Tom Schmidt of Dragonfly on why decentralized AI compute is overhyped and agent payments are the real crypto x AI thesis

right. Double kill. Two ads in one go. Let's do it. Nice work, John. Let's bring in Tom. Let's bring in Tom. How you doing, Tom? There he is. Good. Nice to meet you, Technology Brothers. Good to see you guys. What's happening? How are you? Hey, not a lot. Doing well. I'm uh I'm loving the suits.

I mean, you guys always bring it, but uh you look great. Have you watched the stream at all? Are we asking too normie of questions or are we hitting the right topics? No, it's the right blend of norminess. There's a lot of stable coin conversation, but that's to be expected. You know, people people the stables.

So, no, it was it's it's been good. It's been great. I mean, my my bare case on stables is I I went super long stable coins and my portfolio is completely flat. And so, I'm looking for alpha. Give me some crazy ideas. What are you excited about right now? That's not one year out, but 10 years out, a 100red years out.

What's the future look like? Break it down. 100 years out. Um, I think we're all going to be dead and replaced by a aen. So, it's all good. I'm not I'm not too concerned about that. Um, you know, on the stable coin thing, I wouldn't sweat it.

I think you're going to get a little uh, you know, circle equity airdrop um during the IPO. They're going to just air aird drop some shares to all the USC holders. We're shilling for the US dollar. We're sh We're US dollar shills. Yeah.

Anytime anytime anytime a company sells like a a traditional crypto asset and buys USDC, they're defending the dollar. They're defending the dollar. Yeah, I I agree. I actually, you know, I think you're talking about um stable coin and people when people talk about stable coins, they talk a lot about flow, right?

They talk about um using stable coins for payments, B2B payments, P2P payments, whatever. Um I think the stock thing is actually underrated. And there's a few people in DC who talk about this.

I think Paul Ryan was actually talking about basically stable coins being a way to um you know mitigate a lot of the US debt because now you have all these new treasury holders. Tether is like the seventh largest Treasury holder at this point.

And I'm like this feels like a very pro- US bipartisan topic of yes, we want people to be buying. We want more marginal treasury buyers. Stable coins provide that route. And it's like a little bit of like a you know saying the quiet part out loud but I don't really care.

I'm you know very pro US and I think like using promoting dollars irony of that digital assets which people saw as alternatives to the dollar are now some of the greatest defenders is kind of to dig into it though is is uh like tether's the seventh largest holder.

Is that because they displace like Fidelity by taking like all of the treasuries and they just move them over or are these can we really think of these as like net new treasuries that are being bought? these these are incremental, right?

Because these are people who are offshore, people who are in Argentina or Turkey certainly maybe some of it is is cannibalistic, but for the most part from what we've seen from and actually um Rob Hadock, who's one of the GPS with me at at Dragonfly, just had um a great stable coin research piece um come out yesterday talking about who's actually using stable coins.

And it's yes, some people in the US and yes, maybe some people who previously would have purchased treasuries through the broker, but actually it's it's people who are offshore. And that is actually I think kind of the the killer app. What is happening in the Asian crypto markets right now?

I know you've spent a ton of time over the years kind of focused in that area. Uh how is the regulatory environment kind of evolving? What what does crypto activity look like? How do we get them to rotate entirely into USDC? Yeah. Yeah, that's part two. Uh well, they're waiting for the airdrop.

Um no I mean Asia and the US have always um sort of I would say specialized when it comes to crypto. Like if you look back the past three four years um Asia has really been running laps around the US when it comes to anything CFI.

Um you know we're early investors in Bybit and Bit which are now two of the largest exchanges incredible businesses. I mean we're talking billions of dollars in in revenue. um they some of the features and some of the products that they offered um US exchanges are just now sort of catching up to.

So I I would cite as one example um a lot of Asian exchanges offered what we now call CD5 of um offering a totally normal centralized exchange experience in the front end but wrapping some sort of D5 protocol on the back end to be able to offer people you know trading for new assets or be offer you lending and financing opportunities or yield opportunities or whatever is actually available on chain but without all the complexity and yiness of you know having to manage your own keys which is kind of what normies want.

Coin Coinbase just launched um you know Bitcoin backed loans using Morpho which is an onchain lending protocol u maybe a few months ago. So this is something where the US is now just kind of playing catch-up to what's been happening in Asia for many many years.

Um I think there's also been this trend of more I would say um Chinese um uh previously like OTO founders moving into crypto and frankly also moving offshore because they see hey venture in China is like kind of dead. Um you know the domestic market is kind of dead but I'm I'm a really great technologist.

I'm a really great builder. Um building in crypto is a way to access a global user base and a global capital base.

um but you know with sort of the skills that I already have and so um you know we're early investors in Kaido um which is sort of this now it's been coined infoi um which I don't I don't quite love but I think it's a cool like concept of sounds um yeah yeah basically scraping uh Twitter data right now um ingesting that and and then sort of spitting that back out into this sort of gamified clout system um they also do some really cool stuff with like transcribing all all these podcasts and just like giving you sort of like an alpha sense for crypto um but again that's like um a lot of Asian talent now now sort of moving into crypto.

Interesting. Uh how are you thinking about the intersection of AI and crypto? Personally, we've talked with a number of people about the potential of stable coins within agents, but there's a bunch of other How often should you pivot from AI to crypto and back?

Is that every three months, every six months, every two years? What are we thinking? Um no, but but in all seriousness, how how are you I mean, I imagine you're you're a consumer crypto investor. You need to get you need to get this bet right. Otherwise, you're going to look like a fool in 10 years.

So, I imag this this is what I have nightmares about. I um I'm sure you do. No, I um I think we we've been maybe a little bit contrarian when it comes to AI crypto in being more bearish on this category. I think a lot of sort of what AI crypto has broken down into is people trying to do decentralized inference.

Um sort of this, you know, GPU marketplace. Um and this is sort of an again an old idea in crypto. This was something that um Gollum way back in the day was trying to offer for for video rendering. Now kind of coming back. Um there's decentralized training, decentralized data marketplaces.

I think for the most part this is just really tricky to build, right? Like if you think about how you actually train a state-of-the-art model, you don't use you know a million consumer grade GPUs all around the world.

It's like you want one big colloccated A100 cluster like you know that is actually kind of uh uh the thing that you want. Um and so like a decentralized network actually is just like it doesn't really have the right characteristics of something you would like the type of compute that you want.

Compute is is not a monolith. Um the areas where I think we've been investing in AI meets crypto in the area that we get really excited about. Um you mentioned sort of um NPC and agent payments. Um I I agree this just seems like kind of a no-brainer where you want deterministic final finality.

you want um you know microp payments something that just like existing um uh SAS companies existing SAS payment companies just aren't really set up to provide um you want to be able to say hey I want like one simple API call for a fraction of a penny and they're never going to use the service again that's something where stable coin payments are uniquely enabled um uh are uniquely positioned to be able to do this but it's just not something that you're normally going to see out of out of traditional payment companies.

So, um, we're investors in in a company called Gold Sky, um, which is which is building in in this in this category. Um, we're also, I mean, kind of on the cute topic, um, we're investors in in Exo, which is is sort of borderline crypto.

It's, um, I don't know if you you've seen them around Twitter, but basically, um, they allow you to run, um, sort of sharted inference locally on your own network. So, let's say you have, you know, a couple Mac studios. Um, they're really sort of specializing Apple Silicon right now.

it will automatically discover and and um execute um let's say something like DeepSec R1 across those different um uh computers on your local network. So you can run really great models on consumer grade hardware um given sort of the existing um shortage of of of GPUs.

And so you can imagine that hey if you can sort of distribute and shard compute across your local network maybe at some point in the future you can actually do that across a global network. Um they're building towards that.

I I I don't know if that's exactly the direction that they want to go, but in the interim, you have this really really useful tool um that allows you to actually get access to um local edge compute models, which I I think ultimately is is where that whole industry is going to go um versus, hey, we're all going to be running um our own models on, you know, someone else's cloud.

Uh back to the venture side, how do you see crypto funds evolving? It seems like there's a number of funds with massive aum and maybe not enough uh places to bet on the equity side.

Do these funds end up having you know you know uh be end up 80% sort of liquid 20% equity uh you know sort of early stage focused or or how do you see that evolving especially um I was talking uh with with Anie uh on your team offline and and he was talking about how uh with AI being so deflationary we just have companies now that have massive potential but just don't really need capital.

So you're sitting in the investor seat being like please take my money versus you know maybe 10 years ago it would have been reversed. Totally.

I think um that is definitely a trend that we see in I don't think it's even necessarily a crypto venture thing specifically but uh I mean you were talking earlier about people throwing you know AR in the bio and it's like yeah that is downstream from just operators having so much leverage now that they didn't have before.

Crypto amplifies that where now you can sort of go public. um you can access a massive user base, you can be way more sort of revenue positive than you were before.

And so maybe the history of of crypto funds, just to sort of take a step back, um if you think of sort of like 2017 2018, for reference, Dragonfly was was started in in 2018, crypto was was was a monolith. You give some some some capital to a fund manager and they're going to decide, okay, I'm going to buy some Bitcoin.

Am I going to buy some ETH? Am I going to invest in a um you know, early stage equity round? Am I going to buy a Saft? whatever that sort of blend was, they're going to be managing that for you.

And oftentimes that was in a true closed-ended venture structure, but sometimes it would be in a liquid hedge fund structure with some side pocket. And it was like the whole thing was was kind of no one really knew what this asset class was going to look like.

And so you give capital to someone else and they figure it out. Um, that was also obviously a a a byproduct of the fact that even getting access to something like Bitcoin previously was so difficult, right?

like you couldn't buy it through your brokerage, your bank, no ETFs, most like you had to use some wacky custodian. The whole process was very convoluted and so great, you managed to do it into a fund. That's something you can underwrite.

Um, and they can actually go and determine how much of the Bitcoin exposure that you want. You you fast forward several years, um, and now, hey, uh, if you're a reasonably sophisticated LP, you can go and choose how much Bitcoin you want to buy yourself. We actually don't buy major liquids in in our fund at all.

And it's like I you don't need to I don't need to charge you 2 and 20 to go buy Bitcoin or ETH or Solar, you know, whatever it actually is. You can go do that for yourself.

And so there's been more specialization in in the crypto fund um um area where now you have true dedicated venture funds like Dragonfly that are kind of doing let's say seed through B um in addition to hey maybe doing some some treasury purchases for some liquid assets that we think have sort of venture upside.

And then you have sort of true dedicated liquid funds. Um, maybe those are, you know, delta neutral funds or credit funds or actually just sort of long only discretionary funds, but there there's sort of this this true split. And I would say it's a little bit like the sharks and the jets.

I think the, you know, the liquid funds always talk [ __ ] about the venture funds and the venture funds I I don't really talk [ __ ] about the liquid funds. I think they're great. Um, but it's it's it's u, you know, there's is sort of the specialization in the strategy.

Maybe to what you know, Anie was was mentioning and what you were mentioning earlier.

Um now the trajectory of fundraising goes hey we're going to raise a preede though maybe we'll raise a seed maybe we'll do an a TBD but generally hey you'll launch a token um token will go public and then maybe you'll do a treasury sale after the fact if you need more capital and so if you've raised you know north of a billion dollars to your point where do you actually go and deploy that um maybe you there are some equity only companies um and and certainly we we do a number of those where we think hey there's this company's gonna IPO um you know, we don't think there's going to be a token.

Um it can be uh just just a pure sort of re revenue generating company and I think that's like um actually kind of an underrated area of of the space.

Um but it's sort of unclear where the rest of that capital goes which is why you sometimes see these very wacky rounds where some company raises some insane amount of money and you're like what is going on here?

Um and I think there's just this um lag effect between you know capital raised by VCs and then capital actually you know deployed into startups. What what keeps you up at night about the industry right now? Potential risks.

It feels like the industry broadly, you know, this year has been up and down, but so has every year in crypto to some degree. Uh, but generally the industry is sort of like riding on a high, getting more regulatory clarity.

Uh, there hasn't been, you know, a 20 billion hack in a while, but like what what kind of what what kind of risk do you see? um at a at a high level that that that actually worry you? Yeah. Um a million things honestly.

Uh it's it's an industry that is always uh changing and and therefore there's always more things to be worried about but also things to be you know excited about at the same time. Um maybe to your point around um regulatory clarity.

I'm more worried that we we don't get regulatory clarity and we fail to get a stable coin bill passed. We fail to get a market structure bill passed. Um, both of those are in the short term not terribly bad.

I mean, I would say it's impressive how well stable coins have have grown in spite of being in this sort of regulatory gray area.

In the long term, the question is, can we get a digital version of cash that is they can be they can have have privacy for peer-to-peer transactions or or do we get true ponopticon status centralized control? um that's something that is going to be bad on, you know, a 10 to 100y year time horizon.

M maybe not so short-term bad, but we need to sort of lay the foundation for that now so we don't end up in that bad situation. Um I I'm also frankly maybe a little bit worried about um a lot of these sort of uh Bitcoin treasury structured companies.

I I think there's only so much market demand for these actual convertible assets and I'm like it's great when number go up. I'm I'm a little bit wor worried about what's happening when when the number goes down. Let's just give it up for when the number goes up. Yeah. Yeah. Just stop right there.

You don't need to talk about when the number go down. We want to cut you off right now. That never even happened. So, let's let's not worry about that. But, um overall, I mean, it is a great time in the industry. It feels like um like you're saying, it just never dies.

Every time someone thinks it's over, that's actually the time that we're so back. that is that is the exact time to be back and um if it you know it feels like we're we're back now in in a big way. So um yeah, we're we're really excited about where the future of the industry is going. Well, thank you for hopping on.

This is fantastic. Thank you for coming on. Come back on again soon. I think you might be a generational yapper. You're one of us. Um you know, we live to post, we die to post. It's uh it's all part of the same the same life cycle. So, thanks for having me on. Thanks for coming on. Good to see you, Tom. Cheers.

Uh in the meantime, let me tell you about Linear. Linear is a purpose-built tool for planning and building products. John for modern modern software development, streamline issues, projects, and product road maps. I would bet that every founder that is coming on the show today already uses linear. Don't say that.

They're all about to convert with our coupon code. We don't have a code. We don't have a code. Linear linear is going to eventually get to 105% market penetration because certain companies will have two instances running. That's ideal. Yeah, that's our goal. Totally dominate.

Yeah, we're trying to capture 110% of the Yeah, we'll go into that more later. Uh, next up we have Constantine from Block Demon here. Block Dam, how you doing? Welcome to the