kinda predict what they're gonna do next and it feels like all the major companies are all gonna all fighting for you know things that have a lot of overlap right now yeah i've thought a lot about it there is one friends that i argue with all the time about this because you know look if it is true that good production is going to basically be free if it is true that you're gonna to be able to build any product you want in one prompt like there's a very big f here right but if you really believe that you can build something even perhaps as as like crm in the limit right or like an email client which just a prompt then it is true that like the last mode left and the last barrier to entry it is distribution i think it really means that like production and product is being to basically be with zero and i think we're gonna have this like immense conglomerate of mega apps you know and the kind of conglomerate that like so far right now there's like two mega apps those like the google ecosystem and like the microsoft ecosystem right like the office suite like those are the two players right like they give you email they give you documents to give you directory they give you file system all of that stuff like like team collaboration all of that stuff right it's telling that the only two companies that have been able to pull this this up or today like two or three trillion dollar companies it's just really hard to build that much product right you know and i think like salesforce has been trying to be this bigger app as well but are doing it in a different fashion than like they're acquiring the choir qui for like the office suites they're acquired slack for like the collaboration and all of that stuff i think it is likely that we start to see more mega apps yeah this product is going to be less of a differentiator and less of an obstacle and as a result i do think that it is likely that is the intensity of competition increases dramatically over the next few years you do angel invest in some startups correct yeah great but how do you think about that right now because like i recently invested in a i'm not gonna say the name but like a start up that's doing come you know kinda mixture between like using ai for email and crm and then i see what you're doing i see what hubspot doing i see what clyde just announced i'm like okay yeah so maybe they already have like ten competitors and i thought they had like one yeah yeah how are you thinking about that with your insulin i don't i i think like one of the advantages of being an angel investor instead of a overseas that you get to be delightful un about your investments like you get to do like vibe based investment oh this is a smart guy i like him and i like the you know yeah total kind of his gym whatever the product is great like yeah i mostly like a product guy like i that makes total all sense just is a ride like luca i was one of the phil checks in lovable i was like yeah i can type this thing get a website yeah this sounds great you know think my mind i i should've have anton was messaging me and i like i was busy on though i started writing checks like maybe like three or four months after that actually my first angel checks and i messaged him later i was i'm so busy y'all check out sometime time and was just like yeah and never checked it out i actually met recently vc and this is my point about angels getting to be un i met a vc who passed on them and as they told me you know we couldn't get to a point of view on like the future of the category and who's going to be the winner and what's gonna be the mood and i'm like i totally get it i i get things from like their point of view because they signed like huge checks you know and i'm i'm so glad in that in the shoes like yeah the vibes were great i loves the product i love the guy just just put the check yeah yeah and he was hustling on twitter like super hard back then remember he was putting out tons of videos are like yeah i'm building this i'm building this with it yeah so i i i tried to not think my enjoyed investing and he's working out of us invest exactly you know the mid kill i think that's kind of that it's like invest in good people and good products and the right it's gonna work out right makes excellent what products should end with like some key takeaways of like how can people you know anyone who's listing to this right now when can they try the new lin and like what would you tell people like what are they gonna able do with this that maybe they couldn't do before yeah right now you're go to lin dot ai it those a big text box and you type what you want in there and it's gonna do it for you i think again that's that's the beauty of if it's like ai agents or no longer like a good thing that like you see a lot of people hype about on twitter second you you got a text box to type something you get it it's that simple what people are gonna be able to do with it like think of it as a really smart really cheap really fast turn and you have infinity turns now and so like that's my point to by like the analysis is like i have infinity in terms and imagine i have like a full time there doing like insane competitive analysis that's what they're doing anything that you find yourself doing more than twice basically any task can now take you about five minutes to automate fully end to end and so if you're gonna spend more than five minutes doing this task over the next month just create an agent for it and basically if you do it well over time your job should consist more and more of just upgrading your agents just office orchestrating you on the agent ecosystem inside of an org who do you think are the many people who use lend and like and and do you think you should have someone who kinda manages all the agents or like everybody not as simple everyone should just they're gonna have to to do it on their own for whatever role they have i think it's both we do see companies that tends to be a champion inside the company like every company has got the ai guy right so like that's really helpful to have yeah yes you should have that person that creates a bunch of other ai agents but again if you fully embrace which we do like if you fully embrace the vision of the ai employee should every employee pulled up to the same person inside the company i don't think so right i think you really want everyone in your company you tell them basically everybody got the promotion congrats your major of like a an unlimited amount of interns that's right the the more creative you can be you know more productive you can be you can that's more and more of them yeah that's right and i think so far you have a centralized pi that built jobs should be to facilitate these transformation inside the company i really view it as it's the same thing as digital transformations like we're moving from like pen and paper and fax to computers and from computers on prem to cloud like it's the same things like you're moving to from like humans to like an ai first company and you expect everyone to pitch in and everyone to turn themselves into these managers of ai agents and so you sort of need like a little bit of you need someone until like constantly be like a winged in the backs of people because otherwise the sellers goes is going to win and and people are gonna get stuck in their habits mh but yeah no i think like the functions i i think again go to market i think prospecting is a really big use case like sd seo account executives like coaching people is a really big use case like we've got ai agents now sitting in every meeting in the company and like they report on meetings they coach people who are in meetings engineers like it's really useful for like good reviews for like qa as well to like test your application for recruiting is really helpful like it's it's basically form of sales recruiting for counting it's more and more helpful like any like sort of like batch document processing like we've got a company that receives a lot of invoices they have an agent it uses computer use because it uses quickbooks and the quickbooks api is not good don't know if anyone quickbooks is listening to us but like the the api is bad and so we were never able to make it work and so we just now we're just going through the backdoor and we're using computer use and it just receives invoices by emails looks at like the invoice number all of that stuff finds the vendor on quickbooks like enters all of the information again it's a red spa thing film i i i sit here think about like what this is gonna do to her performance reviews like i've successfully set up ten lin you know someone comparing some sm these other ones they actually got this person got thirty lin going and they're all like producing results that being like how you judge how will an employee is doing based on how many you know lin they're properly you know have set up in in operating you are kidding but we have one employee and like they're killing it everyone's like oh my god like they'll the go don't know how they do it and we as a user of lin you receive an email at the end of every week that's like lin saved you x hours this week and this employee share the screenshot of their report and their report goes like lin saved you like the equivalent of forty full time employees this it's it's like two thousand hours it's like oh my god that's how you do it like we basically have like a whole different org inside the company we even know that so like this was just got like a whole org of like forty eight agents working for them it's actually much more fully in like thousands of agents a second equivalent at a full time employees will of work yeah and it's got while like then do you pay them a lot more money because of that or like the company owns the set up with the agents that's kind of complicated the company does own the agents so don't give them ideas it's like i want you let's saving you forty times money give me ten times more money you know that's no it's also costing us a pretty penny employees at lin have like unlimited lin credits but like it's very expensive actually the tokens are so expensive mh and so our will own lin build is outrageous and we're very happy to do it because like in the end our i will overall burn like compared to how fast for growing actually really slow really small mh and that's because like i actually think we spend almost as much in tokens as we're do in humans wow is most of that money going to ant philanthropic or open ai or ant is the default model today and yes most just being to topic expertise yeah one other thing when you were talking about all the different use cases can you create invoices and stuff with lin yet i'm still using fresh books to send out invoices for a few different things and like it's such a pain it's like why am i paying these people to just generate invoice and email it it can do anything that you can do on the computer okay awesome i gotta try that so computer use man just try it and yes it will create invoices for you awesome cool what it's been great to have you on here i think this has been a useful episode for people and anyone who's listings they should check out lin and try it out yeah i'm percent yeah thanks for having me yeah it's been great thank you

what if your fig design could become a working app just by asking in this episode of the next wave our guest is david ko from fig he showed us a demo of fig make a mind blowing new tool that turns your designs into live working code with nothing more than a prompt whether you're a designer a founder or just someone dreaming up ideas on the train this is the next evolution in software creation no more handoff offs no more designer versus dev debates just collaborative multiplayer magic live in the browser david walks us through with how fig make builds dashboards games and real data powered apps on the fly he shares fig vision lessons learned from their first ai rollout and with the future design and creation might look like in a world of ai native workflows if you wanna see the future product design this is it but before we get started you know we've spend a lot of time on these episodes and it would mean a lot to me if you could hit the subscribe button it helps grow the show and get on better gas to provide more value to you so without further ado let's jump in with david ko from fig cutting yourselves like well half sounds pretty impossible but that's exactly what sandler training did with hubspot they used breeze hubspot ai tools to tailor every customer interaction without losing their personal touch and the results were pretty incredible click through rates jumped twenty five percent and qualified leads quadrupled and people spent three times longer on their landing pages go to hubspot dot com to see how breeze can help your business grow david thanks for coming on the show thanks for having me super excited to here yeah you know kind of a crazy story i've been a user and a fan of fig for a long time actually recently one of my designers my previous start bind it he told me you've got tri to fig my out and i was like okay i'll try it and i've loved it ever since been awesome to see how big fig become a house become kinda the standard and then when ai started happening i was like oh god hope fig gotta be okay we just everyone's using fig now all the designers but what happens when ai starts coming into the picture how is that going to change things and i saw dylan introduce fig make you know maybe a few months back and like okay i get someone from fig on the show and also just to learn you know how it works and like where you guys are planning to take fig in the future yeah i think people wanna start off like explain what fig make is a little bit and then let's just show them how it works yeah and maybe before jumping into make i should mention so fig me ai has been around a couple of years and we have a ton of ai features embedded in all of our products so you can use them in fig jam to summarize stick and help you brainstorm inside fig slides you can generate images and rewrite text inside fig sites you can fill in content as you're building out your site we've totally rebuilt our search engine to use multi modal embedding so it can search by text by frame by image but the thing we're most excited to talk about and focus on today is our biggest net new gen native product which we're calling fig make yep so fig make is a prompt two code tool that lets you get started with either natural language or a screenshot or a fig frame and it will write code for you and show you an interactive playable version of the app that you can rip with we've built it to be multiplayer first so multiple people can be in here chatting alongside each other oh using code visuals and language so just to give you a quick sense of it i mean it's one of these starting prompts for a data dashboard so you can see here the prompt is creating intuitive and visually appealing dash word that provides users with a whole bunch of more specific things i'm gonna go ahead and send this off and you'll see here the model is starting to reason so it's starting to think about you know what different components it's gonna need it's summarizing its action plan and now it's going ahead and starting to write some types files over here we're seeing the code starting to stream from the model based on this prompt there's actually a whole file structure different components here of code that's been preloaded and generated by this and sometimes it can take a couple of minutes depending on the complexity of the thing you're asking for so while this goes ahead in streams i'm gonna pull a little bit of a martha stewart and swap this out for another version of the exact same prompts which i already ran ahead of time so this is same prompts you know you can see the the whole set of reasoning it did and it went ahead and created the first version of this dashboard application you'll see one thing it just immediately followed up with is please note fig make is not designed for collecting personally identifiable information but we can actually add a whole database to this to make the data real oh so we can go ahead and shoot you that in a second but we recognize for lots of applications from prototypes to internal tools to string up things you wanna publish being able to persist data real data and connects the data sources is really important so anyway we'll look real quick at what it actually made so it says you know here's the dashboard you can see there's some interactivity activity and the hover states on this chart here a couple different charts showing different breakdowns so it's essentially actually you know looks kinda nice there's a hover state there's a list of project entries some of this stuff is not fully hooked up but it gets you a started at this to like a great way to like prototype something yeah totally and i think you know one of the things we've really seen too is that this brings a whole team together where you know we've really optimized for designers which i'll talk about more in a second but everyone can come in here and start r and building on this simply with language and so i'm just gonna go ahead and create a new project here this is a test project using super base which is a third party tool we integrate with that basically can host data and so you can see here it stood up a database based on the underlying sc of the prompt of the project i was describing where it's gonna store this data and it's also starting to write now server code not just front code that we'll be able to handle what happens when you click buttons and you wanna add new things and store them and so this is a really powerful piece of functionality to make these things long lived that's awesome and so we'd like they just figured out the sc you need and i guess if there's anything that you might have missed it'll go and update the database and they add have new tables and everything like that yeah exactly and this is something you can iterate on over time say add new columns add another table for users not just for you know revenue as an example right here and so yeah this is a you know really powerful sort of building blocks to make even richer experiences and so a lot of what we're starting to see people you know use this for is building richer sort of higher fidelity prototypes in interactions to prove out ideas and get more feedback on them and so traditionally in the product design process you know making mo and fig great even making traditional prototypes and fig where you hook up noodles to connect you know workflows is really powerful but also it's kind time consuming and there's also a limit to what the fidelity is of those experiences you know there is no data stored on them you know they're not actually code back so you don't get the full fidelity of of interactions and so what we've seen is people are using this both if they don't have the skill set to engage in fig design is fully as other you know professional designers and also for designers to pick up where fig has traditionally ended you know a lot of designers don't know how to code you know that deeply mh and so it feels like a superpower to get something where you can bring in your designs and keep going yeah totally i mean like i'm an okay designer in an a okay code i'm not amazing at either one and i've always like when i used fig mo i was like okay you know i spent a lot of time making something beautiful and then to go actually coating it oh my that's a lot of work and usually when i go code it i end up getting lazy and it looks slightly different and like okay that's good enough this close enough so i was like if i could actually get my beautiful design and like press a button in the future and then be able to tweak an ai that'll be amazing only hundred percent so while this is checking on on the database piece which you can imagine i just wanna to show sort of a couple other pieces so maybe just to pause the database example to give you a sense of how this works with design systems we know design systems are you know really important for teams know they create consistency for users and what you expect you know out of the products and so they allow you to you know have an level little consistency but i'm super curious like how does the design system work with this so i may not know the full functionality of like regular fig because i know you can like save colors and things like that fonts yeah is it beyond that can you like say your button types and all kind of things like that and yeah great question so basically taking your design systems in fig which have tokens and variables they have you know colors fonts themes and converting that into something that can be used basically as an input to this generation and so actually over here i did have an example where i converted the same output into the fig simple design system which we've opened sourced and using those set of styling instead so you can see it's like you know now using this sort of fig branded aesthetic and we're excited to do over time much more here you know here hear a lot of requests to like use my literal components and i can show you the workflows we have for those today as well yeah it seems powerful like i've been using cloud code and factory both a lot more to like create you know prototypes and his play around and you know one thing is like getting them to actually follow design system like the best thing i figured out is like can kinda like map out part of your design system mean like a markdown file or something in hand to them but that it keeps forgetting so if you had something that's directly in fig is it seems better to me i'm blame yeah and i should mention as well like you know people are used to breaking and editing in fig so it's not justin in prompts you know you can use this kind of magic pointer tool and starts to go ahead and just refund things in line and you could do that with your team is the multiplayer aspect of this so you just yeah yeah exactly so i can go ahead and share this file and everyone sees the same thing just like you back from fig design yeah and actually multiple people can chat so you see my avatar here in a multiplayer file you'll see other people's avatars with their prompts and something we've invested a lot in is making that work across modalities so if if someone makes changes you know visually here versus via prompts or even versus via code it all plumb through to everyone else and so we want this to be a surface where an engineer can come and r with the pm and the designer and make their own version i'm like okay here's how exactly this should feel and how does the multiplayer work with the prompting like i i assume it's one person prompting at a time and they're like what their when kinda seems like okay it's i mean they just see it happening and they hit to wait i guess yeah yeah good question basically what happens while prompt is running is it you know you get a analytics state and you pause run new anyone and server everyone gets that same experience can either like wait for the thing that's been running to finish running or you can kill someone's stuff for you and we find that works pretty well her and you know if if i wanna start r on this i could have like an inline contextual prompt for it but i could also like jump on over and see exactly the piece of code that corresponds to this and start changing that directly as well there if i have that context you know i grew group on irs for for me saying this i'm like imagining some crazy collaborative project where everyone just like messing around and making something together and like just sharing the link online it'll be fun but yeah well i think how teams to actually get real use out of this sylvia yeah hundred percent yeah i mean i think that's been the magic of fig always is like fig was born on the web it was multiplayer first it's just sharing a link i think there are starting to be more and more of these you know prompt to code tools but think you know we have put a ton of effort in into making this work really well with your team across different personas and modalities yeah i i'm curious that i mean obviously you know lovable has been taking off and then there's v zero and there's a few different ones so i mean some of the functionality looks similar but you know what what's the main difference the main difference is it's more like designer first and you can actually kinda get in there and tinker with the designs and and things like that you know i'd say first and foremost this whole category is really early yep you know like we're talk about the whole space like none of these things existed like sure and it's like this whole space is getting figured out right now which is really exciting and a lot of that is you know building on the new capabilities of these models and trying to figure out the workflows that makes sense and so i'd say you know right now we've invested a ton in multiplayer as a sort of core thing you come to love and expect from fig and a lot on inter op with the rest of fig you know say you have a mock you're trying to build a sort of interactive visualize for solar system this is a traditional you know fig layer just from fig design gonna literally hit copy paste on this thing and go over here and say make this interactive and what you'll notice here is like there's a little fig mic on here like this is not actually just a screenshot this is all of the rich structured data that comes over with this it's every layer it's every token every variable every property and we have done a ton of work on quality to be the best place at generating high fidelity results from designs as a starting point and we've seen a really high percentage of our users start with designs as well and we've optimized for that you know even in the blank prompt box we say at the top like put your design in here and so i'd say one of the things you know we've really focused on is taking the workflows designers already have in fig and letting it go much further to express things out don't higher fidelity through code through interactivity in a way that's much lower effort than people have had before okay let's be honest your ai prompts aren't giving you the results you deserve but with a little coaching you can transform from basic prompts to engineering conversations that get you exactly what you want from chad gp that's what this playbook delivers not just random prompts but a step by step system with the exact techniques top ai professionals use every day you'll have your own personalized prompt engineering system that delivers consistent results get it right now scan the qr code or click the link in the description below now let's get back to the show when i think about fig you know when it was first suggested i use it i used to use photoshop i still use it a little bit but you know using photoshop to create the designer website is a pain like just dealing with different layers there's a lot more manual work to do small little changes and then you get the fig it's like everything everything's easy to move around and and then only like that but then like you know you guys started giving all the like code or you could back to the css rules and all that kind of stuff you can use and so obviously this is like a natural revolution like okay now just give you the rules because then if you don't have a code that's hard to like know what to do with but now literally just create the code for the designer totally yeah and so here this example finished cranking so this was gonna bring in a you know a mock that how different itemized planets or the solar system and you can see based on literally make this interactive is the prompt and the underlying structure of the design that we brought in it inferred that you know each of these different planets is its own entity moving at different speeds they move around the circles it created you know this functional speed scrubber that works if i click on one of these planets there's an overlay that comes out did it know that because of the layer names or did it actually look at a picture or like how did it know that try to infer that yeah that's right so you know if i go back to the original design here like i can you know drill down a bit and you can see you know each of these different planets has been sort of like itemized with the planet name so like you know uranus orbit along with the arc that it takes and so like we found this kind of context he's like incredibly helpful for the model and so like you take a random screenshot and give it to any of those these tools like we'll do okay especially if you could write a longer prompt but to the point of like where is fig different is like fig is already the workflow design teams product teams are using to articulate design intent to detail out specific workflows and that data is super super valuable to model and they can pick up so much context from that to make the output really really good with all less effort makes sense it feels like you guys have done a good job when i was thinking about what think about doing ai i was like well right now at least from like twitter discourse kinda know twitter's kind of a bubble but you know it feels like the design community is very passionate about things and in terms of ai right now you know there definitely been some cohort like the design community this the kinda anti ai or at least that's what i've been seeing on twitter it feels like you guys are kinda doing a great job of making it clear like hey this is not replacing you like if i was a full time designer this is amazing because like for me whether you're designing or coding you know it it is the art of creating things right you're you're creating things out of nothing and now the idea that you could be a designer and not know how to code and you created something out of your brain and turned into real design and then now you turn to a real product that's just like that's amazing i i don't think like both people relative like how magical that is hundred percent yeah it's been totally uni inspiring and i think we've been blown away by just how much people can make with this and so maybe just to finish the demo section with a couple fun examples from no customers you know this is one of a map and like you know this looks a lot like a mock and i think traditionally designer might end here but like you can actually use this math and like you know that's like one prompts to add that level of banner activity and like you know imagine the feedback you're gonna get in a user study around you know exactly how this works to find coffee shops yeah based on this another fun one this was for a conference collecting all of the different icon geography and the logos for a team that was planning the conference and so it became a little bit of like their micro site for their design system out of the conference that they all put together in and published and sort of added all this wrap around with fig make kinda so we've seen internal tool start to really take off as well and so this was a really fun one of like a color palette generator that someone on a design team built to make it much easier to get the right sort of compliant colors with the exact mixes in different scenarios and like this is the kind of thing that you know is so much more powerful as an app that you can explore and send around relatives like a static thing and you never would have gotten a you know internal tools team to fund right but now someone can actually like go ahead and post their team because they don't like to wait right they just cut the idea and they design it and then and then it makes it totally and the last one like i've just been blown away by stuff we didn't optimize for you know based on how good these underlying models are and so like people have built full on three video games in this which is like kind of insane here's one that someone built which is i haven't actually even finished this game there's like a whole different world a bunch of different spells is like features that are kinda come out and start going after me and i'm like wow this is insane that like you know literally just through prompting you're able to get to this level of expression you should make a like a old school rpg with this is or i used play request that's actually how my career got started yeah yeah it reminds me of playing every request and like kit mobs and like you gotta do the failed and you gotta run little bit then there's this stuff coming at you yeah someone actually recreated dude oh cool which is pretty well very cool my first million hosted by sam par and john per is brought to you by the hubspot podcast network the audio destination for business professionals my first million features famous guests like alex h sophia ama and hassan min h sharing their secrets for how they made their first million and how to apply their learnings to capitalize on today's business trends and opportunities they recently had an a fascinating episode about how you can scale a profitable agency with zero employees using ai agents listen to my first million wherever ever you get your podcast so i'm i'm curious like who is the user of fig make is it gonna be primarily designers or are you guys trying to reach out to people beyond designers as well because i i definitely see how this becomes a super superpower for designers i am curious you know how other people side design would use this or how they get value out of it if they're not a designer or or if you guys don't care about that you're just serving designers only you know and if other people use it that's great you yeah yeah great question yeah we optimized for designers to start and so we've done a lot to you know as i mentioned bring over that context from fig design but also make this a great place for a whole product design engineering teams to work together and so the fact that code is there and you can change that the fact that you have a pointer and directly we you know change different aspects of the experience like all of these have been designed so that your designer your pm your engineer can use this as a collaborative multiplayer surface to iterate on what the product experience should feel like together which is what fig been great at historically as well and so you know that's definitely they've been the core right now i'd say you know we've started with prototyping as a heuristic scenario and we've been surprised how blurry the line is between what is it prototype and what is production okay and it's soon you can publish and start sharing with users and getting feedback and start using code libraries to hook it up to real back end to storing data like you know we've seen people pitch vcs and get funding on the things that they've built in fig make without you know anyone else involved with real users using it yeah wow and so i think we've been just totally blown away at the breath of cases people are already using it for yeah and maybe to share one more example we did a ton of dog food on this before it went live it was really helpful in iterating on quality we often generally have a dog f oriented culture at fig mo but this was over the top that we called it the great fig bake off and we had like twenty different meetings scheduled for the whole company in every time zone where people from the team would be there to answer questions or troubleshoot or help you work with bugs we had a competition like what were the coolest things people made and all sorts of people made inspiring thanks and i think one of the coolest ones for me was someone on the hr team who had zero design background zero technical background they used workday as a source of truth for people data they found it work the api and they built a game that uses the real people data of the company that shows four paces and four names and you went to guest goods who yeah and it like became viral within the company of people like playing this game and trying to you know get no other colleagues crazy and i was like wow this is something that like no one would have funded as like an internal crap what the hell are you talking about right it's why pitch that right but yeah let's go and resource doing this and it's like you know it's a persona we didn't optimize for it like we didn't build finger make for hr people right oh but it's just been totally inspiring to see you know how accessible it has made the process for everyone it's awesome my friend who used to run a design agency in san francisco just imagine you know for him like you know when when you're pitching new clients if he could just like design things and then just had a prototype there it's like that's just nuts right compared to like oh here's a design like here's a prototype that works you know yeah that's just totally yeah still wanna go back to us said previously about like if you guys had any backlash on this from designers because i assume you guys got some i think i saw some online a little bit when it first got announced but i assume you know if i was a designer even if i didn't like ai maybe at first i'd be like oh this sucks at ai and then i'm start using it my day to day work and it's just like oh it's making my life so much better so like how can you keep hating it is that kinda help with yeah you know i'd say just to zoom out from the story of make to the sort of fig may you know we shipped stuff last year at config twenty twenty four we shipped a feature called make designs mh which was our very first attempt at a prompt to design products inside of fig design and so you could say you know make me a landing page for and you would create it not as a code prototype but as native fig frames and we messed up in a bunch of ways on that like we messed up explaining how it works and so there's confusion on that like it use off the shelf models and templates that are customized and it was funny we had called the project internally first draft and the whole point of it was like solve the blank canvas problem get you started make it so that like you know we're doing your laundry new chores and you're doing the creative part as a practice design team and i think make designs even in the name implied much more than that and it's like we're gonna do the whole freight which is not designer you don't need to do any design just we'll do it for you totally and like you know we we are a design company we are four designers and so i think it was definitely a big learning and like how we talk about these things and what they're for and i think with make i think what we did right was we started with things that designers have trouble with and they can't do which is getting into the code layer and so we can sort of start at the end of the process where a lot of people do get sucked today with the current set of tools and so i think it felt like something net new as opposed to automating parts of the process that i have a lot of opinions expertise taste around which was not original intention but you know easy to to come across wrong and so i think that was definitely a big learning you know for us in general as the company that pride itself on being advocates for design as a function and champions of designers is like making it i'm really really clear that this is kind of trying to give you new superpower i am curious like where does fig and fig make where this go like in three to five years how does this work like if you're a designer yeah i can't make forward looking statements why i have two kits i have a six year old and a three year old and my six year old yeah over the last year started playing ipad games he's like addicted to duo lingo he's like a daily active user awesome wearing math and chest and that kind of stuff and he started using make before we launched and i was like curious to see as a very much not target persona right like how understandable would it be and he is like hooks like he on you know begged me to like spend more time on make and you know walking home from school or you know at the grocery store running errands he'll be like i had an idea for a new one betty oh that's so like you know he's made like six games now he'll sometimes come home from school with like a sketch of like a map or a level and he's like take a picture of this like we should use this as a starting point moment and honestly it is like mind blowing and i'm so jealous like right yeah imagine being a kid like you you have any idea like i can actually make this now the ai won't help me make it i i'm an eleven year old son you know he grew in san francisco ran a bunch of like tech and game industry like executives that people like that but he was little until you got see all this kinda like creative world there right like artist artists like like that we used have like artist parties where we draw in the walls and he'd be there with artist drawing on the wall but now we live in japan you know japan's great but he's no longer in that kind of environment i i've been showing him things like a rep i showed him rep he loved that like he made like a little site with that but the websites were really ugly so i was like okay you know and he didn't really like how ugly it was which i think rep and those guys got a little bit better but i think some like fig make where he could actually like go in and and tweak the design itself too would be awesome for him to be able to see the whole process of like making it look good and and then actually shipping it would be incredible yeah i love his this feedback and important parts and chris okay i'll get the done i think to your question i'm like you know what's the future look like you know find years from now i think it's like making the path from idea to production i did so reality so much work and like you know seeing even game toy examples with my son makes me realize like the way people think about the workflow the way you think about what's possible the way you explore ideas is gonna be totally different and i think in this new world design is even more important than it has been and design isn't just pixels it's empathy for users it's understanding of your target audience it's you know knowledge of current workflows and where they fall short it's exploring the whole solution space and thinking deeply about each of the nodes that you might not have had time to you know in the old world yeah it's getting down every animation and interaction of the experience and so one of the things i'm really excited about in that future is like software is gonna get freaking awesome like everything is coming back so much more thought through so much more explored every fit and finish is gonna be better and by the way there's gonna be a lot of low quality software too yeah totally like there's just gonna be more software than has ever been made it's like in the early days of the web that's why website said the podcast before now group and iris like the web you used be like fun you know yeah it was messy it was you know with total mess but it was easy to get there just like tinker and make websites and now you don't really see that eating more so i'm hoping that stuff like fig my make brings that back we're like yeah the web becomes fun again because people can just have ideas and just easily make stuff totally yeah i think we'll see like you know even more long tail personalized experiences for different schools and communities and like people building stuff for themselves yep and the stuff that's built for everyone being like even more sophisticated and you know helpful in in a wider range of i think your applies some really interesting experiments around collaborative creation of things too i think you see like communities like actually creating new kinds of projects and stuff together will be fun totally cool so you know this is a total theoretical scenario imagine you got a time machine and you're in the new york time machine you got out it's twenty fifty what's different anything different it look the same it's still the same bagels and pizza is there other stuff going on in new york or were going on i mean new york bagels since pizza are not going anywhere there's hopefully honestly twenty fifty is so far right like ten years let alone that far it's like man i you know i have aspirations there's whole conferences on this stuff and it's like you know very hard to predict in everything right it's something i like to rip on and like let look hear what people say it's a great prompt i mean stuff i'm like personally excited about way am coming to new york figuring out how my kids get play and stuff is hard excited for ai to get your level of trust and safety and rollout out to like the point where getting around in the world is like a solved problem and like that seems like it's happening which is super exciting like way moe was fantastic health is really hard i personally have celiac disease and was diagnosed about two years ago and it's been a whole learning journey about you know what can i eat and what are my sensitivities and so forth and lots of blood tests and follow ups and you know even for me chat has been like so helpful in understanding what the terminology means and all that and i think in broader access to that level of expertise with verifiable trust and awareness is gonna be huge for society just like people understanding what's happening in their lives and things that are really core yeah from health to finance to dealing with the insurance because they're house burned down to all of those kind of core fundamental things that are like anxiety provoking complicated and scary and it's hard to get experts to actually understand the whole system and so i think that's gonna be like incredibly meaningful in a ways that are like hard to really internalize still right and then there's like the really out there stuff of like are we all gonna have neural lengths or glasses on our faces that like understand the world around us and are chatting with us so we're gonna have moon colonies by then that everyone gonna be living longer like i hope all those things are true and like you know i'm an opt optimistic in a tech twist but you get into like singularity theory type of questions and yeah once you report to singularity all bets are off who knows what the hell happens right yeah the health thing that's something i think about a lot because like you know if you think about a doctor they go to school for eight years or whatever it is and a lot of much about reading a lot of books and trying to retain as much of that knowledge as possible and maybe if you're lucky they retain like five percent of the knowledge or something and then over time they get more expertise by dealing you know with patients but obviously ai could take in all that knowledge and over time actually understand all that we've learned and then start expanding upon that and then combine that with also liking more data on us individually you know even like down lowest level possible it seems like ai will be able to understand us in the future like eventually we will have ai at like okay nathan what's going on with your body right now like you know there's like a be things totally unexpected like oh some cancer started for this reason and you know and then we can kind of solve it this way that sounds like sci f five for people by i i really strongly believe that's coming it may it may take like twenty or thirty years hopefully it's more like ten but that's gonna be such a game changer for people yeah i'd say also you know just on the you know the role of ai and these in the workforce us as well you know i think we'll see as it continues to evolve and capabilities come out i think in every sort of industrial breakthrough and revolution has been concerns about what it means and ultimately people have learn of new skills and been even more productive on things that are even more meaningful and so i'm yep personally really optimistic and excited and i think seeing people like my son start to use you know fig make and tools like it totally new ways makes me feel like you know there's just new frontiers to work on and like what work means will change yeah is a designer i think you know when we have robots building things i mean like what being a designer it's gonna be it's gonna change maybe you're like almost like an architect you're like literally like designing new building you know on the moon or or mars or or wherever we go and you know a designer bill to actually like come up with things like make ai assistant maybe do they do themselves whatever but call a combination and then the things become real things and digital things it's it's gonna be such a crazy time to be alive totally yeah i can't wait for that cool man there anything else you wanted to talk about where we got here or honestly i was kinda curious like your head takes too and like you know you've tried a lot of these tools you know what's your sense of where the market's going you know how you see fig fitting into it so i i've only played with fig make a little bit quite honestly i played with it yesterday i like to have people on and like trying to have as like a real response to what they're showing me as possible versus like yeah i've actually tried all of this i know everything you're doing and now i'm faking and pretending like i don't know what's going on or like i'm learning something new i play with it a little bit to exactly what i was hoping it was especially the fact that you can just like like import the design in and hadn't like start from the design because i feel like the market like you know what lovable doing it's great i do feel like it's saturated there's gonna be there's a lot of players in that space so i think you know i've also had on mat baton the founder of factory who's competitor with devon and and like devon in fact i think that's interesting that's called a different side where it's it's all agents and it's more hardcore and like the enterprise and coding and so like it makes sense that you're not competing directly with level or or v zero you're doing something more for designers and so i i think you guys are on the right path and you know it definitely seems like way better than what you guys announced in twenty twenty four quite honestly like i've ever saying that like oh this is you it's not i don't know that's all i it's kinda confusing i can see why designers didn't like it fig makes seems like the right path like this is something that give designers super superpower i you know haven't played around with the product enough to give you like honest feedback like oh it's amazing and all it all works and but at least conceptually and for what i've seen to the early version of it it looks great cool we yeah appreciate any candid feedback at any point yep i'll play with it more you know as i said we're just getting started you know in general an ai in this market and i think it fig ai as well and we have a lot really exciting stuff coming we're very focused on this so you know any and all feedback from you your listeners really excited for it and yeah stay tuned for some fun stuff ahead awesome it's been great having you on here is there anywhere like people can find you online or are you all active on on twitter or on linkedin or anywhere like that yeah i'm on twitter in linkedin david ka k s m i c k cool and yeah thank you for traveling me on this was super fun really appreciate it bye thanks

welcome to the next wave podcast i'm matt wolf and if you've been keeping an eye on how ai is changing the way we work you're going to love today's episode i got to sit down with nicholas holland the head of ai over at hubspot where he spent nearly a decade driving innovation in this conversation we explore the massive shift happening right now from ai as a simple productivity booster to ai actually doing work on your behalf we talk about this new front here nicholas calls work as a service the rise of the super contributor in sales and marketing and how agents not just assistance are quietly taking over repetitive tasks in real businesses you'll hear a four step framework for getting your team started with ai why structuring meeting data might be your biggest hidden asset and what it means to become an agent manager in the near future it's practical it's forward looking and it's packed with insights you're gonna love this one so let's dive right in with nicholas holland cutting yourself cycle will half sounds pretty impossible but that's exactly what sandler training did with hubspot they used breeze hubspot ai tools to tailor every customer interaction without losing their personal touch and the results were pretty incredible click through rates jumped twenty five percent and qualified leads quadrupled and people spent three times longer on their landing pages go to hubspot dot com to see how breeze can help your business grow well thanks nicholas for joining us on the show today it's great to have you and excited to dig it and nerd out with you about ai how are you doing today good thanks for having me matt appreciate it yeah so let's get into a little bit of back i don't wanna go too deep on this but like how did you get into ai what's your relationship with ai and how did that sort of evolve into what you're doing at hubspot right now yeah so i been in hubspot for almost a decade i started off right in the innovation labs out of ireland then i took over the kind of content and marketing portfolios and eventually became what they call a gm so just responsible for product and revenue over the marketing products and ai has always been something that you know customer platforms like hubspot have had to do but it's really been like on the machine learning side i think of it simply as like scoring things so lead scoring has always been a big part of that mh when open ai came out with three point o there was rumbling in the marketing world about gen ai and so we started to look at that chad gp comes out and of course the world changes overnight everybody's using it talking about it and then we had multiple ai leaders and for whatever reason they didn't work out there was a point where they were looking for an ai leader in three years had gone by and marketing is being impacted the most i think by a our husband right and so here we are trying to serve the marketer and we had just put out a ton of value by applying ai on their behalf and so there was a point where do we go get an academic leader someone who you know major in ai and our product leader came to me and he said i think we need somebody who actually knows how to build things with ai and we would love to have you do it and i you know my specialty is in applying ai so if we do this we will become experts at applied ai and he said let's do it and so i took over about ten months ago and since then that's really been how we've just begin to drive really deep on how do we apply we're model agnostic we take you know a multi model approach and then we apply that you know we're at a point out where we don't even think ai is the limit anymore it's like how do you actually bake the cake yeah yeah yeah no it it's interesting because you mentioned marketing right my background was actually in digital marketing but because ai became such a sort of integral part of what marketers do now i started talking about ai as it relates to marketing and then over time ai got more and more popular and i went you know what i'll talk about ai and marketing but also just talk about ai because i just love this stuff but i feel like marketing is probably the area where i feel like ai is really sort of impacted things the most shoot for the marketers that are embracing it i'm saying like a rebirth of marketing being fun again saying a rebirth of like marketers kind of starting to find you know blue ocean again that was like the hardest part of like if you back the clock like two years ago it was a red ocean for all marketers everything had been figured out all techniques were basically known all the software had started to become homo now it's like blue ocean again you're seeing wild variances is in creativity tactics all that so it's pretty pretty cool yeah yeah i mean all of the really cool marketing that you've seen over the years that you're like oh this probably took massive teams and you know tens of thousands of dollars to produce has sort of been you know for lack of a better word democrat and now anybody could kind of do the really cool marketing stuff that you've seen and you know i love that that's what brought me down the rabbit hole in the first place i don't know how you say it matt when you're talking with others but like there's like a new role emerging i don't know what other companies call it but we call it like the super contributor in just the last week i've seen a marketer who runs a prospecting team now mh so let that think again like the marketer is now running the prospecting team and it's because they're using agents and different ai things to get that done so now you got a marketer doing some sales type roles i've got a sales guy that i was working with he does all of his own product marketing work now because he's using you know image creation he's doing videos he's doing all that seven so like it's just really interesting with all of these like capabilities coming on like what does it mean to be a marketer or what mean to be a salesperson person like this super contributor is now rising where it's really like if you have a lot of creativity and you've put in a little bit elbow grease to learn some of these things you got a couple of new superpower yeah yeah i mean you know guys like us sam alt and da ama die from ant philanthropic they've made comments about how they think the first one person billion dollar company is coming within the next couple years right that's because one person now could be the builder the marketer the salesperson the you know the customer support person they can do that all by using various ai tools and agents and things that are available to them so yeah yeah we're a pretty exciting world right now yep it's gonna be cool so i wanna talk to you a little bit about the shift of ai productivity to ai actually sort of doing a lot of the work for you and being able to use ai to build system because in the early days ai like when chad gp first came out and even before that when we had access to gp three and stuff like that it was really great at being like a productivity productive booster in the sense that it could help you write some emails and you know beef up your copywriting a little bit and and some little stuff like that but now it seems like more and more people are actually using ai to sort of build out their whole business system so i'm curious like what have you seen in that realm how are you seeing this shift from productivity to like now ai is sort of able to help run the show so about a year ago again like in the early state where i was first started to take over we started to notice that there's like this shift from saas software as a service to what we think is gonna end up being w people chuckle when we said it internally because it can seem a little cheesy but like work as a service mh and work as a service can look like you know we've even talked about it like should hubspot actually do marketing for people or do sales for people mh it could go that far but what we think right now is that like that is kind of the key differentiator between of the past like cloud or internet or mobile and then ai those were all enabling technologies that let you go basically faster so it's like you heard the old phrase henry ford said if i built what people ask for i had given them a faster horse right you know so we actually had that internal debate are we providing faster horses or are we providing cars and i think that ai when you start to shift to it doing work is a lot more on that car standpoint because with cars what came along you had new types of professions you had infrastructure changes with gas stations and roads you had kind of knock on effects of cities moving further out right i think that's the kind of equivalent that we'll see with ai and specifically like everybody's kind of in love with the concept of agents agents or a type of ai that get closer to that concept of work that it's basically something that's doing a unit of work i think for us what we're noticing now is like we group everything as an agent now internally mh so even like assistance like a chad gp they used to go out of their way to say like we're not an agent we're an assistant right but as ted gp gotten more powerful it's doing more work it's doing more things autonomously it is an right it's just an agent that interacts with humans so we're kind of seeing agents that interact with humans as assistance we're seeing agents now that basically you give it a goal and it'll go knock that out so that's more like a task agent or a work agent mh and when you look at that you know now you start to break it up into ways that like one you need to find parallels as a human to think about it so if you had chief of staff or an executive assistant they wait for you to tell them what you need right although some of the best ones anticipated but you know you asked them to do something and they will go do it you'll say hey can you take notes at this meaning can you let me know about this next meeting that i have coming up when you book this for me that is a body of work that's starting to happen now and a lot of people are using their ai assistance to do those type of things mh so now we have companies all the time ask like how should i approach ai one of the first things we say is give every employee and assistance mh mh we have one that's very good for go to market roles so marketing sales and service it's powered by chat gp it's got all the context of hubspot but like we say give them assistant and teach them how to start using that to basically help themselves the next thing though we move into are the agents and a lot of our customers are still thinking like what is an agent how do i go about this etcetera and so we have out of the box agents that basically a low lift you know fifteen minutes or less of time to have like a big body of work and so tier one support mh not just of service but also tier one support of marketing inquiries tier one support of sales increase that's like one of the agents we build right and then going from that once you start to now people will start to get an idea of how it works and then they'll start to build their own kind of special bespoke agent so we had a customer who wanted to build something called like an account hand agent and the easiest way to think about it they just had like this little pain in the rear that was a body of work mh that they couldn't get sales reps to do very well right so i'm the sales rep matt you are the service guy you're gonna deliver on what i just sold what happens in a lot of org because i'm like hey matt i just sold this new account please hook them up and you're like i mean what did you sell them what did you promise like can i get all the details and they're like yeah yeah you know what i'll get those over to you or they'll say can you go check out the crm and you're trying to piece me all the calls and all that stuff so again one of our clients was like can i build a an account hand agent that's brilliant yeah yeah it's hours through all the engagements it understands what it's trying to get done it then puts together an account hand and then if it's kinda like an ambient agent upon deal being closed it just now does the hand off like so i don't it's very hard to describe that to like a regular person without going deep because you're like that's a body of work that you didn't have an account hand person yeah but boy did you need it and at the end of the day that's like why i think now we're not seeing like employment collapse right that just makes that business so much better that makes that service person not hate that salesperson that makes the system now do this body of work and so that will then be like a net ad for that whole company so as a long answer it that that's like what yeah i love that so basically you have the the salesperson they're putting all of their communications into their crm you know hubspot for all of the communications they've had with that customer in the past once a sale happens then the person that actually has to handle the deliverables they basically essentially get like a quick reports essentially all of the past communication so i mean you might even have stuff in there about like their birthday or like their dog's name or something and then the person that takes it over can actually jump in and already have rapport other thing that's crazy too is like we all tend to think you're the star of your own movie so everything you do is you think about it from your standpoint we're seeing organizations now where on that one account they had an issue so they called in to support they then basically were talking to team a who has product a about this and then they bought product b right now that is all very important for like whenever you hand off stuff because that continues cross sell upsell conversation you know what things they've been upset within the past you might have multiple stakeholders so it gets very complex really quickly when you begin to think through all the different possibilities but what i love about this is like it's a very very human empathetic problem right i'm now in this value chain i would love somebody to hook me up with knowledge about how we got here and this person has already done their job they don't wanna waste a bunch of time handing it off to me and so i love it because it captures the moment that we're in that there's real value here it doesn't have to be a proxy for an entire job role the right business is happier the customer is happier like that's that's just a very like empowerment moment yeah yeah you didn't cut any humans out of the mix so you just made the humans involved a little bit more efficient at what they're doing totally real quick if you're enjoying my conversation with nicholas you're going to love this nicholas is giving us his three step framework for ai adoption it takes you from complete ai overwhelm to clear competitive advantage in just ninety days if you're drowning an endless ai tools nicholas has cracked the code on analysis paralysis his ai adoption playbook walks you from personal assistance to team agents to custom development with zero disruption to your workflow get it right now for free click the link in the description now let's get back to the show so i wanna jump back to agents for a second there's a bit of a debate right about like how do you actually define an agent i've actually seen people get really fired up about this over on x for some reason about how you actually define an agent and people getting an an argument about that's not an agent this is an agent so i'm just curious like how do you guys actually define an agent what's the difference between that agent and the assistant boy we get fired up internally as well one of the ways that i've learned to navigate this is you know inflexible thinking will be a real detriment to people during this ai phase i think flexible thinking there's gonna be a real boo and so it's not a is agent or is not agent it's more like a spectrum on this realm of autonomy right i think with an assistant we use that term to just represent the interaction mode like it's primary job is to interact with a human when we talk about these kind of task agents their job is to knock out a task you know there's another we we talk about internally as like a customer agent their job is to like work with third parties not in the company external so you have different kind of rules of engagement with those all of those are kind of on different realms of autonomy but in general i think the easiest way that like i've been able to put it together and my brain is like the fact that it can have some autonomy agency is really where this comes from the fact that they can have some agency and begin to do work is when it starts on that spectrum of basically being an agent the debates we have internally is in this i don't wanna make any anybody's mind blow but it's like if you go and you ask chat breeze ant philanthropic whatever if you say make me a blog post there be a lot of people who will dial on the hill saying that's not an agent that's just a prompt i say okay great no rock with me for a minute mh so then i'm like okay if i ask only breeze can do this but if i say breeze make me a blog post and it now does the same thing the first one did but it also is context aware of all of your brand guidelines is that an need there's still some people who will ball saying no that's not need mh but then for the first time i'm like well think about it it's now made on brand content that's that's hard enough to even get like one of your staff members to do that right so didn't you get into if i say breeze make me a blog post on brand and generate some images for it this is why i think jet gp has become more flexible in their definition of it because now it's done the writing it's got now some brand application and it's done an image that's related to it that's like really three jobs that used to be a copywriter a media slash image you know layout person and then of course the editor who would put it on brandon so this is what makes it so hard is because right at what point does it become an agent and so there are a couple of camps one is if it's doing any work on its own so that that would be all the way back to just the prompt as the prompts get models get better like that will work mh some will basically go on the path it's an agent whenever it effectively has a goal can make tool selection on its own and has context that's like our agent leader internally right now that's his whole thing because like it has to have multiple tools available to it right right they can do that and so i think right now that's why people get fired up about it is because the floor is moving up the models are getting better and better like i don't know if when chad t five comes out but like there's some rumors that like it'll now just be a super omni model like it'll just be one you won't choose a reasoning model you won't choose a writing model you won't choose an image model it'll just be that so that model will become more agent in nature and then also over time you'll give them more and more tools that they can choose from so yeah anyways that's a hard debate it's a trick question for anybody that you get us good in this space but like that's just some of the flexible thinking i have along that spectrum yeah yeah i mean it's very similar to the concept of ag right like are we at ag already if we do it hit it will we know if we do with the goal post moves and else that it depends on like like you said the threshold like there are some people out there that are smarter than others we probably got some ag based on those standards we got right exactly exactly i mean if you go back to like you know alan turing or someone like that back from like the nineteen forties i think he would probably say like you guys have ag already if you were to ask somebody back then right yeah if you to look at me in high school how far i was in now do we had ag like i was that was definitely ai is definitely smarter than i was back then yeah yeah yeah i mean when it comes to agents the definition that i've heard quite frequently is that it needs to work through multiple tasks but then also it should have the ability to kind of like double check its work so if it notices that it made a mistake it should be able to realize it made a mistake and then go back and double check it and one thing that i found really interesting is that you can actually get models like o three and o three pro some of chat gb models to sort of act like agents right now with a single prompts absolutely i was giving it prompts yesterday where i was telling it to you know come up with five solutions for this problem and then analyze all five and then pick the best result from the five solutions that you came up with i mean that is going through and really thinking through all sorts of avenues and then deciding on the best avenue you can also tell it things like you know solve this problem for me and then once you're done solving it review everything you just wrote find any logical fallacies with what you just wrote and then rewrite what you just wrote to fix in the the logical fallacies that you just found right at that one prompt is acting like an agent for you because it's double checking its working and then fixing the double checking of it so it's like yeah i mean like you mentioned it's just a a broad spectrum i'd say like some people would consider what o three does where it we'll work through and pick tools to use to help you get to a solution pretty agent already totally yeah and then that's at the model layer you know now what people don't know is like in one chat thread you might ask it a reasoning question you might ask it a basic knowledge retrieval question some knowledge already in the l a fast answer you might ask it to do an image you you know that's why they call them the o models because they're omni models right in my opinion we're already way down the agent pass so i spend less time h over what's an agent nowadays and now more to your point i'm much more fascinated by the work that it can do and so for our ai now and we're still early days in this but like i want the assistant the breeze assistant its mission is to be like your executive assistant your chief of staff that's it yeah and so that's all we says over and so we're doing like really cool things like with that we're like it knows who you are memory and all stuff it basically knows your role but it's also working really hard to like understand your calendar it's working really hard to understand what task have been assigned to you and so that to me is like well we start to get to work there's that and then on the agents you know that we just talked about a minute ago like the account hand off or the the sales prospecting agents like it's something where there's a body of work that it is particularly good at and that's what we're doing so we're trying to just really relentlessly find those bodies of work those use cases hammer those out with agents and then i think what's coming next is i think everybody's gonna be overloaded with agents that's my next theory yeah yeah you know a hundred agents working for you sounds cool you're that billion dollar one person company but on the flip side like man what does it look like to manage a hundred agents who knows yeah that'll be interesting i mean i've always liked the analogy of like a conductor right like you've got like your symphony and you've got your flu player and you've got your violin and you've got that and then you've got the conductor sort of conducting the whole thing and i feel like we're slowly moving to the role where most people get to be that conduct and they get to conduct all sorts of different instruments and the different instruments know what they're supposed to be doing and they do that thing well that's right i think what really excites me about the future of ai is when ai gets a little bit more proactive i think there's a lot of people right now that they sort of avoid ai they don't want to use ai that you know the ethics behind how it was trained and you know all sorts of reasons right but i think once ai starts to get like more proactive where it's looking at your emails looking at your calendars and then like maybe one day you just get like a message on your phone that's saying hey nicholas is trying to schedule a a call with you i see that you have this date open on your calendar would you like me to respond to him and book it in for that time right like once agents start doing that i think the world is onboard i think it's like okay i literally have a personal assistant in my pocket now we're close yeah we're closer than people think on that we're spending a lot of time thinking about like the primitives or like the low level things so there's this concept that's come up of these things called ambient agents they are proactive like you said the concept there is like they're listening to all signals at all times and then they know what to do so we have a pretty robust automation platform at hubspot that kind of is baked into the whole platform so everything that we do at hubspot like one unified code base everything we do it translates to all of our product lines mh so we've got this automation layer that touches everything and so when we were building agents we originally started off with like give it a goal so just think of like a master prompt then people wanted to give it a lot of data in context so this was like how do you feed it a bunch of information then they wanted tool selection so kennedy email candidate web search can do that so we did that so that was working really good but then people are like how do i run this thing and you're like what do mean it's like right here you click run and they're like man i'm not gonna go into your agent page and click run every time i need it like this is supposed to do a job but it should do it at this moment right this is like a really good epiphany for oh my god think about it like this it does a job but all of us have a signal do it at this moment and so we basically pulled the automation platform into the agent builder so that for an example you know again go back something simple like account research the first time a form is filled out you can go do account research you know the account hand the next time it's marked as sold it can go do the account hand and then there are other events that we listen to so it's not just on objects changing status like events so like you can actually say when person visits this page and we know with our customers like when they visit so now whenever matt visits this page it can actually kick off a whole stream of like write an email matt so good where you been haven't talked to you in forever and like so those ambient agents are here now again i'll go back to how good these are is gonna be less and less about the ai and more and more about like who is the shift cooking the meal right now right right so i wanna chat a little bit about real world use cases like what are some things that like business owners listening obviously hubspot has a very large audience of small and medium sized businesses so that's primarily who's probably listening to this podcast what are some like real world use cases that people can do like right now today using ai to help their business whether it be you know other ai tools or hubspot internal tools just like what would be some of your best advice right now well first i do think leadership needs to make a decision what you'll find in a lot of companies is that they'll have you know some forward thinker innovator on their staff who's dabbling and playing with this but until leadership makes a decision that they're going to embrace this kind of future of work it's very hard for a company to move forward mh so that's first i think you should have top down buy in once you do that the very first step that needs is that i think every employee needs an assistant it can be a gemini eye if it comes with your google account it can be chat if you want or claude ant philanthropic good and then of course we have breeze for people who have hubspot and breeze again is just chat powered plus all the contacts mh but everybody needs that assistant and then to your point from that point forward they should start learning how to interact with those assistants so that's just like your regular old prompts like what's the weather tell me i wanna play my trip to bermuda you know can you help me brainstorm on this marketing idea about to contact this customer you know they they have these objections how would you handle it it's like your thought partner mh once you do that the next thing i encourage companies to do is a little strange but it's like get all of your unstructured data into a central system so if you're using a fathom a firefly hubspot has a note taker mh get everything all the meetings all that stuff recorded get it into a system where now the assistant has access to it the future agents have access to it that's really critical because there's like eighty percent of a company's knowledge is locked up in phone calls and meetings yep so now you've got every employee has an assistant you've got your data now unified with unstructured and that's the first time that you should now start to choose an agent i think you should start with an now of the box one first the one that i think it's gonna be most ubiquitous kinda like websites kinda came up in the like nineties like nineties and it was weird when you first had a website and then became weird if you didn't have a website i think right now it's like in that early phases is it's kinda past weird but it's like people are starting to have agents that man the frontline support and it's because they never sleep you know they're helpful they're not trying to rush you off the phone they're not trying to really like turn you into a sales lead you know this all these kind of per reverse incentives whenever you're having a tier one support but hiring a customer agent like for hubspot all you do is like point it to your website if you have a knowledge base you point it to your knowledge base if you have some files you wanna give it you do that and then that's it it's it's literally those three things like upload a few files pointed it to your website point to your knowledge base and then we're seeing that it can resolve you know roughly fifty five plus percent of resolutions mh that saves a ton of time for your employees to now work on other stuff escalation all that and it still has like a hand kind of thing as well it has a a hand off so if you say i'm not happy with you you know escalate me it'll do that the other thing that it can do is like start to find out if you're like wanna talk to a salesperson and you wanna route to sales so you can have a very like informal conversation about products and pricing and packaging all that stuff without feeling rushed and then of course when you're ready it can do that so that's the kind of the first agent that i recommend people do and then after that i recommend that they then begin to bring their people along on what i kinda call like ai fluency this goes in a bunch of different directions but like carve out time like one day a month for them to mess around with this stuff there's like educational stuff that hubspot provides like help them come along and you'll turn around twice and then in a year your people are fluent all of them have an assistant you've got an agent or more working for you and then the software that you use day in and day out if it has ai features like hubspot has like over a hundred ai feature just baked in they're not scared to click it we noticed a bunch of people like ai features are coming into all these software tools and a lot of people were like i don't know i don't know what it does i don't i don't know how much it cost you know like busy and so that's what we think like an ai leader looks like their data game on point their employees are on point their agents are working i think that's it my first million hosted by sam par and john murray is brought to you by the hubspot podcast network the audio destination for business professionals my first million features famous guests like alex rem sophia ama and hassan min h sharing their secrets for how they made their first million and how to apply their learnings to capitalize on today's business trends and opportunities they recently had a fascinating episode about how you can scale a profitable agency with zero employees using ai agents listen to my first million wherever you get your podcast yeah i like that a i think one of the things you mentioned that it's been sort of a game changer for me is i've started recording a lot of my meetings and if you actually have the chat gp desktop app they actually have like a record button now too where to record your whole meeting and then at the end of it give you a summary give you a need to do items so i just actually hit the little record button on my chat gp and i don't take handwritten notes i'm just like very focused engaged on the conversation when the meeting is over i go alright what did i say i was gonna do what did they tell me i needed to do what are they gonna do for me what my action items and in those threads a lot of people again this is like we're all pretty like new to this game but like those threads in chat gp are kinda of the saying like we learned it in like breeze it's not just that moment people chat with it right now but like let's say that like you and i are in a sales engagement we're hanging out we're talking you did that recording thing and then let's say i go cold it happens all the time like we don't talk for two or three weeks and i'm like can we get back together and and chatting in you're like what the hell like what do we talk about and you can go back and ask questions of that conversation yeah this is like a new muscle that like tools like granola if you haven't seen granola dot ai or you know like these are new muscles that we've never even had before which is wait a wait wait minute wait a minute it's not just a transcript it's not even just like some structured follow up or task out it's like i can go back and be like what did matt say about this how did he feel about that like that's crazy yeah it's well i mean you know i think back to you know whatever fifteen twenty years ago right everybody knew everybody's phone number right like i knew all of my neighbor's phone numbers my parents phone numbers like all of my best friend's phone numbers i had them all memorized today i don't even know my mom's phone number i couldn't tell you what it is i it's all saved in my phone now i don't need to remember that i've sort of outsourced that piece of my brain a little bit i know i'm not say people should like outsource like all of the conversations they have but it's almost like that same thing of like now we have the ability to remember like every conversation we've ever had in the same way that we have the ability to basically have storage for every phone number every contact every connection we've ever learned and i mean you could do it in the real world you've got things like the limitless pendant and the b that just got acquired by amazon and you know tools like that where you could just be walking it around in the world world keeping track of all your conversations as well and so it's really in an interesting world to be able to sort of like outsource our memory like i do have concerns that maybe people will outsource too much of their memory so i still think it's important to be very engaged in the conversations but it is very cool to be like i had a conversation with them two months ago what did we talk about again and then you know jump on two months later and but be like oh how was that birthday party you were going to the you know the day after we talked or whatever i know it's really cool that we have that sort of superpower now that we didn't have before you know it's funny the other thing i'll tell you like i think there's a lot of fascinating ant anthropology type things going on here so two months later you asked me how the birthday party is in today's world if i'm not super into ai i'm blown away i'm like oh my god it's amazing he remembered like what a thoughtful incredible person and like over in our lifetime our younger kids may be like in our lifetime there were people like that who were just like so like thoughtful i think now if you said that i would probably know that you had some trouble you know what's funny this is i think it's awesome i wouldn't care yeah the fact that you asked me about it is meaningful we talk about it we build a little rapport and i go back to again the delta that only a few extraordinary people would have remembered the birthday now it's something that we all have and so now you and i can have that moment let's say that if that's not your superpower to remember that stuff and so there's just a lot of like things where i think that like the threshold for humanity of the bar is gonna move up in a good way which i think is pretty exciting yeah for sure there there's something that you mentioned before we hit record you mentioned context engineering and i hate to admit it but that was the first time i've heard that term can you explain what that means what is context engineering yeah so the ai journey that a lot of you know us who are like in the ai industry maybe youtube matt it's like you learn about a prompt then you learn that like the human language has a massive you know capacity for like communicating information so everybody was really learning about prompt engineering and then there were some prompt style that came out there's a great youtube video on a prompt structure called craft mh c stands for context r stands for role a stands for the the kind action mh f is the format that you wanted in and t is like the tone and a can be for the action dress score or the audience anyways point being you used to be like a master prompting you would fill all that out but what has happened is that to really get good ai result it takes a lot of work to give the right amount of context so we started to see a body of customers that naturally on their own would have like the mega pdf file the mega pdf file would be about matt wolf about his podcast about what he cares about it's like this mega pdf out and then you would bring that into every ai interaction saying like hey i'm looking for an idea i'm looking for a script i'm looking for this and here's all the context well where things started to get wild is that there's a huge variance in the context that you bring to all of these ai things and so there's a lot of effort right now in this world of context engineering which is you know you're trying to make a podcast on hubspot you can make a lot of content on hubspot a lot of the content right now if you don't give it the right context even with great prompting it's not you it doesn't sound like you right doesn't know your history it doesn't understand all these things that you've done in terms of previous guests and so we started to see that like context is a layer above data so everybody used to think data was this moat data is getting easier and easier to get very hard especially like with customers wanting to own their data it's very hard to treat data as a moat the context is a way that you basically pull signals out of that data and the way we first stumbled across the first kind of context was companies using hubspot i kept saying like i don't wanna have to keep telling it about the products sell mh mh i don't wanna have to keep telling it about who our competitors are i don't wanna have to keep saying something like who's our ideal customer profile these are all reasonable things that a business would want a new employee to know this is what we do this is who we do it for this is why we're unique that's who we sell to and so we were like dang so people were now having to constantly copy and paste or upload that mega pdf file with that and we were like why don't we start to create all that context in a layer that they can do that and so we first came across like a brand context then people were like i wanted to know a lot about me i'm nicholas i'm a sv product i'm over ai these people report to me these are my peers this is like what i do memories about me so then we built out like a user context model you know now i wanna sell to matt matt's a prospect and so you know you go look at his linkedin page you go searches his blog you know you go find them on reddit it and so you then build out like a whole context profile on matt and that's a buyer context and on and on and on and on and so we started to find out all these context that really were better than people carrying these mega pdfs and you now bring those to the table so what's happening in a lot of the ai leadership stuff is that anybody who's working on ai solutions including i think open ai and ant i think this next year is gonna be all about context and memory i think people are gonna ob assess over that because it will ultimately make all of the agency all of the agents and things like that just so much better yeah i totally agree with that i mean i think the thing that excites me the most about these new models that keep coming out it's less about you know it got a ninety nine point three percent on this benchmark instead of a ninety nine point one percent isn't that amazing like those kinds of like benchmark leaps like they're they're impressive don't get me wrong like cool that you did that but from like a real world consumer that's using it all the time i love seeing the increasing token count like i love seeing that gemini now has one million tokens that i can use and supposedly they'll be making two million tokens available soon like that's what really really excites me because then you get these opportunities to really bake in even more and more and more context and you know before we hit record one of the things you mentioned is like what is def defense well like like you mentioned a lot of the data isn't def def almost anybody can acquire data at this point but like the context that you bake into the models that you're using is i guess what is a lot more def def these days i know for me like i have the youtube channel and i do a lot of like ad integrations i've actually over many many many many months dialed in the exact way i want my integrations to sound and all they wanna be worded and the exact sort of like structure of them and the beats that i wanna hit on them and how do i get them under ninety seconds and i've sort of trained over months exactly how those integrations should go now it's gotten to a point where a sponsor will give me an ad brief and say here's what we want you to talk about i literally copy and paste that brief into my sort of pre built i use chat projects for this so i go into that project and it just like spits out the perfect integration pretty much ninety nine percent of the time where i barely have to do any tweaking and nobody else has access to that specific context that i built up over time listen if somebody were to ask you to change now like you'll be like you'll pry it for my cold dead and and so i think that pick that maybe you do a particular out maybe over time you have some sort of quick hits that are a summary of this like you'll build up three or four or five of these type of you know they're like sub assistance or kind of you know a personal assistance you know customer assistants you'll build those they will have the context that you're looking for and it really does then become a virtuous loop the more you use it the more context it gets the more you dial it in and it's very very hard to then switch that's that's why it's great for the vendor very very hard for you to switch and so you know i was laughing if you were back you know starting a product if you ever started a company like one of the things you hear is like something you build has to be ten times better to get people to switch i don't know if you've ever heard that yeah heard that yeah the thing is is like in the world of ai right now a lot of people are switching without just like whatever the latest the newest you know those swap they're like very very easy come easy go but i am seeing as people get more and more of that context engineering down that that bar is now moving up to five times ten times and there is a world where it knows you so well that it might be harder than ten times better to get you to switch no absolutely i think another thing that's really interesting about the world we're going into now is that even software is kind of gotten to this point now where anybody can create like little tools that they need for their business i think we've entered this place now where if i have a little like bottleneck in my business that i just find myself this is something i do every day it's repetitive why do i keep doing this i can literally go to one of these tools have it generate a little script for me and just take that off my plate and i even feel like software and like writing little apps is no longer really that def def because i can go in find an app that i really like i've done this on past videos where i went and made a spotify clone just to see if i could right i went and made a a feed clone just to see if i could of some of these tools that i've used in my home life and now i don't even feel like having a software is even that def def so just well really interesting saying that i have a controversial take on yeah theory i do think you're right for some people it will be hard to be def divisible however this is like an interesting take the other way i think you have the same access as i do as others to go by paintbrush a canvas mh but what you paint what i paint and what a master painter paints is wildly different and so i think that the part that that i'm also you know somewhat att to is that there's no accounting for taste and this whole thank so i think there's another thing that's kinda fascinating happening in the world of ai tied to what you said which is you have a little problem there's no solution out there this is when i talk about the bar for humanity gets better like you solved it it it's awesome no harm no foul there's another one where there's a tool out there there's not a lot of competition maybe the price is too high for the value get this is gonna be one where you will wanna build something and that's gonna be a fascinating world for those type of things there is another world where it's just so easy you don't even look for an alternative you build it and i'm seeing this where people now love the tools they build like there's a a psychological thing called the ikea effect which is you built this table from ikea you love this table matt this is your table born of your own two hands your friend comes over and he sees a table with a crooked leg and it's chip this side of it but you love this table and so i think there's gonna be a lot of ikea effect happening right now where see that a lot of people are building these apps that they're very proud of that do the job by the way that is key it's like not a the project it does the job but it's also nowhere near commercial software it's not it's not great and so there'll be a whole world of these like personal projects that have been enabled but i still think there's no accounting for taste there will be room for some of these high quality high craft type tools so again the competition build out there gonna change quite a bit yeah no i i totally see that i mean most of the little apps that i've built are ugly as hell but solve a problem for me and i would never try to sell access to them they just play me with one thing you know well i know we're running out of time here but i i do wanna ask about like hubspot and like what are some of the like really cool ai features of hubspot that you're excited about right now and are there any sort of upcoming ai features that maybe you can tease a little bit anything exciting to talk about there i think i'm excited we re architect the breeze assistant to go a lot deeper with open ai and to bring a lot more of the hubspot context so for a salesperson person a marketer a service person i'm just pretty excited because i think i don't know how you feel matt but like there's a huge capacity for using ll out there like a lot of people i know they first start with one but then over time they're using complexity for search they're using right no google for search they're using chad gp for analysis claude for writing g for current events like there's like this ability to handle multiple in your life and so i think there was gonna be a world where there's gonna be work ll now some of the ones i just mentioned will try to go into work as well i'm excited for that like we'll have breeze assistant and i think it'll be great for the go to market professional a lot of the agents we put out were kinda dog water when we first put them out meaning like they they just were like very very immature the models weren't there yet when we first put them out the the orchestration around it but a couple of the ones that we put out are pretty darn good now so the customer assistant that tier one one is pretty good now and it is having meaningful impact to businesses and then the prospecting agent we just came out with the v two it's still in private beta but what's really good about it is that it does account research and it's really good using deep research so it does account research it then goes through and it's this is crazy it's actually deal aware so let me tell you what i mean by that if you put a contact in there that you work with it's watching the interactions and it is ready with the next communication that either you can send or it can send on your behalf and so you send it out and let's say the person replies let's say they get on the phone and you called them and let's say that you have a zoom meeting in the middle of all this right and then you then hang up it's ready for the next communication that's aware of everything that just happened so it's not like a automated sequence it's not that it's like it's crazy so it's like it's all you know i call you up we hang out we talk about three or four different things it's now ready for the next steps and those are plans so i think that's pretty good and then the last part is i think this next year a lot of people are gonna wanna curate their context so we're gonna expose a lot of that context for customers to look at and i think there will be a a rise of a new worker actually i think there will be an agent manager and an agent trainer sometimes they'll be the same smaller companies but in a larger company you can see an agent manager being somebody like the sales leader who says we're gonna use these agents for our sales org and then the trainer might be either the bd themselves the sdr themselves or maybe a team manager i think that's gonna be really wild because you'll start to see people go yeah i managed agents in my last job you know like oh my god you're hired so like i think awesome well i mean this has been an amazing conversation i'm really excited to see what hubspot rolls out in the ai realm and the agent realm coming up and you know thank you so much for hanging out and and sort of nerd out about this stuff with me today it's always a blast yeah thanks for having me appreciate it

everyone's vibe coating right now but why does every vibe coated app look absolutely the same in this episode design legend mu joins us to deliver a masterclass class on how to vibe code with actual taste we break down how to vibe code a website that not only looks great but most importantly to actually converts into real business if you care about standing out in the age of ai this is the episode you need and let's get right into it cutting yourselves like will half sounds pretty impossible but that's exactly what sandler training did with hubspot they used breeze hubspot ai tools to tailor every customer interaction without losing their personal touch and the results were pretty incredible click through rates jumped twenty five percent and qualified leads quadrupled and people spent three times longer on their landing pages go to hubspot dot com to see how breeze can help your business grow manga thanks for coming on well thanks for having me i'm really happy to talk more about what's going on with a world of ai and you know how we can make better prompts better designs especially better designs because that's what the world is needing right now yeah well at first i wanna i wanna to share this tweet which caught my attention about a month ago that you did it may realize you know because you're talking about how to prompt for ui you know and i'm always seeing people talking about vibe coding and one thing i've noticed is you know a lot of the stuff that people build when they're vibe coating it's either ugly or if it looks good maybe they use lovable or something it feels like all the designs look exactly the same to me and they're also probably not really like paying attention to how well these designs convert or anything like that i watched your video i was oh cool finally someone's like vibe coding but actually thinking about good design and so i thought it'd be awesome to have you on here today and to do kind of a you know masterclass class no pressure but like a master class on how to vibe code with good design yeah you know there's no pressure of course but as someone who's built i think i've done way more than ten thousand props at this point i've built really really complex apps including dream cut and or as you've seen mh and that prompt builder that you see is part of the feature of aura which allows you to essentially hand hold you in terms of you know how you wanna do your design okay and using the vocabulary around design because right now as you you've mentioned there's a problem with generic designs ai is not very good with design and has been like this for a while it's getting better but still even if it's getting better as with coding it still needs a lot of context it still needs a lot of these terms and these examples these images these you know animation terms that most people are not aware of which is why when you create an app using something like lovable or v zero is kinda look like ninety nine percent of the rest of the app is as if like you're using a template like bootstrap or right stuff like this which makes your website look more like a blank ui kit rather than something that a designer comes in and customize yeah they're not memorable you people can tell that it's it's template exactly and you know what is surprising is that it doesn't take that much effort more to get something that looks really basic into something that looks a lot more polish and sometimes it's just like a little ingredient that you added to the design like a little prompt or a little animation or a little resource that you add to your design to make it look just so much more sophisticated i think it'd be great just to like show people what we're talking about yeah we can talk about or but yeah know maybe it's better just to show people so they have like think understand what we're talking about yes so you know obviously i'm showing aura but i could have shown any of the neo lovable or habits to be an app that i built from scratch by the way using cursor and what do i try to focus here is if you've seen the age of dr the age of you know before that devi art i'm someone who came from that that background so i really care deeply about you know designs and effects and graphic design and print design so i try to make this as unique as possible so for example this one is quite popular and if you see the composition of a design like this right you know most people they would have a very basic design without you know some of the effects that makes this a lot more unique such as supply background so what is blind spine is a three d resource where you can essentially create or remix existing three d mh that you can put on your website so the the block that you see here is a three d resource which is again not that much more effort but you do have to go to sp design and then you go to community for example and then you can remake some of these designs that are made by the community oh i see so you can use that design and like probably modify the vectors or something like that and kinda customize how it looks yeah and also another one that i like is unicorn studio which is absolutely amazing again it's like an animated asset that you add to your website that elevate your design by ten that's so cool and you know that's kind of a those ingredients that you need to be aware of to make your design so much more unique and less generic so going back to aura so or i just wanna make sure i understand it's getting kind of like templates that you can get the prompts for or you're getting the entire template you i'm trying to make sure i understand yes that's a good question so aura you know there's a ton of templates that you can use and each of these designs that you see here are also code okay and the best part is that it's only one file and it's in html and css which everyone knows even if you're not a code i'm sure you've heard of html css right and essentially all of this doesn't depend on any libraries or react or it's not gonna give you an error you can just copy the code here and you can bring this code to any platform such as lovable or v zero and bolt and so at the basis if you don't wanna use the service you know obviously i'm gonna show the prompting aspect but you know what i wanna show is that this is how unique your design can be you have you know the menu the title description the buttons and all of that stuff and then you have the resource which makes you design stand out and then you have different types of design for mobile for animation for loading and so you bring it all together and you create something that looks a lot more unique and maybe a full page there's just so many ways to prompt to get there and as you mentioned before we do have a prom builder and this prompt builder is where everything comes together in a formula in a recipe if you're are a cook you understand that you have to follow a certain recipe to get the perfect omelet the perfect ramen so in this case it's the same way right like first of all you have a template right and this is the easiest way a templates is like a package of noodle or ramen you know it's gonna be good all the time but you know otherwise if you're more adventurous then at least this is gonna help you with telling ai how to do things and what to do and kind of giving the right direction i guess it gives you like the right visual language to their limbs so they know what to build for you right because like otherwise you're just like oh make me a beautiful side like what does that mean exactly yeah and a beautiful side can be a lot of things and one of the problems with beautiful side has a term is that beautiful is highly subjective and you know it's not very precise and oftentimes l m's are trained to be in a way that beautiful means purple gradient means you know like you know like ours the same super bold font and yeah you know it's very generic as a term so instead you wanna know okay what is a hero and what is a feature what is a onboarding and then you know is it a card is it a list is it a square is it a mental layout and the framing do you you wanna frame it full screen or you wanna you know create a presentation for youtube for example so if you do a presentation on twitter or youtube or instagram you might wanna have like a cart presentation or a browser presentation that gives a lot more context then you can add a background in in the back so you know that's for like the layout and then there's a styling you have multiple styling such as flat design outline minimalist glass ios material and then you know as you dive deeper into each of these categories you get into light mode or dark mode and then you have also the accent color so this is super important because every website has a brand color that differentiate from the rest so obviously if you don't say anything you just say beautiful it's always going to be purple i don't know why ai loves purple maybe because in recent years many designers are using purple so it's trained on that but unless you tell it it's always gonna go with purple so you might wanna you know tell oh maybe i want like a red or you you want amber or orange or cy or blue so there's that and then the background color again background color is the same as the primary color or the accent color but reduce in term of saturation and brightness as a result it's less you know attention seeking my first million hosted by sam par and john p is brought to you by the hubspot podcast network the audio destination for business professionals my first million features famous guest like alex p sophia ama and hassan min h sharing their secrets for how they made their first million and how to apply their learnings to capitalize on today's business trends and opportunities they recently had a fascinating episode about how you can scale a profitable agency with zero employees using ai agents listen to my first million or you get your podcast i think i'm getting a little bit into the details but that's got you know to just to explain so that you can get more value out of it so you don't have to do all of this but if you want to get into the weeds like really tweak and customize things you can exactly and then you get like the border color the shadow shadow works really well especially in light mode as you can see here you know you can have like no shadow small large and then you also have the beautiful shadow which a lot of people like to use and beautiful shadow typically you don't wanna have a border with a drop shadow like this one you would maybe also have inner shadows and you know maybe a three d bevel kind of thing which is also really popular on ios especially with a new liquid glass as you can see a lot of people love kind of like the more realistic looking and it's kind of like skew more ph is getting back into the game it's revived but in a way that is more modern in a way that is more contextual rather than having you know something that it always like wood or always like stitch letter and now we have a liquid glass that you know works well with any background any color so it's that's really cool and then obviously one of the most important parties is the typo and there's millions of type phases out there but i like to break it down to a few that most websites are using right now so we all know enter everyone use inter but you also have the sister phone which is a lot of people like to use guys and a man jak carter space on and again you have examples on the right side so you can see how it looks like it fits your design or not you know what is a sa font what is the san like a every font is more traditional it's gonna have more curves and mono space is more for coding condensed expanded you know and then you have the rounded which is more for playful such as for kids app and then you have the handwritten if you wanna be more like a school or something like politics yeah yeah yeah exactly but right now i would say focus on sounds on ser and maybe sometimes rounded and then for each one you have a list of fonts that you might wanna use such as play fair or instrument which is really really popular right now those are the main ones that i would use and then you have like the font pairing what is font pairing essentially is okay you have the title which is one font and then you have the body which is a different font typically a title is gonna be more like big you know bold and all that stuff and sometimes you might have like ser as a title but the body font usually is not ser or at least normally because it it needs to be readable most people are gonna spend most of their time reading this so it has to be sound as much as possible so that it's more readable and then you have the sizing the subtitle the body text if you really wanna break it down but also you have the weight of the font so especially really important that sets the personality of your website is the heading so do you want something that is super bold or you want something that is light right the light seems a little bit more sophisticated a little bit more modern and then bold is a bit more like in your face you know it needs to be readable and all that stuff and then a lot of people need to be aware that title fonts because it's so big they can be much more closer together in terms of letter spacing right seeing traits like ke and stuff probably exactly exactly yeah and so yeah typo that's kind of like the gist of it a little crash course into type poverty and then you have the animation so you know most apps yeah let you can see the preview of everything like versus just going to to v zero and it it gives you something looks cool counts like a versa website or whatever you can actually like kinda customize things and get it how you want exactly and you know this by the way you can just use it for free anyone can go to this ui you know all of these you can just note that you can write them manually but right now the whole system yes you do need to have an account you do need a pro if you wanna add the whole prompt right if you add the whole prompt has all of this but clicking here is free everyone can just reference this see what's going on what are the terms all that stuff you don't need the you know any account for that and then you have the animation so fade slide scale all that stuff you can also select multiple you have the sequence animation right and then you have the word by word letter by letter right when you're doing these animations in the prompt is it going to tell v zero or lovable or whatever which library to use or or something like that no because animation doesn't require library okay the only reason you're are using a library is to make coding easier mh but javascript in itself and css itself only has these animation possible right it's possible it's just that it requires more code but the library sometimes will simplify like frame emotion for example is a really popular library it makes the animation and the coding for that animation much simpler and it allows for some tweaks that are maybe hard to be possible it's possible to do with coding and so that the l can do it themselves yeah yeah because if you think about it all the libraries are all using html css in javascript right they're all based on the same framework it's just that their packaged into react but react you know you create a react website you export it and you publish it in essence you inspect the browser it's all javascript and and html right right so the same way with animation all of this is just animation using css and html and javascript okay let's be honest your ai prompts aren't giving you the results you deserve but with a little coaching you can transform from basic prompts to engineering conversations that get you exactly what you want from chad that's what this playbook delivers not just random prompts but a step by step system with the exact techniques top ai professionals use every day you'll have your own personalized prompt engineering system that delivers consistent results get it right now scan a qr code or click the link in the description below now let's get back to the show i think it'd be good if we could do something like make this practical for people like if we can actually come up with a website landing page you know we can even make up a company or whatever but like a make a landing page right now like whether we use v zero level whatever you wanna use and if we can kinda like tweak the design and then show how we're you know actually making something absolutely one of the most important component that anyone's gonna build is the hero yeah the hero is essentially the first thing you're gonna see when you land on a website you don't really need the context where hero because the hero already has a a a much of a context it's gonna be full screen i like to adapt from a template the reason why it's really important to adapt from a template is because it's going to have so much more context on top of it but it's still gonna be a hero and then you know for the primary color for example a lot of people love dark mode right now yeah you don't really need an accent color if you're using dark mode because white and black are good enough in term of tie face definitely some inter is really popular and big titles subheadings i'm gonna make it a little bit smaller more minimal which is you know a style that i like and then we have you know letter spacing tight and all that stuff in this do we actually change what the text says or would we do that separately you know by ourselves or with an l or right so all of this is just the visual language it's not gonna tell you what the app is about which is why we have a custom prompt right here mh where you're gonna say something something like create a hero section for a saas app i was thinking maybe we could play around with my my website lord dot com you know it's currently an ai newsletter sure you can create a full page of course it's just gonna take longer because the more that you create the longer it's gonna take for ai to channel can just make like the hero or something maybe like a hero yeah so hero section for a saas app for laura dot com so that's kind of like the basic of it i didn't need to mention it but it's good to say it and then what is it about the name of the website and what is this website about so this is going to determine what kind of content is going to show yeah let's give it a try and hope for the best so we have all these things we have the context which is the html from the template i'm gonna use cloud four and it's going to start generating and you typically go to just like claude first is that is that and then use like the artifact to see like a preview right so while that problem is cooking let me show you some examples of ui that are created with claude and what i would say is that claude is my recommended model for few reasons one is that it's the most creative one versus four point one is more reliable but it also generate mostly the same result it is fascinating that o three to seems smarter but also claude still seems more art articulate and better design stuff for some reason it's so interesting all the time o three is the best at vision so o three is smarter but doesn't necessarily necessarily create better code or better html but right now i'm using cloud four now one thing that is really important to me is how easy it is to customize after so one feature that just launched is to be able to quickly change the fonts so you can go to fonts right here and you can change to different fonts just the same way as the prom builder you can just change to you know play fair or instrument or jak carter but you also have inter and then you can sort of like play with the bold of the fonts so light or you want bold or you want it metallic so you also have the tightness and one thing that is cool is to be able to see what is going to apply to so you you see like these green outlines and also for the body text so you can have like a bold body text or a medium body text which is really popular or regular body text and then because we're using tailwind code it's really easy to use the code and transfer between colors and boom you have dark mode that's cool you know traditionally if you wanna have dark mode you if you've done fig more for example you have to set up the variables you have to set up like the light and dark colors but in this case is generated by ai and i was just able to just transfer the colors between light and dark just like that and the other thing that i love you can see here the neutral color right you can change the neutral to be like blue ish which you can see now it's a little bit more blue or you can choose one of the colors that is setting for your brand so for example if your brand is more red or orange you can see that you can really play with these tones of colors and this is thanks to the ability to just take the code and cross match and then just replace all the tone of tailwind colors awesome and back to our prompt is just finished and we have this beautiful blob in the background which is from the template that we have included yep and also it's in dark mode it's about data centers it's about lord dot com is using the same layout that we did and also applying the fonts that we specify as well as the beautiful animation right here and of course you can customize this by using design mode and code mode if you're are more familiar and you can change the text change the styles or change all other colors for example light and dark mode as you wish you know obviously anyone could use this but it feels like for designers you know this is like a a superpower and i i was sit here think about it the fact that you know you've created this in a few other apps now i just wonder you know where do you see the future going like our our designers now going to be like amazing at creating companies or or know how do you see things playing out i think one of the key differentiator for designers is going to be taste but not just for designers but for apps out there if you think about it right the more companies that you have the more apps that you have you know there needs to be a way for an an app to stand out and the way it's going to stand up is how good it is and how tasteful it is is the same with restaurants if you have too many mcdonald's then mcdonald's is no longer special so you know now like in japan for example they do it really well where the food is amazing and it's very specialized and so that's how i see the future of ai is that you know the key differentiator is going to be designed because every website is gonna look the same every app is going to function on the same so it requires someone to think through the problem to really solve these problems for themselves as a customer and to use it on a daily basis and then with pace and knowledge they would create something that is more unique and that is different from the rest and you know as a human we always have millions of special things that we want to tweak and you know now that it's so easy to create your own app to design your own app just do it right like just do it based on what you want and what your tastes are and i think that's the future of ai i totally agree it's like the best time ever to be a creator that i've ever experienced in my entire life so it's exciting so like manga it's been awesome having you on here and just maybe tell people where they can find you online or or or you can find me on twitter at man i do tweet a lot about prompting about you know the whole education around design code i've been doing this for a while traditionally it was before vibe coding but nowadays everyone's vibe coding everyone's wife designing and so i tweet a lot more about the education around it i have made several hours of content around you know how to create the best looking ones and you know give hour a try even if you're not gonna use the app you can still browse all the templates and you can find the ui that you want just copy the code or get the animation that you want and all of this is free right you don't need to create an account you don't need to pay for anything and the code is usable anywhere it's super easy it doesn't have five or ten components that you have to copy one by one and then bring it to lovable or v zero you could copy the anywhere right could use it with cursor whatever you little factory something else yeah and that's the beauty of ai nowadays as a language is that you know you can bring anything that is a word even image you can attach an image you can attach a fig url you can also use the prom builder and all of these are techniques that's going to really elevate and the same with the html code that i just copied i just paste here and say create a website based on this right so if i do that then you know ai has so much context it has so much taste that's already embedded into it right and you're gonna create the best looking one that's awesome manga been awesome appreciate you coming on really appreciative of the time that you spend talking about this kind of stuff i think it's really important and i hope these techniques going to be helpful to your audience thank you thank you