Opportunities in AI You Can't Miss this 2024: Insights from Founders

welcome to the Cav mines [Music] podcast now what are you what are you paying attention to when it comes to to AI I'm trying to it's tough you know I've never I've been in tech for 12 years and I've never seen anything move this quickly uh and if you want to pay attention to everything you get no work done so we try and we try and balance like paying attention and building what we set out to build um but I'm like today Gemini Gemini came out and one of the things I'm very excited about with it is that they are multimodal first and I personally believe that multimodal is the future now I don't know if it'll be multimodal models or if it'll be single modality models that some framework Architects together that's a big question in my mind but it's one that I pay a lot of attention to because it will affect the tooling and what the customers inter faes are a lot so as someone building tools like I would like to answer that question as quickly as possible so that I can optimize in the right direction um this is this has been an open question Forever by the way like even at Cruise like there was teams trying a one big model that did all the modalities versus like all of the different models like an image model and a light out model and a radar model um and typically the single modality one's one up until now but everybody knows that maybe in the future one big multimodal model could be better and this is why Gemini particularly interesting because they've Google is clearly taking that bet and they're leaning in that direction they want that to be the way and if they're right um it'll be very exciting what are the missed business opportunities or uh problems within AI right now that you see are unsolved you know maybe it's something that you're just like why is this why does this not exist yet I don't know if it's a Miss it's if it's an unsolved problem but to me it's like it comes back to that gap between uh what AI can do today and what we see it do in production today like AI can do so much today but it's just not in production yet and because of that fear fear driven stuff that we talked about earlier and so my biggest frustration is that like if we had if most of the companies out there think the big companies Nike Disney like the consu big consumer companies if they were actually putting AI in their apps like it could be so cool but it's not there because they're lagging big Enterprises are Lag lagging behind in terms of what they can ship um and so that to me is the biggest opportunity um there's what what small startups and what like top Tech startups can do in the valley is exciting but like the real world is comprised of a lot of much larger companies that move slower um and they're just like much further behind than I wish they were like as a consumer my life would be so much better if AI was embedded in more places yeah that's a good point what what do you think Founders are missing or entrepr are missing when it comes to AI based on all your conversations that you're having my help take is I don't think they're missing much I as I've never seen as many people found companies in a space as I'm seeing in AI I was in the YC batch the amount of AI companies is insane both in our batch and in previous batches pivoting into AI um I think it's just it's fiercely competitive on all layers of the slack like consumer layer application layer Dev tool layer infr layer deep more or less deep like all of the layers have competition um I think that it's unclear who will win but everybody knows there's a lot of value on the other side and like competition is great I'm a big fan of competition um and so yeah I I don't know that there's much just being unexplored that the one thing I thought was being a little bit unexplored was multimodality like there weren't that many Frameworks to try to tackle it uh but I think Google changed that today I think we're going to see a a big flip in the next few weeks and we're going to see people building multimodal Frameworks uh I'm curious what those look like because we've thought a bit about it at a so it's PR pretty hard problem and then what what are you paying attention to when it comes to AI like what's what's exciting for you right now yeah I don't get particularly excited about like the I I don't really call them rappers myself because I know there's like that that point that people always make like look at Expedia right like Expedia was ultimately a rapper and it's now like a massive company that's much larger than like the data source provider that they ultimately built around so I'm not one of those like Skeptics that say like you know AI rapper companies can't work um I'm not but like what gets me like most excited about are the vertical use cases for AI and like industries that um you maybe wouldn't think of um so a couple like examples come to mind out of like companies I've invested in uh one is like using AI um within their software Tool Set uh they're essentially building like a uh a workflow and uh process management software for uh companies in regulated Industries so pharmaceutical manufacturers manufacturers of like highly regulated food or or beverage products um and that gets me excited also like I I recently invested in a company that's that's using AI to conduct uh user interviews um which for the last 75 years has been an incredibly like intensive like human run process um so I I get excited about like the very vertical use cases for AI and I it's not that I'm not excited by the rappers but I I I would be worried about there is just like um they'll be disrupted quite quickly there's maybe not so much of a Moe and it maybe becomes more of like a go to market and branding play and that's probably how Expedia actually won um that's my two cents that's an interesting story I didn't know that about Expedia um and then I think there was a lot of other businesses that are now massive that started out as like essentially built over um uh an API or some kind of database or web hooks or something and then grew in scale and became a leader um over the period of years and then what what are you paying attention to personally when it comes to to AI like what are some things that excite you and interest you yeah uh I mean it's moving at like a break neck speed which is really really exciting um you know I feel like every month stuff is becoming obsolete so for me things like being able to like um you know process synthesize images for example with open AI new release that's really really interesting it's uh going to give our model like access to context that it wouldn't have had before um there's a lot of startups uh coming up uh as well that are trying to compete with some of the bigger players that's pretty interesting um I I don't know I don't know if there's like a specific thing that I'm like looking forward to or like being on the hunt for but just generally keeping a pulse uh on the greater AI ecosystem and seeing what people end up building Beyond just rappers which are are very much not defensible what are some missed business opportunities or problems that can be solved with AI right now that you've just seen because you've talked to a lot of other Founders you're in space yeah what's some interesting ones interesting one is honestly like legal for example uh I'm sure there are people working on this but like legal stuff is extremely expensive um when you deal with contracts you deal with agreements you deal with documents it's like overtly complicated the language that they use and sometimes I want to look at a document and have a co-pilot that basically explains to me like I'm five like hey read this document and tell me what it means in plain English like I I'm sure some tool exists out there that does this maybe maybe it doesn't um but like that's one area where I think there's a very clear um uh opportunity space um another one is this um like chat Bots and help docs where you know having someone be able to ask questions and be directed uh exactly to the right resource to the right interactive demo right interactive guide um I know there are companies doing this I just haven't seen it as much in like in the wild using the software that I have um and then another area is honestly like therapy um I think therapy there's a lot of applications of of generative AI because like we as humans sometimes don't want to offload our baggage or we're embarrassed by telling like other humans about what actually frightens us um and I think people would be actually more open talking to like an AI or a robot because they can fully like Express and tell their whole truths without being judged right so if you can have a model that is specifically trained on you and your past and it remembers um you know like I think pine cone was like uh one technology that allows your your model to have like retention of memory um that would be awesome and it could like live on your phone or live on your device so that there's no like security of like hosting your past data up in the cloud um and just being able to talk to your like AI uh therapist because I know that is also something that is not very accessible to people because it's so expensive right so what if you can make that democratized and accessible to people that um that don't have access to those types of benefits or that kind of money what what are you pay attention to when it comes to AI like what what are some cool thing cool uh why common a or non-wide Comm air companies you've seen um yeah what's interesting for one directly related to our business we're always paying attention to ways to save money like if there's a new model that was just created that um that that like for instance a three billion model um three billion parameter model uh we we like we're going to evaluate that because if we can charge our customers half as much and make the same margin then of course we're going to do that um uh but like Beyond just related to our business what's interesting to me in in AI is um the integration of uh of Agents with useful workflows like I said one of our first uh things that we built out was taxi and we saw the potential but we also thought that it would be a while before um it would be ready for the market um so hearing from some of our agent friends uh what they've been able to build uh is very exciting and then on the flip side uh hearing about employee dislocations uh is also interesting and a little frightening um one of my friends just told me that uh using his platform someone was able to uh someone was able to build an agent that turned their back office from 30 Financial analysts to Three Financial analysts so that's a 90% layoff uh which um which which I suppose from an economic perspective I'm I'm just an armchair econom Economist here right but um like that's good for the economy because in theory those 27 people can find other jobs that also produce value the same value that were initially producing is now being produced at a much lower cost so prices go down um total production goes up but um that's terrifying if you are very comfortable at your job and you want to keep working at it for a while um it's unlikely that the sorts of White Collar jobs that many people um a lot of whom I know uh in America America and around the world uh are comfortable with right now are going to exist in the same form that they do right now a few years in the future that just it it seems very unlikely that we'll have people who spend most of their time typing and clicking across different web interfaces because um because there are agents who are being built right now specifically to automate that process and and just take like normal text instructions to complete the same thing um and like I said when we built taxi that was more expensive than person but with these fine tune models taxi is now significantly less expensive than a person that people don't realize yet but they will very soon is that data analysis is now incredibly cheap uh my my co-founder just analyzed all Hacker News posts I don't know how familiar you are with Hacker News but news form run by y combinator um my co-founder just analyzed all Hacker News posts across a few dimensions for like the last 10 years um and it cost him a few hours with with a couple gpus with these fine tune models so um like there's a lot of data analysis that hasn't been done yet that is now incredibly cheap and is interesting and potentially useful to uh to like business use cases so um yeah I think I think we'll see a huge uptick in data analysis uh as as people realize they can use fine tune models to do that um as far as specific businesses that can be built from that I don't have a great answer um uh but I I think that there are a ton that came up from there in a world that's like increasingly leaning on AI like are you taking a stance that's more AI plus human or AI minus human and do you think that it's going to grow increasingly more AI minus human or AI plus man we could do a whole podcast on just that question or that series of questions um right now I think I'll just share a couple of like philosophical thoughts on AI that hopefully answers all those questions I think that as a species we know very very little about intelligence just like it's a word that we all stuff a different meaning into and so because we don't know a lot about what that actually means what intelligence actually means um I think what AI is going to do is help us understand our own unique intelligence even better so I think anybody has like a really strong conviction on oh it's definitely going to be AI or it's definitely going to be human I don't think they have a full appreciation for what what what the different kinds of intelligence are that we're going to see emerge right you know like like we have an intelligence called adaptive learning it's intelligent at knowing like what what how to make how to curate content that's going to make the next newsletter better you know and gong has one that helps you do for sales calls but like there's it's not just AI artificial intelligence all these sub sections of what intelligence looks like and I think the more we see of those the better it will reflect back to us what we are that it is not and inside of that I think we'll have a deeper appreciation of what our intelligence is actually capable of which is dude extraordinary things so generally speaking I'm AI plus human right now I think there's been enough use cases across over long enough time to show that the AI plus human use case is still still the winner um also just you know following Moors law and Technology adoption like techn technology enablement um yeah it's like it seems inevitable that like computational intelligence from a human perspective is going to be pretty much made obsolete just like if all I did was took computational intelligence out of what we do like AI is already better than us at that not totally like there's on the extreme ends there's always exceptions but that's what that's what I'm saying is that like what's beautiful about human beings man is like we do it all like I I I can use my my EQ to feel what you're feeling and I can use my computational intelligence to make an assessment of what that means and then I can wrap that all up into something we call intuition to make you know uh deep deeply accurate decisions that shouldn't know but I do you there's so many things that our intelligence is it's a it's a combination of so many kinds of intelligence that I don't think we have a deep enough appreciation for it um so yeah I I'm I'm generally AI plus human so I think people are not you know sometimes when I talk to our customers or potential customers sometimes they're just not ready to you know adopt to the new world or oh I think AI tools cool but it doesn't really capture what human does right so I think you know um understanding what they want to see and uh so that we can ultimately bring that value to them is important and also but at the end even even though he saw everything that AI does and it really makes sense for the team just kind of afraid to you know um because I don't really understand what goes behind AI right so I'm not ready to adopt this yet right um I think those are the things that you know we need to work on I think we need to go from step by step um gradually make the transition easier and also you know um let them see what goes behind actually AI right um because when they see it it's like oh AI is really cool it spits out all these different metrics and all these you know insights that I I had no idea it's about but is it correct is it right you know they always GNA have the doubt and skepticism in their mind because AI is you know something is it's it's not something that they've been working with right for the past I don't know some rapid fire things actually so what are you paying attention to when it comes to AI what what gets you excited is a slightly copy answer but the answer is sort of everything in the sense that certainly whenever something new drops in an open AI update email you just know there's going to be this huge uh Rush of kind of people trying to get that first M advantage and being the first to Market with either a new product or if it's kind of related to to seing content I i' being the first one to kind of generate that feature um with what's now possible um um and I think if I if I didn't have by word and I I didn't kind of have any product or anything my strategy would probably be just kind of refreshing my email waiting for for new models to come out and just trying to be ready and to be the first one to build something that that kind of takes advantage uh of the gap of what's now possible through apis versus kind of what what users can do easily and trying to kind of Arbitrage that essentially would be the you know aside from UI what do you think could be the next let's quote unquote say like most impactful AI business that no one is building because you've seen a lot so what would you say would be something that you're like oh okay I I think the businesses is a is the big sort of question mark word there if we can take that out and forget the business part of it cuz there's a bunch of business dynamics that might prevent extraordinarily powerful things you know transformative things to occur and that could radically change the world and by the way then change how you would even build businesses like what does it mean to build build a business as we were talking early in the podcast right they said that Venture capitalists are struggling with sort of the pace of disruption uh and I think that's only going to accelerate and so how do you build a business what is a business you certainly don't write a business plan do you have time to incorporate no you don't I meaning you should but you do that later uh the first thing you do is you throw a lot of stuff at the wall and see what sticks and once it sticks that's when you build a business around and so like finding product Market fit becomes something you do multiple times a day on different things until you find it and then you very quickly go do it and by the way you'll probably get disrupted shortly thereafter because other people can do that too once they have the idea they'll do that too and that's going to be really hard to build businesses in general in that kind of an environment yeah so well if we take that off the table then then it becomes arguably even a more difficult question because everywhere you look their AI is going to create transformative like m massively transformative change look look at medicine people radically transform medicine access to diagnosis uh uh ability for personalized treatment plans uh do things proactively so you doing research while we sleep at wire speed right and that's by the way only accelerating and and so uh whatever that's medicine you can pick anything else but I'll point you to one that we talked about earlier which is um the Creator economy is about to be destroyed I believe as an economy now this is going to be provocative and because I built a company in the Creator economy space it becomes even more provocative and I'm not trying to be provocative I just want to point to something that we think is obvious which is humans are obviously creative beings and so we will never stop creating for the sake of creating and using our Creations as a means of communicating with one another self-expression Etc check but making that into a business and monetizing it when you make something into a business you start to deal with the cost of goods how much time does it take me to produce a piece of content versus how much money can I get from producing that content as soon as you look at that equation it becomes very clear that there's no need mean no possibility for humans to compete with automated content creation personalized automated content creation so we already see automated YouTube channels mostly the stuff they're producing is kind of junk soon it won't be junk it'll be as compelling as anything produced by a human but it'll be produced on demand for each of us individually and that's obviously going to be radically more compelling than anything else and so I think the only YouTubers in a sense that might get a pass for a little while are the YouTubers where the content is about them the human not about anything else because that goes back to that thing we were just talking about where I don't want to talk to your AI I want to talk to you I don't want to see your AI generated you know Vlog videos I want to see you okay great then you're safe but for everything else I don't care if it's you who created it or not I care about the topic and there that topic is going to be again because it's personalized it's going to be so much better than anything you could possibly afford to produce and that's going to be very very disruptive that's true cuz then it's well it's self-explanatory like it's just personalized for me I don't need to search for anything and that's the ultimate yeah personalization AI interesting okay and proactive ai ai we shouldn't be asking AI what what to do generally again this is the dark agees another way of looking I've tried to explain to some people like if you met a genie and the genie said I can I'll answer any of your questions what do you want to know and some people say What's the meaning of of life and all that that's great but I say that my question would be what question should I ask what should I know that I don't know what's the most important thing to know hey thanks for listening to today's episode of The Cave mines podcast I really appreciate you for taking the time to listen to the AI insights the golden nuggets the value and I hope you got a lot of that from this conversation that we had we're going to be interviewing a lot more amazing AI Founders startups experts bringing you these deeper insights so that you can apply to your company and take it to the next level and adapt it for the age of AI if you like this episode please make sure to subscribe download this episode tell all your founder colleagues and friends about it we do this all for free you're getting every single thing with the Deep Dives the podcast webar workshops all in an effort to adapt our businesses and help us grow together as part of the community last thing I'd like to ask please leave a review if you can something that's one two sentences it helps us rank higher so that more people can find out about us and then it also means a lot when you're able to take about 5 10 seconds to leave a nice two to three sentence review for us as well I look forward to seeing you in the next podcast episode and where we're going to keep making things even better more valuable for you thanks again and I'll see you in the next one thank you for listening today's episode make sure you subscribe to the Cav mines podcast the only podcast Founders and CEOs need to integrate and leverage AI into their businesses every week we sift through the buzz and serve you crystal clear insights that you could plug into your business today visit our website cavs.com for show details and many more products this show is copyrighted by caves written admission must be granted before syndication or rebroadcasting

2024-01-03 00:50

Other news