How Web3 startups can succeed as AI and blockchain merge

0:00I'm Ashen Addison from the cryptocoin show and today on blockchain interviews we have naven agnot CEO and co-founder of Cumberland Labs navine welcome to the show and thank you for taking the time thank you so much for having me you're very welcome excited to dive into your expertise in incubating web 3 startups what makes a good web 3 company uh a little bit about AI as well which I know

0:23you have a background in and and speaking of that let's just before we start into uh current events I'd love to hear a little bit on your background blockchain business Ai and then we'll dive into everything Cumberland labs and and blockchain absolutely um so I have a varied background I um have a PhD in uh computational neuroscience and AI from Columbia University after which I was a

0:46scientist at MIT I left MIT in 2005 to create uh an AI startup uh that really became a series of AI startups where basically the idea was here's an amazing technology uh can we uh apply it to the public domain can we do some good with it can we apply and solve real world problems using AI right um so the the one of those companies actually um became big it called lendo where we

1:14it was a fintech company uh where we were using AI to help assess credit risk for young people um and that sort of brought me closer to um fintech and closer to sort of the technology um that was around that would ultimately become blockchain um that where I know I I bought my first Bitcoin in 2009 actually 2010 and sold it as well U but but the

1:44the um the the point it was fascinating to see blockchain as a technology grow up uh through the 2010s um and we finally began to dabble in it um at my company um in 2018 2019 um and uh I I exited out of that startup and um connected with Don Wilson uh at drw and together we began this uh incubator Cumberland Labs where we said

2:11you know uh it's an amazing technology uh what interesting things can we make with it um and let's start an incubator where we do that so that's really what Cumberland Labs is it's the Genesis isn't um taking technologies that are sort of very very Cutting Edge um uh blockchain AI and solving a real world problem that's really what the focus is on wow very cool and it sounds like

2:37you've got into AI pretty early I feel like we're going to circle back into that because one of the use cases and narratives of blockchain at least in 2023 has been finally I think people the light turned on with with chat GPT sort of realizing hey you know this is a technology that's a little bit ready for more mainstream adoption and also can be potentially combined with blockchain for

2:59even more benefit but there are you know since you got involved and and since 2018 there's been uh an expansion of different use cases and different industries that are now able to start understand the potential of blockchain and start integrating into those Industries and with you guys incubating all these different projects you must be well aware of all the

3:19different Industries and you know what are sort of the barriers to entry what are the first use cases that will actually penetrate into the market and where blockchain is sort of starting first um so yes absolutely I have I wrote my first neural network in 1992 yeah so I've been doing this for a very very long time um before a lot of your listeners were probably born yeah uh so

3:43the the um um use cases of AI I think and broadly of Technology have always been tricky to identify early on right I was around in the '90s and watching sort of the internet uh come up of of age and looking at the internet and I remember being part of conversations in the '90s with people where people were like so what can we actually do with the internet and honestly I didn't have very

4:08good answers because um you know online shopping was sort of a horrible thing in the 1990s you couldn't barely get online um you know you had you had this modem that you had to uh log in on and it was just really cumbersome really horrible right um there was really really there really was not that much good you could do with the internet in 1996 right or

4:311998 um so and it you really couldn't imagine a service like Uber or servers like Facebook or any or Airbnb right you uh the advantages of the internet came much later and it's difficult to predict these things right so something similar with blockchain as well you can see the potential of the technology you can see the potential of creating um a way in

4:54which you can uh publicly identify a fact right I can publicly identify the fact that um Ashton owns um this thing whatever this thing is um and there's all kinds of benefits to it but it takes a while for the benefits to permeate and for us to be able to create products that are widely useful and people are comfortable um using right because onboarding is an issue there's all kinds

5:19of um things that we have to tackle with blockchain similar by the way to um how we tackle it with with uh the internet MH yeah that's that's a great point and I feel like you're probably seeing some synonyms of of that internet days when it's like well what what are the use cases you know there's a lot of things that people are trying um but it's going

5:40to take a few years for them to build up many will die and then some will sort of prevail and then in hindsight people will be like of course those were the main use cases of of of blockchain um so being in Cumberland labs and seeing a lot of web 3 startups in their incubation phase what are some of those synonyms that you're seeing or what are some of the things that you think might

6:03stick as we move into the later years of blockchain um so I think um there's several different types of um things that are happening in the world of blockchain I mean early on we got into these um this phase of um you know these icos and these uh token offerings and companies using tokens basically to sell um I I don't want to say equity but something that akin to equity um it's

6:32almost like a Futures Contract on the company right where um your your token will get be more valuable if the company does well or the protocol does well or something like that right so that that is uh simultaneously a good and a bad um it's good because it allows you to access liquidity as a Founder um it's not good because that liquidity is actually very expensive because um if

6:55you think about it um you know you're typical investor your typical VC or somebody like that when they when a VC invests in your company um they're they have a four or eighty year outlook on that investment right they know that it takes that long for any young company any young product to mature to become profitable to have um you know some kind of uh good outcome

7:18right um whereas um your typical uh guy who's buying your token uh you know Joe Mo in Idaho or Indonesia or wherever um they don't have any four-year Outlook right they have a two-month Outlook they're like dude this is it going to the moon or I'm selling right so as a result there your your token is actually extremely volatile right so the majority

7:44of tokens out there um you know for typical projects are extremely volatile and companies are not set up to deal with that right um your typical founder you know lots of things are going wrong if if you're a Founder I I can tell you all about the things that go wrong at startups I'm sure you're familiar with some of those yourself um fires are burning every day as you go home and you

8:06can't deal with everything now you also have a token that is just going up and down doing all kinds of crazy things right so uh it's it's it's it's tough um it's it's a you know yes you get liquidity you get access to liquidity um but it's a it's it's high it comes with the price um and it's a price that not all founders I think are prepared for you know so that that whole way of using

8:30blockchain to create tokens I think um we're going to be defining this in the next um in sort of the the next wave of um of this thing um the of of of crypto adoption right um where adoption is going to come from is um we expect it to come from uh solving typical use cases right um it could be as simple as money transfer or

8:58payments it could be other interesting things like um establishing facts right like the there are these um proof of attendance protocols and other kinds of things where you want to prove something that's online um with Trust app which is a project that we are very proud of that we just started um we are actually uh taking sort of tackling misinformation

9:21on the internet right so this is our counter to the problems exacerbated by AI which is people say these things and you have no idea is this real is it not not real um did they make it up are they lying are they misrepresenting some facts like what's going on so our approach there is we um use blockchain to create permanent pieces of content that are publicly accessible um and then

9:45we give you a little tool it's a browser extension it's a little tool that'll put markers next to your social media post saying sort of red green or or gray which means we don't know right and the the biggest um the biggest idea here is don't trust us right um You don't have to believe our rating um every time we rate something or we we we uh put a marker next to something uh you can

10:09actually go check the nft yourself you can actually check that the fact that um uh yes this is real the source is real it's publicly available um you don't have to trust us right so it's it it creates a way of um you know this trustless um we want to bring it to the idea of preventing misinformation that's uh and again that's a typical example of the type of use cases that we expect to

10:33become popular where blockchain is used not necessarily to create tokens that people trade but instead used to uh create tokens that establish establish facts about the world that are then irrefutable and unchangeable oh that's a great example naven and I think you're right there needs to be some way to align expectations when you increase that liquidity for and then and you also

11:00break down the barriers of Entry to investing now you're getting more people that you know previously couldn't invest because they didn't have a million dollars they had a small amount of money in part was in part those old regulations were to keep people safe and protect their investment um but now people are throwing in $10 and expecting it to turn into $100 within two days and

11:24they don't really have the the financial Attunement to understand that yeah a startup especially Founders and long-term investors you know they have five to 10 or more year investment Horizons and none you know Amazon wasn't built in a year uh and it took them 10 15 years for the revenues to start going exponential you know and I think a lot of people

11:50they when they get into blockchain they want to go they want to make it all right away and I feel like uh it's to when you do that it's also really easy to lose a lot right away absolutely and uh I I think what happens in these sort of crypto bull markets when everything is just sort of going up into the right um it's uh it's easy to start believing that um you know um your

12:14investment strategy is working whereas in reality your investment strategy buy any token is all going up you know uh so the um and people have outsized gains uh which leads to further sort of them telling telling their friends and everyone comes in and everyone's investing 100 bucks and whatever right so um and now you in you invested in this token and it's a great product but

12:39um you're not looking at the revenue of the company or the potential of the product or the potential of that product to dominate the market or anything like that you're just looking at what's what's the my token doing is it going up or not is it going up is it down oh my God it's down let's sell let's sell right so it's a completely different uh type of uh game that that people play

13:02and um for Founders like we we we we have um I've met um like two or three of these companies where they're like look um you know we had a great six months and now we don't know what to do because now our token is down massively and they have trouble now raising even Venture money because people are like well you know you have a bad reputation out there

13:26because your token is down you're like no but actually it's a good product you know so um that that's like those two things conflate with each other yeah it's a bit of a a curse of of the market and in in some part uh beyond the control of the company when you know some of these companies I I'm sure you understand there's a lot of blockchain projects that are still looking for a

13:50product Market fit or they haven't really found it you know they're sort of a solution in search of a problem but there are some that actually have great Solutions and they are solving things but when the market as a whole you know if there's Bitcoin crashes and you know the rising Tide Rises All Ships but also syns All Ships uh and how do you survive

14:11you know as a startup uh you've probably had experience with this personally where you know there there was you start your company right at the the peak of uh excitement and then all of a sudden there's four or five years where nobody wants to invest any Capital even though you have the next Uber in in your hands so uh absolutely I have experienced this

14:30myself um um a couple of times at least once in 2008 when that big crash happened and once in uh 20145 when sort of the fintech investment slowed down and what saves you as a startup is the fact that uh you have a good product that your customers love right and you're able to if you're if I'm able to show to um to an investor or to a VC um that look it's it's a solid product it

14:56has this potential we're addressing this Market um uh we have this customer we have this customer growth uh you know it's like um I don't know if you've seen there's a talk that Paul Graham um gave uh nine 10 years ago on um like should you have a startup or not and one of the things he says is the most important thing in startup is to actually do good right if

15:16you're doing good which means if you're if you have a product that your customers love and um you have traction um then you don't you generally don't have uh trouble raising money valuations get trick right so um yes you raise money at a $50 million valuation and the market now just doesn't bear it that's okay you now you'll raise you'll have a Down Round and have a raise money at a

15:38$20 million whatever right um and and that by itself by the way also is hugely problematic because now you have to go start reissuing your stock options to employees and number of other um complications that happen is a bit of a red flag for future investors but still ultimately um You Will Survive and come out of it if you are actually doing good if you if you actually have a product

15:58that Works um so but you have to hunker down we um at one point in in one of my companies we went from uh we sort of did a very radical pivot and we went from having um you know 110 employees to about 25 employees right we just really hunker down and said what what is the most important thing that we're doing can we just focus on that um and uh you know you you come up against hard times

16:25and uh you have to make hard decisions and um ultimately what what matters is do you have something that that your customers love M yeah great point svine and now to go back to the example that you gave of uh the new Venture that you're working on with the uh the Trust app and and AI you know there's been discussion around fake content for a few years now and I think you know the Tom

16:51Cruz deep fakes was one of the first example um but we haven't really seen it with everyday people but in the last year I've seen the Technologies uh grow drastically with the expansion of you know Chachi BT into video and audio and and it's starting to get to that point where you know your everyday person might be impersonated um or you know worse political leaders fake videos uh

17:15images I feel like this is something that is an Ever more growing threat that I don't think anything uh like any Global potential catastrophe has happened yet but I feel like it very well could and I'm curious on your insights into you know how serious of an issue you think this might be and how important this technology could be for the future um absolutely so I'll I'll sort

17:40of add a little bit of historical context um and this is I want to warn you it's a little bit opinionated but I'll I'll put it out there anyways so um you know there was a time when uh sort of trust was very high espe let's just talk about the US right so in the 1970s was was a time when trust in institutions was generally high it was like uh you know your news institutions

18:05or the government or whatever right and since the 1970s no matter what your metric is trust has been eroding in the US at least and actually more recently all around the world right different countries have different cultures have different uh ways of establishing trust and um you know even within the US it was not Universal right I actually lived in um in in Georgia for a couple of

18:26years in the '90s and uh have a whole other uh system of uh establishing trust um and their level of distrust of the federal government is not new right so the the so but but trust was very high and when people came to the internet in the 1990s they sort of brought their trust with them right so um the phrase um uh the phrase was well it must be true I read on the

18:53internet it was actually a true thing like people actually said that in the '90s it must be truth on a website right it must must be true um uh it wasn't until the Advent of user generated content with Myspace and Facebook and these um other things where really anyone could post anything right again um the default is that people tend to believe things that they read right um

19:18that's because they're they're bringing that from from um uh from their uh sort of that's an overhang from a different time from a different era where people tend to believe in institutions I think the place where we're getting to now is that people are now beginning to or should I think adopt a very defensive uh posture where you need to first start by

19:42not believing anything that you read yeah so if you see a photo of whatever it doesn't matter what it is it doesn't matter who's posing it let's start by the that that's wrong right let's start the assumption that it is that someone's lying to us um and let's question their motives right um um then we have to gradually expand our circle of belief and say who is it that we want to

20:06believe what content do we want to believe how will we decide what to believe or not right so there there's a few different initiatives that are trying to do that so going back to your your point about deep fakes and fake things so Adobe has started a um a um a Content authentication in initiative um where it's a Consortium of many different companies and the idea is that

20:27original that you want to be able to label original content um um in a way such that um a consumer can decide of course the problem is that most consumers aren't deciding the consumers are just consuming it content on social media where there's no way of verifying it right so um with with our effort with Trust app we're trying to address that head-on right so we We Begin by saying

20:51what is a very very narrow slice of uh truth that we can get at right so let's let's um not tackle um uh let's tackle Providence right so you're saying that look guys the New York Times is claiming that Israel bomb the hospital in Gaza uh this is something that happened a couple of weeks ago um well let's just start with is the New York Times actually

21:13saying that right um so at least when you browse your social media uh we can just put little red marks next to everyone that's saying that New York Times saying that because New York Times is not saying that you know um um so that we can we can at least stop you from forwarding the retweeting or whatever those those pieces of content that are clearly misrepresenting the

21:37sources right um whereas somebody else who says yes this is something that whoever um the BBC is saying or CNN is saying or whatever um we can put a green thing saying yes they're accurately representing their sources right let's just start so we um you know from a product perspective I'm a big believer in sort of taking small wins right you make something small and you and you get

22:00it right first before you start to go broader right so um our first attempt is an establishing um sort of uh information provenance right which is the idea that um we can confirm that something uh that a source that you're quoting is did actually say the thing that you're quoting right um then we can talk about well you're not quoting a source you just saying something right

22:24um How likely it is that what you're saying matches a number of other sources right um what is how do we establish ground truth so for example the um the the the the Cai the initiative that I just spoke about they actually publish some tools and we want to be able to use some tools to establish authenticity of content itself right and and we want to

22:43use those tools to then add that to our little green and red check marks and say um yes um this is true according to the authen authenticity established here um uh with with the with with the tool that Doby created um yes this was actually photoshopped not um you know it has been photoshopped in a way that is not consistent with it being the truth or it

23:08being real right um and we uh right now the Cai initiatives don't plug into blockchain but we want to be we want to do that so we want we believe that blockchain represents a sort of uh fundamental technology that allows you to establish facts in a way that is platform agnostic in a way that is um un beatable and unchangeable and once you establish that yes this piece of content

23:33was published by this person at this time um that's it now uh they can't go and take it back or you they can't that cannot ever be deplatformed or whatever you know like that's not a um the whole point of blockchain is um it's your own thing um you and it's a public good in that sense right so we're trying to use blockchain as a public good um in that

23:56sense um to be able to publish content that is um uh accessible to everyone at all times MH very interesting naen and you're right about you know the the the recent news and it's hard to tell what's what's real and what's not um and there's so much information coming out you know even on on YouTube and videos is like days and days of content every second that comes out it's almost

24:21impossible to monitor the entire internet uh I'm curious how would you measure the you know the attainable Market size for this kind like monitoring this kind of information and using AI to determine the authenticity um uh so we are not so we're trying to tackle small use cases and we're trying to empower users to uh protect themselves that's really the the

24:47attempt uh that we have right now right we want to give the user a tool by which they can protect themselves as they're browsing the internet um uh um is this information um true or likely to be true or something like that um the market for that um empowering users is as big as um the number of users who browse the internet right but it's difficult to say what the

25:12um what you know like what features we need to have to be able to attract more and more people right because um you can imagine that there there's a class of user who's interested in establishing providences and a different class of user who's interested in finding dat different class of user who's interested in something else right so we need to keep adding features to attract

25:34increasingly more users in response to your question which is I think you were looking at from the other side like there's so much content how will you ever process everything um uh that's where I think uh uh sort of that's where blockchain really begins to work because it really opens up us up to decentralization right um I have zero hope that just as one entity or one

25:56company or one product will be be able to process all of the world's information right but our goal is to make um uh to to create a platform where lots of people are um publishing content on the internet um are publishing content on chain and um that um is again it's decentralized in the way that it is put on chain um which is really the only way that you solve a scale uh you have

26:24tens of thousands or hundreds of thousands of entities that are putting content on chain um and the same thing when it is consumed you have to consume it in ways so that it's not one single API but it's served in a decentralized way um supported by blockchain that's that that that decentralization actually becomes really important to achieve scale okay no great point and if I

26:45wanted to try this AI is it already available for the everyday user um not yet it's coming out at the end of this month so we are planning a uh post uh Thanksgiving uh type release uh in early December but before the holidays start sounds great and uh what other advice would you give to people that are trying to stay on top of uh blockchain news Ai and all of this kind of

27:12Technology um I think uh they should um they should read um sort of you know listen to informed podcast but really look at what the AI experts are saying I think the AI experts are um there's a lot of concern around regulation and um the potential of AI to do harm and these types of things and I think those are very important issues because um if we

27:38regulate AI in wrong ways or in um ways that prevent Innovation I think um that's a recipe for bad things to happen right that's a recipe for protecting incumbents and uh bad things to happen so it's very important for AI sort of research to be able to continue and it's important for everyone to um understand what the issues are understand what generative AI is capable of doing and

28:01what um how these AIS differ from AIS of 10 years ago for example and um what is it that it can do that and how can it how can we mark it for example in a way that it's clear that this is AI generated you know um all of those things are are um important considerations and what's going to happen in the future and it's important for everyone to stay informed definitely

28:25well thank you for all your insights nain and is there any place that uh we can follow you and your team's work for this kind of stuff um at Cumberland Labs our website where we post all of our projects and um our Twitter handle as well Cumberland Labs sounds great I can leave those links in the show notes as well appreciate your insights into all of this uh technology and Innovation

28:47naine you're really on top of it and uh it's going fast it's sometimes it's a little hard to keep up absolutely absolutely happy to be here thank you so much for having me appreciate it