The AI x Digital Assets Trade Nobody Is Making

Transcript

We are live here with Brian Wong is a venture capitalist and an entrepreneur. We are in New York City. >> Yes, we are. >> We are in the Empire State Building. Welcome to the tokenization tower, Brian. >> Thanks for having me. >> Brian, you have a interesting background. A co-founder of several companies. Uh raised a ton of money for your consumer brands. What are you doing in in New York? I also saw that you're digital nomad. I'm a digital nomad. I've been living in Brazil, Asia, Europe, all these other places for the longest. What do you do in the city, man? >> I live here now, actually. Okay, cool.

>> Yeah. Yeah. Yeah. Uh, I love it here. I've been here now technically 10 years, but on and off between San Francisco and New York. >> So, you're So, your nomad careers came to an end. >> Nomading has now ended. Yes. Yes. >> What was your top spots, man? >> Uh, I mean, I've always been all over, but I still love uh Mexico City. Okay. >> Um, Vancouver, which is where my family is from, where I was born. Uh, San Francisco, where I spent a lot of time building my last company. And then of course here in New York. >> Nice. >> New York is still the best city on earth. >> Yeah, 100%. >> Tell us about your fund. Is it Aski Ventures?

>> As ASKI Venture Aski Ventures >> ASI Ventures. Um what's your thesis? How did you get this started? Um and yeah, man. Welcome to the show. >> Thank you. Uh well, ASKI for those who are nerdy enough to remember is the precursor to Unicode. It was American Standard Code for information interchange. It's like meant to kind of be like a filter for the nerds that come in and and pitch or that that we work with. But ASKI is a fund. It's a It's a boring tech fund is what we call it. It focuses on behind the-scenes technologies, things that are powering other things but aren't really seen and are meant to be used daily, but you just have no idea that it's being used daily.

Like those kinds of things. So, obviously a Plaid or what was a Stripe, you obviously wish we would have been a part of those companies, but no one would have even known those were businesses and they're behind the scenes most of the time, right? You don't really see them. Uh they've built consumer brands on purpose to, you know, enact trust, but most of it is behind the scenes. We focus on those kinds of technologies, picks and shovels. Um, we have a unique strategy. We're doing primarily early stage, so things that are before revenue mainly. That's the goal. And then we're also, so anything under 15 million valuation and anything over 15 billion is what we say. So people kind of pause for a second like what do you mean by that? We're doing a bifurcated fund. So we're focusing on two different asset classes that are on

the way end of the spectrum, but it's designed to create a shorter timeline as a fund. You know, most funds as you know right now in traditional VC has been returning 10 plus years. A lot of criticism. I'm 35 in a few days. I don't want to wait until I'm 50 before I have any kind of return. Uh this fund is designed to return around 5 to seven years, which is super super super short. Um the one thing we're highlighting though in crypto and in blockchain, so we do look at different categories. Crypto and blockchain is one of them is really exclusively in stable coins and the boring tech layer of stable coins, which is actually the app layer. So when you think about it, there's there's really no innovation needed, right?

There's a lot of trady things are just being mimicked in stable coin, right? So think uh payday loans, right? Forex, right? Uh uh business lending, uh you know, even neo banks for, you know, people who are historically not able to get bank accounts, right? Those kinds of things we love. Um and they're just app layer. People like to criticize the app layer, but I think the app layer is where the exits are going to happen because you have a lot of incumbents like the big banks that are not going to build, you know, a payday loan service on chain, right? And are going to be able to do this by just buying a company that's been building this for the last couple years maybe. So those kinds of quick flips, those base hits, those are the types of companies we're going after.

>> Yeah. So it's like a barbell approach, right? where it's like you've got the extremely early kind of preede seed companies and then these large kind of proven Goliath looking to have one more fund raise before they really >> well that's more in secondary so we're actually doing it to find these companies that may not have become like an anthropic or uh you know an open AI which by the way there are many right so a big bulk of capital is focused on those guys but there's still like you know 100 plus companies in the 10 billion to plus valuation some of them have not been doing so great at least perceived wise but their core fundamentals are great and there's a strategy behind that if we buy into the secondaries direct and we're able to build the relationship with the companies there's actually a direct

relationship to buying the companies we're funding in the early side right because these guys need to catch up they want to buy talent they want to buy good companies that have that innovation down pat because the issue we're seeing is a lot of companies spent five six years building stuff right now which by the way in a month or two months god forbid AI could just build on a tech standpoint right So it's the go to market mo that really is what these guys are highlighting because they're early and they might have come from these places. So we're seeing a lot of founders come out of the banks that left right that are starting these things because they know how to do it. They can create a private credit product on chain in like two seconds and then get it in there. So that's the goal. And so I saw a tweet

from um Chimath talking about I think there was it was like the seven most valuable private companies and he was like I invested in two of these or or sorry I know I know five of these. I invested in three of them in the early stages and then I bought secondaries in two of them. And he was like basically going on to say that as an investor you clearly can't win everything. You can't get into every deal in the early stages. But the point was, as an investor, you also have to recognize when you did make a miss and you have to be willing to kind of put your ego aside and be able to buy these things at a much higher valuation than probably the first time that you saw the deal or the first five times that you saw the deal.

>> 100%. >> Do you agree with that? >> 100%. And that's why like secondary markets have been blowing up lately. But I would say again, >> I always run a different direction when I see a lot of people running another way, right? A lot of people are like, I want to get open AI, I want to get anthropic and d and these things are valued at insane. I mean, 600 billion, 700 billion trillion. It's crazy because you're thinking the public markets need to withstand uh, you know, sort of three major IPOs and everybody knows these companies at 23 trillion. So, you're talking 9 to10 trillion that were needed in this public market cycle of of these guys going public. SpaceX, OpenAI, Anthropic. Um, and I just don't think the market will be able to hold that and the market will also need to do that

after the lockups end, right? And so I think there's again I'm advocating uh and you know the theme now of these not so not necessarily the top of mind guys but they're also building I mean I'll name some like Canva, Rippling, Gusto, right in the crypto space, you know, Kraken obviously hasn't gone public yet, but Kraken is, you know, one of these guys that we want to win. We we would they have to win for us to to see another boom to the VC. uh VC world that has bet so hard on these guys recently obviously we had like the bit goes and the figures of the world go public but these these types of are figment these are the things that we uh need to have successful so >> yeah on the specific to digital assets investing side you mentioned kind of the neo

banking the stable coin side what is your thesis here like is it really just as simple as democratizing access to US dollar um savings and then the US dollar savings rate globally like is it as simple as giving more people access to the dollar at scale. Is it a payments thing? You know, how are you thinking about these things from a thesis perspective? And then where are you seeing like like the business case for investing here? >> You're absolutely correct. So, the access to the dollar is obviously a big one, but everybody knows this. We're looking a lot at again these kinds of uh products that are catered towards people who are historically not given a lot of these kinds of products. And this comes from corridors uh that are emerging, right? So this could be between two countries that are typically again not

not really served, right? And it could be from corridors that are emerging and rapidly growing, right? Like I this is not a company we're going to invest in, but there was a company that I recently met that does something specifically for Venezuela and the US, right? Let's just put it that way. And you know, again, if you really want to bet on that, you think about how Venezuela could actually in the future become a part of the US economy, right? Like what does that look like? >> And this is all completely apolitical statement, right? It's just like these are the things that are worth taking it into account. So emerging markets are super interesting because the way they adopt these products are very different than we do, right? You and I, much like if I were to use the comparison of desktop PCs to laptops to smartphones

and tablets, we went through that evolution. But in emerging markets, right, they went straight to the smartphone. They went straight to the tablet. They're digital natives, which is why live shopping, all sorts of different trends worked there really well. Live shopping, like the QVC evolution that we thought was going to happen, right, where we could just watch a live streamer talk about stuff and buy from them. That's not happening as big in the US as it is in Asia, right? And there's certain markets in Asia where this has blown up. The same thing happens with money, right? When you look at their first access to a digital dollar, they're not getting credit cards. They never had credit cards. We were thinking this is normal for us, but for them, what is the concept of credit?

What is the concept of spending money? As you know, in a lot of these markets in Africa, for example, the the main currency has has been for a long time your phone credits. Literally, you could get paid in how many credits I could get for my phone. I don't know if you knew this, but it's like literally getting more minutes. was a currency in these places for a very long time. So currency is different in how they perceive it. So these markets are better because in my opinion the extreme exponential growth curve rates are more possible in these markets where adoption is a lot more likely by people that didn't have something that they had been used to before. Right. And that to me but we're borrowing from concepts that have worked in America and other developed countries where we know that that people will

respond. Right. >> Right. Yep. Yeah. Yeah. Actually something that the audience may not really understand as well or pick up you know intuitively is that the adoption of a lot of this tech is actually happening faster in in continents like Asia. >> 100%. Yeah. 100% 100%. It's because there's no preconceived notions. There isn't oh I I I used to use it this way, right? Like you know it's funny cuz like you know when we used to text with the with our like the numerical pad where we had to press the button that was because we had to learn that and then jump in. But if the first thing you used was a full keyboard with your thumbs, right?

It's that kind of thing is what we see with these regions. >> Yeah. Interesting. >> Okay. Back to some of the uh macro stuff here. >> So, you think that those OpenAI and Anthropic rounds at the $380 billion valuation for Anthropic now into the 700. Uh Open AI just raised that 820 or 850 billion. You think those are just drastically overvalued and that the market won't be able to kind of manage the float on IPO? I'm not saying they're overvalued. I'm just saying that the markets need to be able to uh provide the liquidity uh for the people that have been waiting quite some time for that. Right. And we saw recently that that cap table published of OpenAI and the early investors and all that. Go Ashton Kutcher. Uh clearly he did very well. Uh and Gaio Siri. But the thing is

there's this um there's this dynamic that I'm observing around liquidity and this has really been the complaint for a very long time with ventures that we're waiting for liquidity. So these three events are a big deal, right? Because they will return a lot of liquidity to the markets. However, uh what I'm saying is there are uh companies that are probably not as over not no sorry that don't require that kind of liquidity event that will still IPO that will do well that you can still play in right um I think what's happening is because of that this crazy I just had this conversation earlier today I loved it this this crazy need for energy this crazy need for capital uh to buy chips is really only important if we believe that the outcomes that are a result of these inputs so inputs outputs the outputs

actually matter, right? Right now, there's still a big question mark, right? We know AI is incredible. >> Are the outputs really giving us that commercial use case that people see that will warrant those valuations? So, that's where we we do the calculations on valuations. Like, I I and I don't know the answer to that question, right? Because it's like the jokes you see on on online are always like, you know, the amount of lakes we've drained just to have people make cat videos. What is that going to to actually mean in the long run? And the beauty of markets is it will always figure itself out and the markets will will will rationalize this, right? How much commercial output is coming accreatively to the rest of the world that's adding real value to yours and mine my's lives that actually lead to more commerce as a result of of of

this crazy equation we're dealing with because the inputs are really energy and chips, right? And then the reason why this is happening and people probably know this or don't is that the investors that are already in of course have an interest in telling you about how pervasive AI is going to be. Everything's going to be AI, AI, AI. And I do believe a lot of things will be AI, AI. And like with anything pendulum shifting, um, everything will shift over, try to come back, but usually in the world and human progress, what's beauty about about our race is that it'll always bounce back just a little bit ahead of where that equilibrium was before, right? If you think about it, like the internet did that, right? It's not like everybody, you know, every like, oh my god, when the internet came out, everybody's losing their their

minds, right? But like, but of course, everything kind of snapped back here. We're snapping back there. But on the other hand, there's a lot of other things happening. I don't know if you knew this, but record vinyl record sales hit a like literally a record, pun intended, last year, right? It surpassed CD sales because CDs were dying anyway, but like there's a lot of analog stuff that's happening right now that's like showing how the world will always have these these kind of snapbacks, these balancing acts. And I know it'll bounce back here, but how much of that hop ends up happening net after everything calms down, who knows? And that's the exciting part about being a more long-term thinking investor.

>> Yeah, man. I mean, we're seeing a return to in-person experiences% >> like these kind of gated exclusive events inerson face to face >> clubs have never been hotter, right? Like like like members clubs, right? You you've heard all the new ones. Shargo, like Zezys, all these >> Shargo is great. >> Exactly. Yeah. Because And then there's now clubs that are starting to lock your phone up, right? Like like Sanenti Bungalows, right? So these guys are Yeah. these things are are starting to create a new dynamic where I think the luxury is to not need to be connected and tied to these things. But again, like I don't want to sound like a lite.

These things will always be around and they will always create a lot of progress. Um it's just where does it snap back uh just a little bit ahead of where it was before. >> Yeah. And I mean the question of what are we producing for all of this consumption of ships of energy as you say that you know there's Elon's posting like some weird form of hentai on X as [laughter] a means to like you know trigger uh growth. >> I don't think people understand though the the scale. It's funny like we're it's I don't know if people pause to just think okay >> about the the trillions right like and and you you know you remember when Instagram got acquired for a billion everyone thought that was that that was the equivalent of like $30 trillion today it just felt like a lot of money

and that wasn't that long ago right and when you think about it >> we're just not fully grasping I think most us mere mortals are grasping just how much money is being moved back and forth on these things right meanwhile we're fighting on 5 $10 million venture rounds $20 million venture rounds which now kind of sound like a joke and then meanwhile there's like you know couple hundred billion being raised on a regular just to fund these chips. So this the scale is a very different conversation. >> Yeah. Yeah. >> The scale of the average seed round >> correct >> in AI and in AI venture has just gone astronomically large >> and the expectations are actually squeezing out those who are trying to innovate. Let me explain why. Because of cursor, the expectation is now you need

to do a h 100red million in a year basically, right? Like it's crazy. And and what if your AI company doesn't do that, right? Like what if you do 20 million in a year? You should be proud of yourself. But that's not enough. I just met with a founder a couple days ago, exited to another AI company. I invested in him a few years ago. Um and he was like, my gosh, we what world is doing 20 million in in ARR like a bad thing? Yeah. But meanwhile, they did burn 32 million to get the 20 million, but that's not like that crazy bad, right? But like it the bar has been set so damn high though and this is the worry. the use cases now that people are going for are going to, you know, sort of try to whittle to that like where's

the fast money, right? Which is the vibe coding, the whatever seems to be trendy, but I want to look for long-term gains. Which is why another thesis I did post and I do want to share with this with the audience is the fact that I believe that services businesses, believe it or not, which have been poo pooed by VCs for quite a long time, right? Entirely >> entirely like don't do services. >> Yes. >> Is making a comeback. >> Okay. >> But the the perspective is different, right? I don't know if you got saw the YC memo that was kind of leaked. was like about what they're looking at now and they they do talk about this with just the the return of services but really rather than providing individual services you are the service okay >> if that makes sense so you're literally like like an Accenture and a Mackenzie

and a Bane because they're so damn big they're doing everything for the clients when they come in and it costs a lot of money right but when you are a startup that has let's say I'll give you an example let's say you're one of the companies we invested in was is called Nexa they do agentic autocad so it's a 3D software for construction right um historically it was like hey log in and do your thing use the SAS software and we'll kind of train you and whatever one time and hope to god that you end up knowing how to use it and you keep using it, right? That that's the premise of of SAS. >> But now because of Agentic, I can actually provide you the service of walking you through it, right? Doing the actual work for you and it's just as automated as it was before. So this

whole army of consultants that you would have needed doesn't need to be there and the cost basis is not the same. So you actually are able to provide that service. So I think if there's a vertically oriented business again could also be in stable coins let's say again you're creating a a genic layer for for stable coin transactions to to to clear them for risk or whatever it is for small businesses right like for fraud and you run it the whole thing just for them the banking layer including accounting tax blah blah blah and it's all in one service that's like bre with stable coins with a service agent built in which could be what airbor ends up doing right when they when they end up rolling out their products to a wider uh customer base is massive Right. So the expectation now is that services will

make that comeback but it'll just look differently than what we were trying to avoid before as VCs. >> Right? Because the frame of thought around services was scale your offering, scale your services, scale your team, you know, decrease margins. >> Well, it's linear growth. >> Yeah. But that's been broken now. >> Of course, because there are things you can, you know, again, the agentic orchestration, right? Uh, again, I think the biggest profit margin companies will be agents that help you put together agents, right? Like a network, your your a network of agents. Cuz most businesses, let's be honest, we think America, well, we're in a bubble, right?

Like you have to admit, right? We're here in New York and there people in San Francisco and LA. We think our commerce sort of behavior at etiquette is the same as the rest of the country. No, there are people who are busy building companies that are, you know, doing construction, doing, you know, garbage disposal, doing carpentry, factories, industries. They don't have that time, right? So, if you orchestrate that those agents for them, they they they are able to do it without even thinking. And this is also where we then creep into promptless, which is another big area that we're looking at as well as a firm. What are things where the prompts are not required anymore, where the AI is prompting you, it's a big deal. And this closed into crypto land. Open Clause really shifted that too because it went from this kind of a lot of the previous

iterations of LLM or just any sort of chat flow was um very kind of like uh descriptive and like giving you a uh you know you have to tell it what you need it to do versus like uh prescriptive like here's what you could do and like do you want me to do that for you? that that flip is once you kind of experience that flip in terms of your usage, it's you can't go back to what you were doing previously with the basic >> LLM. But commerce, like we don't scale agenda commerce if I have to tell it to buy things every time.

>> Right. Exactly. >> Agenda commerce only grows if it's buying for me. Obviously, in trading, this has been the case for a very long time with just, you know, quant firms and algo firms. But when you think about consumers and their access to buying products Yeah. >> I had a friend again yesterday at dinner. We just talked about it. He has agents doing replenishing of homegoods. It's that simple, right? That that's that's very simple. Sounds innocent. >> Yeah. >> But it's a big deal, right? And the thing is without it needing to prompt you and buying things and then using obviously the ramifications here, using stable coins to transact and to buy things for us is the wave of the future.

And I do believe that that's a big big area we should be focusing on. >> Yeah. So, let's get into this prompt list uh >> kind of investment thesis. So, we've had this era of LLMs that have been prompt first, right? You >> you you are specializing in the ability to prompt, correct? >> Such that you're trying to optimize for the correct result with the least amount of back and forth. Yes. And headache. Um, and for so long we had just like no that no that's not no like it's just this constant kind of battle. And so promptless >> talk like talk me through the the functionality here. What does that actually look like in practice kind of let's understand that and then let's get

into some of the thesis on the investing side. >> So Anamoka created a product called Anamoka minds recently. Not a lot of people know about it. It is their effective version of prompt list where you can put together a lot of agents that work together and agentic orchestration is the equivalent of promptless in v1. But I'm going to take it a step further. It's really funny though that as human beings we have built this incredible AI and the only way we can really communicate with it is through these lines of of English which as you think about it language is also very I just be very philosophical for a second it's a very inefficient way to communicate right think about it right like me is basically a bunch of fleshy you know bones and cartilage in our mouth making vibrations to communicate with you if you think about it we talked

about this you and me are doing this it's going to take us like 30 minutes to communicate something that a computer could communicate theoretically they can do like that, right? So my belief is where the best prompt list is having the best access to the right data sources at the right time, right? So setting up your your data so that like when you set up your open claw, right? You obviously want to make sure that it's pretty like, you know, guarded, right? Like it's in one little box literally, but you're feeding that box with things like you have your file and directory structure in a way that you think, you know, OpenCloud will be able to understand and be able to parse through and then and then be able to do the things you want it to do. So if you think about it, it

was the setup of the data of what it's reading through of it what it's being trained through that is making it do what you want it to do without it prompting you or without you prompting it or it it prompting you right and then of course you teaching it that you know if there are moments where you need to ask me for clarity you should be right but that's really where I think the future lies and that's why data or these data cleanups data dictionary data consulting businesses are still so powerful and also where your sovereignty comes from with that data is is is key Because once we get into prompt lists, how clean that data is, how comprehensive it is, right? How much of it you own, how much of it you can use is going to be the key and will continue to be a massive part of the business.

Like it is own like again it's the LLM, but it's almost like the LLM was trained on the patterns, right, which is what you know the GPTs were on before. The new patterns still come from the same data sources, but you need it accessible so that when it's trained already, now it's able to do the prompting again back to you. And I think we have to also get our minds out of this equation that we think is like >> cause and effect. Me give it something and it gives me something back, right? When you think of agents having multiple work streams now is just like the world, right? It's not like I asked the world for one thing and one thing only.

There's a lot of think processes happen like just getting your Chipotle burrito. A lot of processes already happened before you came and prompted it. So when you think of your life as that where there's a lot of processes that are already running and then that prompt was a result of all those things, that's how you think about the the kind of agentic non-prompt world and evolve from it. But >> I also believe there's a lot too much going on. You can over complicate it, right? And I think trying to making sure that you only kind of build in your own life that hierarchy of things that you know can be impacted and then yet not using it as a crutch is the reminder I give to younger people because one thing that you and I I hope you're not that much younger than me than then I feel

bad. But like the idea of like you know if you don't if you don't use your own mind to initiate thoughts anymore you actually lose your only reason for existence as a human being like initiating true clean pure uninfluenced thoughts is really quite rare and it is really the essence of being human that I think we should try our best to preserve. >> Yeah. Wow. question on kind of like how you have people thinking about prompting and then prompt list transition. Yes. How do I go from like a uh a habitual execution of a prompt to get a result to operating now in a more promptless world? Like what does that transition

end up looking like for like the end user? >> Well, find uh well, in my world, it would be finding businesses that don't want to be bothered with the idea of needing to prompt all the time, right? So when you think about it, there is a company we're looking at that is mainly doing promptless executions of experiments of growth and marketing just for small to mediumsiz businesses. So you think about lawyers, you know, accountants, you know, sort of practices like you could even have like a dry dry cleaner, whatever it is, but what are things that they're doing, but they don't have the time. Again, they don't have the time to to think of the prompt. They don't want to bother learning how to prompt, right? That's great. Within your own business, however, there are things that are similar. What are processes that just should be automated,

that should be constantly analyzed, right? This could be employee satisfaction. This could be treasury management. It could be taxes and expenses. These things that actually as a CEO, because I was a CEO for like 10 years, like what are things that you're constantly needing to think in the back of your mind that you really don't need to be consciously thinking of and that you wish you had someone doing, those are the things that you you you regularly want to have this promptless model working off of, right? And again, this will become more popular. Again, mark my words. I do believe both perplexity computer and then cloud work co-work eventually all these things are going to have promptless models that are running in the background that are create that are again working on tasks that you've you've given it as an ongoing thing and then it essentially

being a human being that can improve, give you suggestions, you can talk to it, that kind of thing. >> Yeah. And when people are using these kind of more promptless models, are you are you seeing a world where there's kind of like parallel workflows happening on for all parts of someone's kind of like let's just let's just focus on on the business side like is this like a set it and forget it like type of thing? Just like talk to me about how you actually like give us some concrete examples here. Well, there's remember the last time you let your your AI just run free and then and then not monitor for a bit. It's disastrous, right?

Because this thing still is is a little bit like a I would say it's still like a c like an undergrad is really the the you know you think about undergrad lot of knowledge execution poor right like still like I gota I got to check in with my prof, right? >> Um tangible examples in the businesses where you're running this stuff. I I think we're we need to overcome this trust gap which still exists with these models where I've tried many times to be super clear in making sure that everything's set up the way it is but they it will break. It will just and then needs you to clarify and step in which >> sometimes the the sort of conspiracy theory is this is built by design so that we do have this natural uh you know sort of workflow where we're used to

checking in right because we just let things run right now. I don't think these AI companies will ever admit that these things are uh you know going to work without error right and then unfortunately if this starts to automate things that where errors do create catastrophes like air traffic control god forbid we don't want these things to keep so I think that that the the relationship we have with these machines where we have to check in once in a while is still pretty important right so again down to the business layer if it's employee satisfaction if it's checking in with with with goals right and checking on performance >> um what if it calculates bonus is wrong.

What if it looks at progress and and counts something that wasn't counted, right? And people naturally want to game the system. So, you do want to be a part of it. So, that again, you know, the human input that that actually matters, which is knowing when people are trying to trick the system uh really really really counts here, right? So it unfortunately goes back to human fundamentals where where will there be gaming of the system and how do you prevent that from happening through check-ins and ultimately why I think again the valuable skills that will still be required in ALAD is still discretion ethics morals logical thinking and creative thinking those are the things and then questioning right those are the things that will always continue to remain valuable in the in this in this growth area

>> yeah I think it's fair to say that if you think about things in this lens right with just the actual state of uh how these tools can be used there's you know we're kind of we're kind of calling them an undergrad here but it's going to get to a postgrad and it's going to get to an expert level it will >> but even still >> I feel that it's accurate to say that um the AI is going to take all of these jobs claim is still very very far-fetched in the sense that I don't view it as a replacement of someone as much as I view it as an amplifier of someone's already existing expertise or skills. Y >> do you agree with that?

>> I completely agree with you. And the the the moat here is people still don't realize that AGI >> is like a decade away, right? Like some people are still are saying it could be shorter but latest, you know, it could be 10 20 years actually even, right? And but what do they want us to think? They want us to think that it's closer. They want us to think we got to pour in, you know, hundreds of billions of dollars more to keep this kind of, you know, thing going, right? And so again, the that AGI is a replacement of critical thinking and thought, which is really where we're what I just said is pretty important for us. Now, I don't want to be in 10 years like eating my words and be like, "Oh god, like I'm useless now.

I'm basically like a piece of piece of flesh that's just roaming around on the earth." But ultimately, that is what will preserve that important part. And then like again, with every crazy prediction that was made, oh, you know, the internet will you we no one need to travel anymore, right? Cuz we just stream video. It's like no, people are still flying around like this. This will never be replaced. And then the trust factor is one thing I will I will point out. Trust is built through in person, right? Yeah. There's it's a very very untouchable kind of thing about human beings that you cannot really explain. That will always be needed. And so I think there's going to be that luxury of like how did I build my relationship with you before, right? And so there's a lot of people that I've known for like 10, 15 years

that our trust actually seems to be compounded more by the fact that there's so much reliance on this virtual connect that the fact that we had in person experiences before is actually compounded. It's making our relationships even stronger cuz now I go, "Oh, wow. This person actually shared, you know, I broke bread with them. We spent a lot of time together." even it was literally just like meetings in person because we were in an office back in the day because we did our office actually used to be um one of my last companies was was a couple blocks from here was um you know about that you know just seeing that person and and I think that it's so simple >> but you can't look at an AI in the eye >> no >> not going to happen >> so then if we're if we're you know of

the understanding and agreement there what would you recommend to people who are watching this who want to get more from their people that want who have folks in a variety of different parts of their company or who who have uh or who are investors who want to get more from their founders like when it comes to actual like implementation of these tools just like generally curious how you're thinking about like the progression that you're seeing founders companies staff employees undergo that is like the most you know logical and thoughtful progression in using these tools >> easy so obviously offer to subsidize and fund, you know, all the subscriptions.

That's one, right? Doesn't matter which one. Just go nuts, try them, have fun. Number two is actually a gentic day, right? So, get a lot of my companies, I've seen them do this where we spent a whole day, build your own agents, have them work with each other, have a task, that's an outcome, and then give it a whirl. Right? This is the equivalent of back in the day when it was like build an app, right? Was like go and just do an app or build a website. Just do something just so you can test it, feel it, understand it. Does it need to be a part of your everyday life? No. Right? But of course this at least gives them that taste and then a lot of ideas will come out of it. The other thing is having the engineers and the non-engineers interact on these things

right. So hackathons where they work together for a minute mini business outcome. That way the crazy stuff that the engineering team is starting to to touch and and I can confidently tell you 100% of engineers right now I can yeah hands down 100% are using AI to write code and they're really just proofreading it right and QAing it. So to me it is really about accepting that it is inevitable right but making sure that people don't use it while being lazy and then having a crutch for it right but then also realizing that quantity is not the name of the game right right like claude research super good at giving you this entire list of stuff right but you still have to filter through it you still have to use your brain don't use it as a crutch right but it's it really is a soft thing and as a

call to to arms for all humans is don't that that really kind of you know in your your devil and your angel and your shoulder be like, "Oh, no. Just use the claw to check it or just use cloud to do to prompt the checking of the checking." Don't check the checking. You should do the checking, right? Because at least you you using your brain there, right? But you you've had it like, "Hey, just clean it up for me." Like, no, you clean it up, right? But so in the moment, use your brain for something, >> right? It it's the meta the meta use case of it is really dangerous because then we become literally like an idiocracy. Like it I I worry that we we all become vegetables because we're not it's it's atrophy. It's it's it's it's neural atrophy.

>> Really, >> you know, if we don't use our brains, think about it. >> When you combine that with the consumption of short form content, >> oh, we're all just brain rotting. But that's literally the definition. But even this like mental math >> Yeah. Okay. >> Let's just use that as example. >> Sure. >> The ability for people to do mental math has like declined drastically, right, since obviously iPhone calculators and just us not why would I need a calculator? I can just do whatever, right? But there's a reason why mental math and I don't want to make this a race thing, but why as an Asian we were like drilled to to I know I'm sorry we memorized everything mental math wise. I know but we there's a reason [laughter] I'm so dead as the Indians. I would be I would be like in the '9s 89 92 range and

then you guys would come in at like 97 every time and it'd be like what the going on? >> Well, the mental math was actually not about route memorization which is what I thought it was and I hated. I resisted it so much for my parents. It was actually being able to understand why having that part of your sort of daily t so when you're doing a math equation on a daily uh thing like like tips simple things right or like oh I'm calculating a few days out or whatever dividing it's always part of something grander. So if you can shorten or make a heristic of one part of it you then become just more efficient and you you understand that that piece is is something that you you actually had to do the actual mental work on. So I think understanding how the sausage is made in that way is

really still important. So that's why I think you know having actually most of your employees take a basics of LLMs and understanding how you know GPTs are trained models literally they're trained from previous you know generative uh ideas. That's what GP stands for. That was something that people should understand because if you don't then you you don't realize everything will revert to a mean until AGI completely comes out. Right? So reversion to the mean is is is is is terrifying because then a lot of the ideas and concepts end up being kind of around that same thing that was based on popularism. And we're trying to avoid popularism in the first place. And then also you have to understand these models were trained to please you, right? They're trained to make you feel good, right? So now we

have people who have no idea of actual proper constructive criticism, right? because the guy that I'm talking to or the gal or whatever the gender is of the person I'm speaking to on on on my chat GBT is never going to argue with me or tell me straight up that I'm being dumb. >> And I think that's why we need to bring back bullying. I'm kidding. But like [laughter] yeah. Um okay. So on the LLM psychosis point, are are you actually worried about that? Right. this idea that you get this psychosis from that grandiose treatment from these LLMs and kind of uh always being told you're right, you know, all the way to the point that I've seen tweets that are like your your your worst like your worst enemy is a

narcissist and is getting told on chatbt right now that they are so right and that they're perfect or something like this psychosis of treatment from these programs. How bad is it? [snorts] [sighs] >> I think it depends on the maturity of the individual, right? And then how much of the world they understood or how much they developed in a world that was preai, right? And like you and I, we had a world where we do expect, you know, healthy discourse, varying opinions, different kinds of discussions that are part of ultimately informing you, right? The reason why I have these discussions with people who disagree with me is because I want to learn, right? not because I want to get into a fight, you know, or just be combative. And so I think that's the issue is like if the

ultimate objective is to learn, uh, then that's really great. But these AIs are not designed to do that. They're designed to make you happier because that way you keep using it more and they get more money as a result of your usage, right? And so this to me, I mean, obviously usage is not tied to the amount of money you're paying it, but you get what I mean. It's it's really this drug. It really al always comes down to that, right? what kind of it what component of that digital service are we using that gives us that that dopamine hit right and then on Instagram of course it was the the likes it was the comments it was the friend requests it was just the the fact that you felt like you were loved or that people paid attention to you where default you couldn't dislike anything you just like

everything it was always about positive right so we all know this dopamine hit comes from that right so I think we just have to be conscious of it and that's why I think companies and so this is another thing that we've been looking at and we actually invested in a company we haven't disclosed yet that does AI ethics right so this idea of deep fake detection, right? Understanding if this is actually biased, if the output of the AI is biased, right? Like if it's it's it's discriminating because the people that were prompting it, we're discriminatory in the first place, but now we're degree removed away from it, right? That like, you know, obviously I prefer this kinds of this kind of employee, but now and there are AI, by the way, they're screening employees now or prospective employees and hiring processes. these kind of biases, we have

to systematically go in unearth these issues before they get too crazy because then everything's going to shift slowly and slowly. It's frog in the pot. We will all get boiled alive >> if we don't try to stop these things from getting too far cuz we will not realize it once it's gone all the way out here. The concept of an Overton window, I'm sure you might have heard of that concept. Maybe people not, but the fact is we live in a shifted overton now because there's been a lot of extreme conversations about extreme things that now seem normal to talk about, right? And yet, you know, 10, 15 years ago, we would have never talked about that. We would have never fathomed that kind of conversation publicly.

>> That is the that is the thesis >> of many different products and companies that have grown substantially. Yes. Anything from like a hyperl liquid that offers 247 trading because the amount of headlines are just absurd now all the way to like an AI ethics company because of bias response based on traumatic life events from a individual and anything in between. It's just this over window has shifted. Uh I saw a funding round for a NSFWI uh protocol yesterday. uh you can take a photo of yourself naked, put it on there, and then you can then go and basically pimp yourself out uh in the AI

world as like an emoticon uh type of thing. Are you worried that people aren't having as much sex because of AI? Like >> Well, we do know that the fertility uh it's not fertility, but the birth rates in certain Asian countries are declining. You have Singapore paying couples to basically have kids. uh the uh fertility and and birth rate issues are real, right? Because if we stop, you know, having next generations, the entire society is messed up, right? So, there's a lot of other things that I don't even understand that will happen.

Go ask an economist, I'm sure, or more more so a sociologist. But these kinds of things are, you know, pretty pretty scary. I also think that we have still this this uh uh extreme end of of the super rich and the unfortunately not so super rich folks where uh if you have too much money and you have too much knowledge about this kind of thing, you're actually disincentivized to have kids because the world you're bringing them into may not be the most ideal for you and what you want for them. And then those on the other end know that they need the extra hands, the extra pair of feet and so they want to have more kids as a result of that. So I think we're going to have this, you know, sort of divide. I don't know what that's going to look like. Lower population of those

that are on the upper ends of society and then on the other side of society you have a lot of people but not a lot of enablement. And that's why I I mean this has become a very philosophical conversation which I love by the way is the fact that education and really using our minds to rethink the education model is still pretty real, right? We had the alt school. Uh we had the Adastas of the world, but that that wasn't a full experiment that that went through as if you recall them. This was Elon's school in in in the Bay Area which shut down. The thing is no one's really sat down and said, "Hey, this industrialized educational thinking that we've really done like this this elementary school, middle school, high school model was for the industrial revolution, right? It was to put people through a conveyor belt

like we were doing our products." And we're still thinking the same way. We're still assuming we can just shove kids in a school where all the professors have actually never held a real job. No offense uh to professors. Still love you guys. But like a lot of them have not actually held real jobs in the real world that are teaching kids about the real world. And there are schools that are better now that are having, you know, actual faculty that came out of the real world that I think is still required. But these that the curriculum is not catching up, right? They're learning on their own. That's why people want to drop out. And that's for some reason now that's become a mark of of like, you know, like a bragging right that I dropped out. Exactly.

because it wasn't serving you. And I think this system is not serving a lot of kids that have that critical thinking that do know that by the time you're finished learning, right? This the stuff you were learning is not even relevant to what the world is like after four years. Like four years is is a lifetime. An absolute lifetime. >> So, do you have kids? >> No. >> Do you want kids? >> Of course. >> Will you put your kids in in public school? >> Probably not. >> Will you homeschool them? >> Probably. Yeah. Really? >> Yeah. I I would find a a school that, you know, I' I've, you know, I I know of and have friends that have sent their kids to schools that I've looked at that I think would be great. Now, again, I'd have to be a billionaire to send them to

these kids. So, to these kids to this this school, so I have to work a lot harder. >> Couple more investments, man. >> A couple more investments. Yeah. But that that is the that is I'm not quite sure actually. It's really tough because I would rather like I have a niece who's who's turning six and you know, she's got uh an incredible reading comprehension now. I don't know if you have kids or nieces. Are you or nephews? Are you? Yeah, >> I am. I'm 26. You're okay. So, you're talking the thing is >> they're growing up much faster than we did, right? Like, and they're maturing much faster than we did because reality of the world is much more upfront, right? Like when we were younger, because we didn't have that that right in front of our faces through screens, right? We had to go and actually see it,

right? Mom and dad took you to somewhere to see it, right? If you're at home, it wasn't here. Now reality is so close in proximity uh they know they need to grow up faster. This has happened subconsciously, right? And this is not just me observing. This is a well doumented thing. And because that's happening, again, I don't know what the ramifications are for the kids later on, right? Because now there was no period of innocent play. And we all know that play is a great way to develop your critical thinking, right? Unstructured problems come from play. a structured problem of oh barrels of oil are now over $100 because of a war is a structured problem that is placed in front of you that now you are faced with some type of thing you need to try to come up with a solution for obviously

not to the 5-year-old but that is not a result of play no >> uh and so I I do think we we're losing play we're losing the sense of fun >> um you know again I'm not trying to bring all these crazy contray theories but there was this you know there's been a theory for a while that the world ended in 2012 I don't know if you saw this and then we're now living in matrix. Obviously, I don't believe that. But I do believe that there is there was a shift. >> It could have been technological, could have been this collective neurosur I think it was neurosis of this idea that we are no longer having agency over our own thoughts because we are so plugged into this non-stop information stream that the moment of pause >> just having five minutes to meditate. So I meditate daily but I only really can

afford 5 to 10 minutes but that is more than enough according to my therapist. But the idea is because you can create this peace and calm for your your brain to rest, for the thoughts to rest. Much like how you can't weightlift, you know, non-stop. You have to give yourself breaks. Hit as high intensity as it is still has rest of course >> but we are not giving ourselves those intervals to rest. >> Not our brains. >> Our brains and it will fry our circuits and it's happening already. That's why the number of people that need therapy or have like you know anti-depressants or drugs in general is not a result of uh I think anything but the fact that our brains are being fried by the just everything being non-stop. So I think slow and and and not I'd say like slow and and dedicated is is what is

important for us right now. And we need to be disciplined in that regard. And I unfortunately do not think that when you're young and you're impressionable and you don't understand a world that was before this you have that discipline. >> Yeah. >> That's the fear. Yeah, man. Whenever I find myself in that state, I often just try to go to a place where I can surf. >> And you have that you have that place of peace. And I think I urge everybody to find >> those moments, those places of peace. You don't have to be working 20 hours a day. In fact, in our fund, again, not to do another promo, but in our fund, we actually pay for people's gym memberships and their therapy sessions if they haven't done it before, right?

So, the idea is to turn them on to it, right? And we want a sustainable founder, >> not one that is trying to burn the candle from both ends. And I learned this personally the hard way, right? I was a founder, like I mentioned, uh, two companies. And, you know, it was it was I was destroying my body and that's not sustainable. And what for, right? Like, exactly. There was no actual outcome, even financial, great. But you're trying to You don't want to have all the money in the world and not have the body to enjoy it in, right? And people would always tell you, if I if I gave you $10 million today, but the only requirement is you wouldn't wake up tomorrow morning, >> right?

>> You would never take that 10 million. You want to wake up tomorrow morning, right? Sure. >> And so that tells you how much your your morning is worth. [laughter] >> Ladies and gentlemen, Brian Wong, what a conversation. Got quite philosophical there. Thanks for coming into the studio.