Hivemind: Can Bitcoin Reclaim $80k, State of The Market & The AI Trade

Nothing said on The Hive Mind is a recommendation to buy or sell securities or tokens. This podcast is strictly for informational purposes only, and any views expressed by anyone on the show are solely our own opinions, not financial advice.

Hello everybody, welcome back to another episode of The Hive Mind Podcast. I am your host, Cedrus. I'm head of research at Delphi. Um, we've got José back, who's been out for a while. Uh, of course, José is, uh, head of Delphi Labs. Um, and then we got Yan, who's a venture partner, and Jason Yanowitz, who is head of markets at Delphi Research. So we're recording this on Wednesday, March 3rd, 10:30 AM. Bitcoin is at $73K. So we'll probably top tick this, but like it's the first— the last couple of days it's been pretty strong. And even on like the down days, it was kind of chopping mostly as stocks sold off a lot. So, and this is with all the war stuff going on, you know, of course, the US launched a strike on the 28th. Which was last, which was on Saturday, I believe. So we're like a few days into that now. But yeah, you've seen pretty good. Hasn't just been Bitcoin either. You've seen a bunch of alts across the board do well. It's been pretty broad in general, at least regards to like the majors and stuff. So yeah, I don't know, how are people feeling about the market right now?

Sure, I guess I'll start. Sorry, my cat is just freaking out, running on the thing over there. Yeah, no, I mean, I'm happy with the— I'm happy with the reaction crypto has had, right? Like, generally speaking, if you had like a day like yesterday in equities where we were down, I don't know, like 1.5 to 2%. Generally Bitcoin and crypto has just died, right? Like minus 5, minus 7%. And it's kind of like the first time in the last couple months you've seen crypto kind of have a good reaction to bad weekend news and then kind of hold strong while equities decided to not do too well at the, you know, at the open of the week and stuff. So like all of the, like, I guess qualitative signs you would want to see for like, you know, some kind of appetite returning back. Um, is there. So like, you know, it's, it's definitely constructive, like market structure-wise. You're breaking some, some, some important levels, so you want to, you know, see that hold and not immediately just retrace in the next day or so. Um, I did read something this morning that came across my like little brief that I have set up. Actually, this one came through Telegram from Scott, but, um, you're starting to see like, uh, kind of, kind of like these, you know, correction technical signals start triggering. Not like they're not saying that like the, the equity market or whatever is like going to, you know, pull back or, or go through like a mini correction. It's just saying that it's seeing some, you know, stresses in, um, credit and, and, and vol spreads and things like that. Um, uh, so just kind of like an early warning of like, hey, like, you know, if, if these shocks, you know, persist for you know, a, a, you know, prolonged period of time as opposed to maybe like a week or two, like you could start to see some, some stresses manifest in, in equity markets. And this, this, these types of signals, I guess like they're, they're, they're better when there isn't like an external shock, like an Iran war type thing. It's like better when it's all like internal market stresses signaling these things. So like that's why you kind of want to wait and see how, how this conflict sort of plays out over the next week. But like, I don't know, I'm— seems like equities, seems like crypto, seems like everything's kind of just like not really giving too much of a shit.

Asian markets got kind of destroyed, like, uh, Chinese stock market and Asian markets generally took a hit. But yeah, US—

Truck and Miller's cooked.

Yeah, Korea.

I think Korea was down like 10%, right?

I'm sure he bought in at a good entry and obviously, yeah, he's still one of the GOATs. But no, it is pretty funny. I think shortly after that, it did like double circuit breaker.

I mean, the Korean market is up like 2.5x or something over the last, uh, like year or 18 months, something like that.

The Koreans, no matter which financial market, the Koreans just like, they just buy. AI.

Yeah.

Yeah. I mean, it's interesting there. I think Korea and Japan are two of the only economies that, that really benefit from like everything that's happening in AI, uh, just because of how they're positioned in terms of like memory and, and, and like just high-tech components that go into the AI stuff.

Yeah. But they don't have the energy, right? And that's why, uh, well, Korea, I guess, I think Korea gets what all its exports, like most of it from Qatar, I think. And isn't that like the reason why Korea sold off over the last day?

Yeah, I have no idea, honestly.

I'm pretty sure that's why, but yeah, I don't— don't quote me on that.

Yeah, I have no idea.

So yeah, I don't know. I've, I've, I definitely took some, some longs, um, over the weekend when I started to see what the, the reaction was to everything. And, you know, I'm just kind of watching. I don't know, I'm cautiously optimistic. But, um, even, even without that, like, you know, short-term stress signal starting to, to kind of pop off, equities were losing momentum for, for a while. So it's still something I'm, I'm generally keeping an eye on. VIX is elevated. So like, if we get like a VIX blowout, like an actual one, like, I'd probably just like shove most of the cash I have into, uh, into risk assets because I think like the general backdrop still is, is relatively strong. Um, so if we got some kind of, you know, opportunity like that, I would, I would fucking shove probably.

It's just, it's just crazy how many stocks have got killed and the index is only like 2% off.

Yeah, right.

Yeah.

And, and that's kind of like what that, that correction signal thing I was talking about, it kind of like measures like stresses and credit spreads and volatility with respect to where the market is trading. And since the market's trading effectively, you know, within a percent or two of all-time high. Um, they, they tend to be, you know, pretty strong signals, um, when, when they do start popping off. So something I'm keeping an eye on and we'll probably write about, um, in the next day or so.

Yeah, it is unique though in that, um, you know, they, they, the amount that they drove the market is so massive and their slowdown isn't necessarily, uh, a function of like slowing GDP and growth and there being kind of economic weakness, but rather just the sustainability of the spend and the, you know, maybe we need to be discounting future cash flows a bit more if like there are delays and, you know, basically the debt schedule doesn't align with revenues and all the leverage that was built up to kind of fund this build out. So people kind of got worried and just the unknown unknowns of AI. Um, so that they're like, you, you do have this dynamic where, you know, you do have the leaders falling, but you also have a setup where you don't necessarily need— there doesn't need to be this massive follow-through, uh, all the way through the market because it seems pretty contained to those names for specific reasons.

Yeah, I guess, um, yeah, I guess unlike the Iran war specifically, this has been something that we've been talking about for at least like 2 years now. Um, so if it does, if we do get a resolution relatively quickly, you gotta expect that to be a pretty big, uh, tailwind. And I feel like that's kind of why Bitcoin has been reacting. The way it has, not that like war is good, but that people see a potential end to this. Um, but yeah, I guess we'll see how the next like week or two goes.

Uh, I mean, the effectiveness of the campaign has been nuts. I think the other element is, is, you know, there's always a concern around the, the closing of the Strait of Hormuz. And, um, Trump came out and basically said they're gonna be patrolling it and trying to create as little disruption as possible. But it's still a pretty delicate balance and any disruption is definitely going to kind of reverberate through the global economy. But I think that's probably helping it a bit where it's not this, I think that the tail scenarios have kind of come in a bit.

Yeah.

Do we want to keep talking about markets or I know José, you want to talk about AI. Maybe we want to just like go talk about AI from here.

I think there's some value in, in like crypto AI tokens that have done well too. If we wanted to stay on the markets topic or we can pivot.

Yeah.

Yeah. I mean, I think it's mostly, it's mostly Venice and Near, right?

Grass.

Grass, right? Virtuals.

So pretty much.

Virtuals.

All right.

Yeah. So Venice is because Venice is like, it's like this private model and then People have been using it with OpenClaw, right? And so that's why it's kind of got bid. And then they have like a dual token model where if you buy DM, it equals like $1 of compute a day. And I think that went from like $100 to like $680. And then like you stake it and then you get the compute, I guess, that you need. And then VVV, I guess, is the token that gets all the income and everything. Uh, I wasn't really paying attent— that much attention to Venice over like the past year. I know it launched like over a year ago. It launched right at the basically pico top of AI agent season. Um, it was like a pretty good airdrop for most people that were like using Base and had a lot of these agent tokens and everything. Um, I also think like, did you see the, uh, I'm not sure how it's done since then, but the OpenClaw, like, the OpenClaw founder hates, like, he hates crypto, like, and I think he, like, specifically, like, blocked Venice, or maybe just took it from the default.

It was listed as the primary and, and go-to, and so that created, you know, obviously a lot of excitement. There's a couple things, so there was the privacy narrative and that, which were really driving the price action. And then you had kind of FUD on both sides where people are coming out and trying and saying it actually isn't private. And so there's quite a bit of back and forth. I don't really know where it's settled out, but it's, I guess it's worth noting that the people that are coming out against it are kind of building competing products. But yeah, that part hasn't really been resolved. And then the Open Claw element, he just basically removed it as the default, but just listed it as one of the many options basically. And his argument was just that he wanted to remain neutral, is a Twitter comment the Openfall founder made. And so the idea was just to not bias any individual one and rather just leave it to the user, which I think is pretty fair. And isn't— it's like a downgrade in terms of the narrative for the token, but it's not necessarily a net I just appear negative in that scenario. Yeah, the DM stuff is interesting, right? If it's a dollar of compute a day, you know, at $120, um, dollars, it pays for itself in, in 4 months. And then obviously as it really balloons, it makes that payback period quite a bit longer. Um, and then you, you basically deposit VVV and mint DM, uh, that allows you to mint it. So you could buy it directly or mint it. And so there's an interesting kind of relationship there where as demand for free usage goes up, it's, you know, you basically, it's easier to mint it than it is to buy it. And so that drives demand for VVV and kind of creates a loop.

But how do you— wait, wait, how do you mint it though? It's like, I'm assuming it's not like every VVV has like a certain amount they can mint in their life. It's like you need to stake VVV for a certain amount of time or something.

Yeah.

You stake it.

And then the minting function is based on the supply of Diem. So as the amount of, they target a specific amount and as the quantity goes up, you mint less and less. And then as the quantity of it comes down, you're able to mint more. So it's this dynamic mint rate that's based on kind of the amount of Diem that exists.

Okay. And then what's the incentive for like burning Diem?

I think to unlock your VVV, basically.

Oh, okay. Okay. So your VVV, I see. Your VV is like locked. So yeah, so you could like mint it and then sell it, but then it rips and then you're kind of like, oh shit, it's gonna cost me way too much to get back my VV.

Yeah. I mean, it definitely has a bit of a natural ceiling because of the payback period, right? Like, well, at that point you'd just rather pay for the compute directly.

Yeah. Okay. Yeah. I've seen a lot of people using it. We have like an OpenClaw chat with a bunch of people at Delphi. See, a lot of people have been using it recently, which is cool. It's like, I don't know, it's actually— AI and crypto is definitely— I like, everyone's soured on it over the past year or so, uh, because agents like pumped and dumped. But it's still obviously kind of like the area that makes the most sense. Um, even like I was talking like with regards— and we'll stay on this topic, but just because we're on it, like this whole solar founder era, right? Where you've never, for the first time really, it feels like you can have a solar founder create a product that's like 10 mil, 100 mil ARR.

Yeah.

And that's like, you haven't seen that though. We haven't, we haven't, but that's also perfect. Like that's really literally perfect for some of these, like, like MetaDAO, Soar. Um, I think there's like a tokenized equity one, like these platforms where those, those kind of founders are not venture-backable, right? Um, maybe, maybe, but like for the most part they're not venture-backable. They're like, it's one guy, it changes too quick, all this stuff. They don't need to raise a ton of money. Like, it really does seem like this is like— capital formation has always been crypto's killer use case. And it's like, I mean, I know we've been saying stuff like this forever, but it Sometimes stuff just takes time and it really feels like this is such a big opportunity for crypto.

Yeah, I really like, um, did any of you guys read the 2025 like Ribbit report on tokenization?

Yeah, the token report.

Yeah, I really liked it cuz, uh, it was like quite elegant the way it basically described it. It was a report on tokenization which covered both AI tokenization, so like LLM word tokens, and then also crypto tokenization. And the way it defined tokenization was tokenization is making the world legible to, for computers, which I think is like very true and very like elegantly applies to both kind of, right? Like LLM word tokens make the world like literally legible to AI models, right? And then crypto tokenization, I think makes, um, like capital to some extent Uh, like legible for computers. And in the case of crypto, I kind of change it to making the world legible to computers to make it legible to capital. And I think it's like an interesting way to think about it and to think about what it makes sense to, to tokenize. Um, yeah, cuz we've kind of seen it with crypto, right? Like meme coins kind of make this, this like culture or belief or, or attention or whatever. Suddenly tradable because they make it legible, right? They turn like a dog meme into something that's discrete, priceable, transferable, you know, enforceable, and now you can bet on it. And then like governance tokens, you kind of make power over an open source system like legible and now you can bet on it, right? Like you can control the parameters of an open source system or get fees from it or whatever. And I think you can go down the list. Yeah, I just think, and I think AI tokens to some extent, yeah, I think like agents are like a reflection of this and Hasee wrote that post and I think A16z has a good one too of just like how these rails were actually made for machines rather than humans. And I do think if an AI wants to be truly sort of sovereign, like it just makes a lot, like you, you can deploy a smart contract without needing an API key or without interacting with any centralized services, right? Like all you need is to have a wallet and like a, a server to some extent. And you're already seeing these like emergent behaviors with like Open Claw agents. So I do think the thing we were all bullish on a year and a half ago, whatever, is gonna, is gonna happen. But I think it hasn't happened yet. Like you, you kind of know That's one thing I've kind of learned in, in venture. I think you, you kind of know when something has reached the, like, like that's happening, right? Like you, you haven't had like the OpenAI, the ChatGPT moment for this, I don't think. And I don't know what it will be like, like what the use case is that actually drives this, this stuff. Like it's hard to, hard to say, hard to say right now, but I do think it's gonna happen and they're gonna be interesting plays.

But I'm not sure if any idea on what it might look like since you probably got your claws in a bunch of, uh, a bunch of this AI stuff.

Uh, um, no, honestly, I have no idea what the crypto, what the first like crypto AI use cases will be like. 'Cause I, I think agents can use crypto rails, you know, already.

Um, well, I mean, Venice is like. That's something.

It's interesting, but I don't get what the private— do you guys— did you guys look?

I don't know what the privacy comes from. I don't know where the privacy comes from though.

Yeah, there's like different definitions of it. If it's a mixer, if it's some degree of obfuscation, they, they chat, they, they mentioned way back in early '25 where they were trying to do, uh, to encrypt the kind of the prompts and the outputs. I don't know if they've gotten there yet.

Yeah. I don't know what the, where the pri— because the only reason to use this is privacy, right? Like, because Cursor is not just a model aggregator. Like, they have their own models that do a bunch of interesting stuff.

Well, I mean, like, the token model is like something kind of interesting, like owning the, the econ— like VVV, I'm assuming owns, gets like revenue share from people paying for compute. I mean, that's, that's, I don't know, that's cool. And it's pretty open source, right? So, but yeah, I know what you're saying.

There has, there needs to be a fundamental reason still to use VVV. And it's also like, it's still not on-chain, right? It's like, um, You buy this stuff on-chain, but then there's a centralized service that you're reliant on to like, you know, host the models and honor this conversion and stuff. So I don't know. It doesn't, yeah, I guess no one really cares once these things start going up.

Yeah, I don't necessarily know that it needs to be on-chain though.

I think that like the point, like crypto helping AI here is for really like the capital formation and like that whole side of it. I don't know if like things are going to be directly on-chain outside of like the token and like the economics of that and everything.

I also think like USDAI is pretty interesting on that front where basically—

we're investors, by the way.

Yeah.

Asset-backed loans.

Right.

And so you create liquidity for these assets where otherwise it's locked. So the ability to kind of borrow against your hardware to unlock additional liquidity.

I don't know. Near is also like, I think what they're building is pretty interesting. And I think people, I saw a lot of people hating on them on the timeline for like how often they change their narrative or whatever, going back to algo stablecoins back in the day when they had their NearUSD. I think that's just like what you're supposed to do. To try and find product-market fit as an L1, you know. And I think with the AI stuff, like, uh, they've always been thinking— they've been, they've been building this for like years, you know, like the, the same narrative of like user-owned AI and what you need in terms of the stack to do that, privacy and stuff like this. And I know, uh, you know, I, I, I, having spoken to Ilya, like, he definitely has a lot of opportunities to not be doing— like, a lot of founders in crypto slow rug or, or, or, or just, just literally rug with a lot less opportunity cost than what Ilya has being, he's like super deep into AI and one of the, you know, one of the authors on the, on the original Transformers paper. And he's like consistently just kept shipping on Near, kept building cool stuff. So I don't know, he's, he's definitely one of the most credible, like, well, he's definitely the most credible crypto AI founder in, in, in the space. I think, like, uh, and yeah, like a lot of the stuff he's doing makes a lot of sense, right? And you're kind of seeing it now, like you do want, and if this TE stuff works, I do think there's going to be like some, some, some pretty serious demand for it. Um, yeah.

So no, I, yeah, like I agree. I don't think that, I mean, I used to make fun of all their pivots too, especially when they did Near DA. I was like, okay, come on guys. But, um, it does kind of highlight that they're able to like ship like actual products pretty quickly and they're pretty adaptable and like, um, like Near Intense. Okay. They, he basically got pilled on Intense by, uh, Adrian and Enoma and was like, oh, this is actually pretty cool. Let's do this. And now Near Intense does like all the Zcash volume, does a bunch of other stuff. And so they continue just to get shots on goal. And then once they get products that work, like Near Intents, now they're doing other stuff too, right? And like they keep doing this, they can build a pretty big suite. I've been impressed with like, I've never really held Near. I've held it like a couple times in the past and it never worked out. But like I do respect their, they are like technologists too, and I do respect not actually launching like multiple tokens for all these products and everything, right? And continuing to do it under the NEAR umbrella and trying stuff. And their new thing is this, uh, it's like Ironclaw, right? So what is it like OpenClaw with TEs, right? Is that essentially what it is?

Yeah.

Private. And that's a good, that's another good shot on goal, right?

Like, and even now we just with Intents, like they're, they're, they're now slightly deflationary. Cause they burn a portion of the intense revenue and they have some emissions for network security and they're actually slightly inflationary, which like is pretty unheard of for an L1 that's not kind of predicated on token transfers. It's like a business model in the L1.

Yeah, it's definitely worth looking into them deeper. I know I saw I saw a bunch of TAO guys on the timeline angry that people were trying to make Near AI the, the AI play.

But I mean, that might be why they're so hated though, right? Is 'cause they've just like jumped in on narratives and then the existing ones are just like, hey man, stay in your lane.

Yeah. Like, no, even like Intents, like Anoma was so early to Intents and they were right about that becoming such like a dominant way of using blockchains. but then Near just came and was like, oh, we're going to do, uh, all this Zcash volume. And they took it and it's like, you know, as an Enoma investor, it's kind of like, damn, that sucks. Cause like, yeah, they were, they, that would be, that's such a, Enoma's whole thing was always privacy and intense and everything. And it's like being the Zcash trading layer is—

what's up with Enoma actually?

What's, what's Enoma up to?

They're launching, they launched, they're launching their router on a bunch of the EVM chains in Solana. I'm not sure when yet, but then you'll be able to do like all the cross-chain intents and everything across all those chains. And then, so it will be a competitor to Near Intents. And then they also like, Anoma has always been big on the privacy, right? I'm not sure, like trading privacy, not just like trading Zcash, right? Um, so they'll definitely be a competitor to Near, to Near Intense. Um, I know that they were also working on some in-house products that I'm not sure the status of, um, right now. But, uh, yeah, they're still, I mean, they're still, yeah, they're still launching a bunch of their stuff soon.

Jose, how are you? Uh, I know you mentioned this before we, we started, and while we're kind of still on the, you know, AI topic of things, how are you, uh, positioned nowadays? I know, uh, you've made some moves over the last couple months, um, and as somebody who, uh, doesn't have enough AI exposure, this is kind of a selfish question here, so please spill the beans.

Yeah, um, I mean, initially it was, I was pretty overweight, like the, basically just the hyper, like I, I, I can think we talked about it actually on this, on this podcast a bunch of times. I was just basically overweight the hyperscalers, both the US ones and the Chinese ones. So like Nvidia, Microsoft, you know, Google, Meta, this, this kind of thing. And then Baidu, Tencent, and a few of those names. I think over the last few months I kind of started seeing, like, I don't know, I used Copilot the other day. And it's, it's, it's like shockingly terrible. Like I use it because actually my, my, one of my, uh, uncles is like, that's his experience of AI. And I was like, dude, yeah, this, this, yeah. And, and I just think that, that in general, the, the hyperscalers have kind of like, other than Google, have failed to execute on AI, like at the model or application layer, right? They're just owning the compute, like, you know, they don't have like, and, and it, and it's kind of like all their businesses. I'm, I'm, I'm not sure, like, for instance, like Office is still gonna be a thing in an AI era. Like, are you still gonna do PowerPoint and, and like literally AI can just code you up a, a presentation in HTML as a website, right? And like Word and, and all this stuff. I just don't know that these things will, will, will still be like going concerns. Um, and I feel like they just end up in the same with like Amazon. Like, I'm not sure you keep, you're going to keep, uh, ordering things on Amazon. You know, I think like agents are just going to order directly and have access to like all this information. Um, so then it's like they basically are like compute businesses, right? Like that's the biggest thing that keeps accelerating. And I think it will keep accelerating because they, they, they basically have like own most of the, of the compute. Um, but at some point, if, if the application and model companies are making all the money, they own the user, they're just going to like vertically integrate up the compute stack, you know, kind of like what, what, uh, what, um, Grok is doing with, with SpaceX. Like you'd imagine they just, they just own the compute themselves. And so I think it'll, it'll, the margins will go down like quite a lot on the, on the cloud businesses over time. I don't think it'll happen yet. Like I think. Over the next 12 months, you're still going to see like revenue accelerating and stuff like that. But at some point, I think the market starts repricing them, um, where they're not these like monopolies of, you know, these massive like monopoly businesses they have in software, but actually just like hardware businesses with not much differentiation, like very price sensitive. Um, and, and like, yeah, so, and, and I also think a lot of stuff's going to happen on-prem too with AI, like not just with, with, uh, you know, people are gonna run like these open source models locally too. So yeah, all in all, I'm just kind of like not as like initially it was, it was like these guys have the talent, they have the data, they have the capital, they have the distribution. It just seems like they're gonna crush everyone. But now the model companies have out-executed them so badly. Um, I'm seeing so much stuff in private markets that's, that's going after like all the different parts of their business. Um, that it just feels like you have to kind of go, I up the stack to some extent. And so the way I'm playing it is just, um, you know, by all, like, if you look at Epoch AI or whatever, it seems like the next big bottleneck is, is going to be like energy. And I, I'm like very AGI-pilled, so I think we're going to have like compute build out for 5 years. I don't think that slows down any, anytime soon. And so just thinking about all the things that, that need to be, that need to, to happen for us to have this compute buildout that would, you know, that I think is going to happen. It's like one of them is energy, right? Like, uh, the gas turbine companies, like I have like Siemens Energy and I have a, you know, a company called Bloom and a few others where it's just like they basically produce gas turbines. Uh, their revenue is like inflecting. They have massive waitlists and like gas turbines are the way to scale energy just because you don't rely on the grid. You can actually like produce it yourself. And then I have like Constellation Energy, a bunch of others. Then I think the infrastructure companies are interesting, are interesting too. Like Constellation Energy would be in there. I think some people would go long CoreWeave and things like that, or like some of the suppliers, like the cable, for instance. Like I know people have done that, like the electrical cable that you need to build out this stuff. And then just like commodities needed for for chips and electrification. The biggest one's obviously copper, but then like rare earths and minerals as well. And then yeah, I have some other stuff. I have some, I have a lot of gold. And I have some defense, like especially European defense, because I think that's like up only from here. Like after what Trump's been doing, I think it's kind of, yeah, Europe, no matter what they, they, they think they have to rearm, they've committed like €800 billion by 2030. In rearmament. So I'm just, yeah, I'm, I think this stuff, yeah, it's, it's sort of not as exciting as crypto. I don't think it's gonna have as big upside as, as crypto, but it seems like, yeah, it's just the only way I could play that, that makes sense to play like the, the public market stuff, at least for me, for me right now. Yeah. And then I'm, I'm still very bullish on, on China too. So I'm expand, like we're, I'm about to go there for, for a few weeks, so I'm going to be meeting with a lot of like founders and actually some public company CEOs and try and figure out like how to play that. Cause I think it's really interesting what's happening there. Yeah, that's basically how I'm thinking about it, but curious how others are playing it if you are.

No, that was great. I'm still trying, like I said at the beginning, that was it. A fairly selfish question as I'm trying to put together all of, all of the moving parts. Obviously, I think that the Europe defense allocation makes a lot of sense, whether or not. Right. That's, that's what Trump wanted or not. Who cares?

Right.

Like, that's what they have to do given everything that's happened. So it makes a lot of sense. Yeah. I mean, I need to look into a lot of the, uh, energy and, and other AI adjacent type things you were, you were mentioning outside of the hyperscalers. Because last year I really didn't do a good job, uh, of, of getting a ton of exposure during the tariff tantrum. And honestly, that was probably my biggest, um, fuck up last year for sure. Um, so I'm trying to play catch up and trying to figure out like what the, what the ideal, like, I don't know, thesis or allocation looks like today, uh, without any of that foundational portfolio positions from last year. Um, gold also has been something I've been buying a lot of as well. Um, just kind of collecting different little gold bar things. Um, and yeah, just actual physical gold, not like GLD or any of that bullshit, like actual gold. Um, yeah, and then crypto, like I I'm actually fairly bullish Bitcoin. Like I, I bought, um, like I said, I bought some over the weekend. Um, I think, I think if things kind of keep going the way they're going, Bitcoin specifically makes a, makes a, another kind of resurgence with, with institutional capital. Um, so yeah, I don't know, similar. I just don't have the AI sleeve or allocation yet really. And It's kind of killing me inside.

I definitely think international will continue to do well. Like you're also hearing people like international. I'm like a lazy stock investor. I just go in the global index basically. Cause like I don't have time to like, I spend too much time in crypto that I'm not going to dilute myself into thinking that I can like play all this. But international has now been outperforming for like close to a year and a half and people are taking it more seriously, you know? And the more it's kind of like the longer it's like kind of Lindy in a sense where like the longer international continues to do well, the more and more people are going to start allocating to it, which will like in turn cause it to do well. And like, if the dollar is going to also just get weaker over the next few years. That again, just boost international returns from like the currency gains, right? So, and then to your point, like the defense spending, they, I remember there was like March or April of last year when Germany, I think it was, came out with that massive like defense spending announcement. And that basically since then has like set off a lot.

Uh, yeah, the index has 2x'd, the European Defense Index.

Trump's also forcing it, saying everyone needs to up to 5%, uh, for NATO spend.

I think it's going to happen. Uh, it also creates a lot of private market, like, uh, opportunities too, because like, yeah, private, privates in defense are just like inflecting revenue like crazy right now.

Like, what do you mean in the privates?

Like, what do you What are you kind of talking about?

Startups, basically. Yeah. Um, are like, a lot of them are, are just inflecting like, like crazy. Uh, we've also been spending a bunch of time with like some of the Ukraine drone startups, like super interesting stuff happening.

Yeah, drones, drones seem like a pretty good— like, I don't know what multiples they trade at. Um, so obviously valuation matters, but that's like such an obvious defense sector that is kind of just up only in terms of like usage, right?

In general, I've been trying to simplify it. Like the way I think about my portfolio is I think there are like 3 important trends happening in the world in my mind. The first one is like AGI is coming. To me, that's like the dominant trend, like the thing that's most important that's happening in the world. Cause I think it's gonna drive the most change, the most capital, the most everything, and it has a bunch of offshoots. The second one is the world is moving from, from like unipolar to multipolar, I think with, you know, the rise of China, you know, what Trump's doing, forcing Europe to rearm and, and, and just become more independent, like all this kind of thing. And then the third is like currency debasement and debt and just like fiat going to zero. And I basically think like I want to have a portfolio that like I'm basically trying to construct an index. That, that is like levered on those three trends rather than just a world index, if that makes sense. That's how I think about it. Like just, and so yeah, I think Bitcoin and crypto has a place there. And then I think the other stuff I was talking about is, has a place there. And I do think private markets, so like the stuff that venture investors invest in and to some extent private equity is going to have a crazy like a few years too, just because of the amount of like displacement and innovation that's happening. And, and it's so hard to, to pivot an existing company to adjust to it. Like we, we've kind of even seen it at, at Delphi to a, to a smaller degree. Like we've been trying to, obviously over the last 3 years we've been trying to get more and more active in AI and, and, uh, you know, get everyone clued up on this stuff. And it's, it's hard, right? Like it's, it's hard to get people to, to research these things, to use the new tools, to, to, to learn something completely new and outside of your comfort zone. Um, and that's, and we're like our, you know, we're like 70 people or something, right? And if you think about like a thousand or a few thousand or tens of thousand company, it's, I just, it would be such a hard task to, to get it to, to pivot. And I think some companies are just cooked. Like India, for instance, has a $300 billion. I was researching this the other day. It's a $300 billion IT outsourcing sector.

Yeah.

And that, that to me just seems, it's like, it's 8% of the economy, you know? And that to me just seems like it's gonna get like, it's gonna get decimated, right? Like you just don't—

and quickly.

Yeah.

Like it just happens.

Yeah.

The crazy thing is Q4, it was up like 8%, you know? So these things just happen like, and like Google search revenue is still up, like, uh, you know, so these things just happen slower than you think when you're at the edge, but like still, it just seems inevitable that. So even I was looking at Leopold Aschbrenner's portfolio. He like, he shorted Infosys this quarter, which is like the biggest India IT outsourcing company. Like, I think that's a really interesting, I think like in general, it's like the return of active management and like hedge funds and like people who are, who have their finger on the pulse right now in public markets are just going to print like long short, I think is going to have like a resurgence, you know, before you could just like with passive and stuff, it was just everything was up only you just got cooked on shorts. I think you're going to be able to make a killing on long-short and just like analysis and all that shit again in this transition.

Oh yeah, definitely. This is like the active manager era. I think it's really, I think it's really, we said at the beginning, like a lot of stocks are getting hammered, but the index is close to all-time highs. Like you have seen all like all this rotation. It is very, I think it's a, I think most people are kind of excited by the prospect of this new market environment that we just haven't had for like 10 years or so. Right. I think your point of like companies using AI and all this, um, like we've definitely noticed an increase over the past, probably like just like 2, 3 months. Like a lot of people internally are using OpenClaw now. People are like vibe coding their own things. And, uh, you know, Neil's been really pushing a lot of, uh, AI stuff internally and like we still feel behind, but then it's like, I think if you think about a lot of the, the, the banks and like financial firms, they're still not even allowed to use ChatGPT. For like compliance reasons, you know. And so there is a part that the usage is like accelerating, but then there is another part that is just lagging, um, mostly just due to like legal or like compliance or just slow-moving corporations.

Yeah, those compliance things are, are kind of both features and bugs, right? They're like, on one hand they will slow them down, but on the other they are, are one of the moats in AI that, that can kind of insulate those businesses to some degree. Not all of them, but there's certainly some regulatory moats and, and, and kind of compliance issues. And so like, it, it does on one hand stop them from using it, but on the other hand, it is a barrier against their disruption.

Yeah, I think fintech and like bio and just hardware are the things that are like more defended just because they have like regulatory moats. And then just like hardware modes as well, at least for now. But yeah, it's just coming after.

It's just crazy how fast things change too. Like everybody was shitting on Apple for the past like year about how they were just so behind in AI and everything. And now I've seen tons of people talking about Apple as like a really good hardware.

Yes.

It's like that Homelander meme where, where he killed a guy and everyone's celebrating. It's just like Apple spending zero on hyperscaler buildout and it turns out to be be the right thing to sit on all that cash.

Well, cause it's like, I think it was the pod with the Anthropic CEO who basically talked about like all these people committing, um, to compute. It's like if your estimate is off by even like 5%, you can totally bankrupt your business, right?

I didn't follow that maths. Um, I need to go chat to—

basically the amount of leverage you need to put on relative to the amount of, um, revenue you generate from all that buildout.

But why? Because you're, you're, you're pre-committing in advance, but it's like a debt covenant or something.

And you're just levering up as well. Like it's not a problem now, but it is a problem later on as, as you get more levered, you're, you're just the, you know, you're sensitive.

But they're just raising equity, right? They're just raising equity.

No, a bunch pivoted to debt.

Yeah. Let's say like you commit to, to, trillion of compute and then you only make like $800 billion. That's like a massive shortfall. Yeah.

Right. Yeah.

It started with, uh, Larry El— why am I blanking on Larry Ellison's, uh, Oracle. Oracle. They were like one of the, the, the early ones that, that pivoted and that was when you had the first big tantrum.

But the debt is on Oracle, right? That's the, that's the thing. Like it's not, they're taking the credit risk for, for ChatGPT, for OpenAI, not, not OpenAI itself.

Yeah, but it, it, it relates to a bunch of these other builders where it's like if, if you don't, the revenue just, just doesn't flow through to the degree that you think it will. 'Cause it is all kind of intertwined.

Mm-hmm.

Yeah. Yeah, I guess so. Yeah, the models are looking, are looking great now. Hey, like people, people weren't sure whether these things would be good businesses. They're looking, they're looking like they might be the, sometimes I think they might be the only businesses, like the only software businesses at least left. Like, yeah.

I'm not totally convinced that software just goes to zero.

I don't think so either.

I don't think so either. Like, like a lot of the, think about it, uh, like these software companies are going to be using AI to make their software better too, right? And then you get these contracts locked in. There's definitely, there's definitely a lot of, there's going to be a lot of like compression and there's definitely a lot. More competition now, but I don't know. I'm not totally convinced that like all these software companies—

it's kind of just been an overreaction. Like Claude puts out a bot for this, this sector is dead. Claude puts out a bot for this, all this sector is dead.

I mean, the SaaS index was down like 80% at one point. It was like fucking crazy. I was like, man, if you had good agents, if you had like some, yeah, if you're like a hedge fund, you can go look at the individual names and see, cause some of them are going to probably get better with AI, right? Like more profitable. They have the relationships. They can add like agents on top. Like even what Salesforce is doing with their like sales agents, I think is pretty interesting. You know, if they can literally become—

the thing is, Salesforce, it's two things though, right? Like you're, you see this name down X percent and you think, oh, no way it's that bad. But the problem is you have both like earnings take a hit and you have multiple compression.

Yeah.

And so it's just like you get hit on both and then it's like, all right, maybe this is the right price for it.

Reality. Yeah.

Yeah.

I don't know. I was thinking like enterprise software is probably a little bit safer than most of the other bullshit, mainly because I don't know, they'll probably, I don't know, enterprises aren't going to use vibecoded shit, right? Like at least not for a long time.

Aren't, but it's like, it's an option at least, right? Like it's, they can, like some of them, they can budget internally to say how much would it take you to build this? And then when they're. And, and also there's like two things when you renew, uh, they're like, okay, we're gonna increase the price by, you know, normally SaaS companies will just like, uh, this year we'll increase the price 5% or whatever. And, and they can just be like, listen, no, like we can't, we know we can build this internally for this much. You know, you're gonna actually like take the price down by this much. And then the other thing is just if you need less developers to do, to, to, to deliver your product, then you're just buying less seats, right? For these like SaaS things that are, that are leveraged to seats. Like, um, and so I think if all these companies get smaller, it's just like reflexive, right? On the downside. Um, so yeah, I haven't taken any stabs at SaaS. I don't have time honestly to look at the SaaS companies and like try and figure out which ones are cheap. Like I'd love to find a manager that's specialized in this because I think they'll make a killing, but I don't have the time for it to be taking punts at Salesforce or whatever.

Yeah, I feel like I'm just the dumb money in that scenario, kind of just taking a stab and I'm the dumb money in most of these scenarios.

That's why I'm outsourcing my thinking to you guys.

Yeah, but it's, it's, it's, it's exciting and, and like terrifying too, uh, sometimes. Yeah, it feels like energy is a really good spot though, like just because you, you have like such capped— it feels like your downside is very capped on even the most expensive things in energy aren't trading that crazy. Like maybe Constellation Energy, okay, that's not really an energy company, it's more like an infrastructure, but they're not really trading that crazy. And it just seems like you're going to have a persistent bid from this buildout. I don't know.

Yeah.

And a bunch of the— you see Leopold bought a bunch of the Bitcoin miners?

Yeah, he's repurposing them, right?

Yeah. Yeah, I thought that was interesting, which makes sense.

You have the bones.

Maybe the Galaxy will get a bid. Duncan's Galaxy coming back from the dead.

It's up today, it's up 12% just with everything else. Is this true that, uh, MicroStrategy has an, uh, MNAP premium again?

They've had one for a couple days, I thought.

Yeah.

Oh, really?

I thought they did.

That's good.

Cause it got down to, I guess he's probably buying.

I guess this is the pump.

I was going to say, where did he get money from? This dude's a wizard with just finding money in the banana stand.

There's always money.

Crazy.

Yeah.

I think Bitcoin is like very, like, looks good. I bought, I've been buying Bitcoin like last few weeks, months. The only thing I've been buying, honestly.

Like, I mean, gold's taken a hit, right? I think part of it's there's like a bunch of apparently Korean funds with exposure to Korean markets that had margin calls and had to sell gold as a result. I think that was part of the source of the drop down. Um, but yeah, now, now you look at Bitcoin as a percentage of gold and it just, you start doing some moon maths and it attracts them again. It's great again.

It looks cheap again. Yeah, it's nice. Um, yeah. And I think if Bitcoin does well, everything else will probably get a, get a catch-up bid. I still, yeah, I basically still have the same thesis on crypto I did previously. Uh, hopefully something in AI gets product-market fit, but it's unclear to me what that'll be.

Yeah, hopefully that, I mean, and there's a world where I'd like, I'm fully bought into the world that, or to the scenario where it does. It just makes too much sense.

I think we'll definitely have some like insane moments of like an AI agent trading or like playing Polymarket or something like this. Um, being like self-hosted on a, on a TEE or something and just like, you know, you're just watching this thing like make money on chain.

The real goat.

I've, I've also been just speaking to to some, uh, I've been just trying to speak to some more hedge funds to, to again, try and get this exposure. And a lot of them are using LLMs to trade already, which I thought was interesting.

In what way?

Like, what do you mean?

So the one, the one, uh, I, one of the ones I spoke to, they basically just plug in all their data, um, so to, to the, to the LLMs and, and then they're like literally refreshing it. Every, every 20 or 30 seconds. And, uh, it's actually generating like, uh, so this is, this was a systematic hedge fund, not even a like a discretionary one, more of like a quant, uh, and it's, it's generating like tradable insights like based on their context and the, and all the data they have access to and stuff like this.

So they're giving it like past trades and then like parameters or things that they want it to find?

Yeah, yeah. I think live prices, volumes, like, uh, correlations. And I, I don't know enough about quant to to really know what go— what goes into making the, the soup. But yeah, a lot of that stuff they're doing, which I think is—

that makes a lot of sense. I think it makes a lot, a lot of sense to focus on quant stuff first. I feel like discretionary is a bit harder to do, but yeah, it's probably—

I would almost argue that it's like more incrementally valuable for discretionary because the quant funds are already very algorithmically driven, right? And so it's just like, slapping another algorithm on top of their existing ones.

Oh yeah, yeah, it would, it would make—

yeah, yeah.

Whereas like on the discretionary side, it's just like helping you process this information faster than you'd previously do it through an analyst and, and, and spreadsheets and all of that.

I'm sure that's happening like on all of them for sure. I heard Goldman's also like not hiring junior analysts this year. I don't, I don't know who, where I read that, and it's like they're just telling all the all the partners to like just figure it out.

It's a lot of sectors. That's being a new grad now sucks. Yeah.

Yeah.

Like that's the area where, like, if you're a lot younger, like you're still in elementary school or like high school, whatever, things will be so much different. And then obviously people who have established careers right now are in a decent spot, but it's the entry-level jobs that are getting cut. Um, which is definitely seems like that's the biggest risk to the economy.

Maybe brings us to the Citrini thing, not to belabor it, but I'm curious where you guys sit on that.

I think my, my biggest takeaway from the Citrini thing is that the authors up front said how this was like very low probability and was kind of like this exercise in just thinking about another scenario just to like think about it. So like people are aware and everyone like really fucking hated them for doing that. Right. Like people do not want to even read or think about the bear case. Even if upfront they're like, hey, this is mostly for like, you know, illustrative purposes, like a nice piece of writing. Um, it's well-researched and everything, but it's not like what we necessarily think, uh, is like most likely to happen. Even with all those disclaimers, like you had, first of all, it impacted the market, which is insane. Second of all, you had rebuttals from like Citadel, like, yeah, the Economist, like all these publications, like everybody felt like they had to come out and say why it was wrong, which is like kind of a little bit, um, unnerving, I guess, right? That like everybody had to come out and And say why it's wrong.

I think it's based on the amount of attention it got. I mean, what he said in there wasn't necessarily new, right? It's kind of what everyone's been thinking about as the, not everyone, but like people have been thinking about the bear case. I think he just packaged it really well and into a digestible way for everyone to really kind of see the various kind of how various areas are all going to be affected and kind of basically at the same time. But yeah, he mentioned, you're right, is it an edge case, bear case? I think the biggest counter is there were no bullish elements discussed, right? And like, it's not, it just can't be all bearish. There's got to, obviously there's going to be, there's like an offsetting gain in productivity and growth and all of this stuff. And so I think it was a good piece to trigger a bunch of people. It was well-written, well-researched, But like, I, I, I assign a pretty low probability to it. Not, not because I don't think that some of that stuff is, is gonna happen, but I just think there are a lot of bullish elements that come along with some of that stuff happening. And then even in like the biggest edge case, I think you, you probably get some, some doomsday scenarios where, you know, governments will intervene before some of this stuff gets really bad. And in a world where we're ignoring the, the bullish side, or you just get like mass printing or something.

Yeah.

Yeah. I think the majority of the argument is like whether you believe that, um, it will be like every other technology where it's like a complement to labor, right? And like eventually you find new jobs that people do and they can do more and it's, and increases productivity and stuff like this, which I think, well, I think it'll increase productivity regardless, but are there new jobs basically? Like with, there were with all other technological revolutions, or do you think like these jobs are gone? For good. Um, I, I'm like, and that's, that was kind of what Citadel's rebuttal was like, you know, smarter words, but I think the TL;DR, and for me, I'm, I'm like a Citrini short-term and Citadel like long-term basically. And that's, I think in the short, medium term, like the path dependency. When I just look at some of my, some of my cousins or, or like people I know here, here in Portugal who graduated like with a business degree with, with good grades and went to work at like Deloitte or somewhere, or one of the consultants, one of the, and, and I, I just think those jobs are cooked and extremely hard to, to get back, you know? And, and I, and I think there's gonna be a period of time where you just have these people like unable to, to get new jobs. And at the same time, they're rich, getting richer and richer and richer, like in, in private markets and, and public markets. And I think that can, has historically been, you know, not, not a great combination for like social, um, like social tensions. And so, and I think in the long term it'll be fine. Like everything will be really cheap. We'll have abundance. I'm an, I'm a full AGI pilled person. Um, uh, and, and people will also find new things to do, but the short to medium term I think can be pretty ugly. Like, 'cause I really don't, it feels different than the previous ones where like, This thing can just, is so much better than me at most things, you know? And, and, and it's much cheaper than me. Like I have to eat and, and be housed and, and like I, I have health issues, you know? Like this thing is just, is just the cost of inference and it's only gonna go down and it's the worst it'll ever be. Like, yeah. Uh, we've seen it pick up quite a bit already, right?

Like it's, it's actually, you know, running some of these agents on an annual basis isn't, you know, 1/100th the, the cost of the employee for some of this stuff. So it's like there is somewhat of a floor because it just takes so long for the energy build out to happen. You know, that is going to be the bottleneck. And for better or for worse, that slow build out will kind of, I think, act to temper the growth, but also probably ease the volatility of the transition. I don't know what people are going to do, though. I like it. I think in previous transitions, you know, Industrial Revolution, you were doing labor in one form, but now you're just kind of still doing labor, but in a different setting and increased output. And then the next generation shifted to more white-collar work as tech revolution happened. I don't know what, where there will like this glut of unemployed kind of early-stage white-collar workers are going to do.

Um, I mean, the social unrest has definitely been increasing. Like, you've seen data centers get laid—

yeah, that was like my big takeaway when I was reading this. Number one, right, like, if Citrini is right and it kind of goes down this path, there's no way, number one, the government doesn't try to slow it down in some way, shape, or form. And number two, there's no way people don't— people aren't just going to like sit there and let all of their careers and livelihoods and everything and all of their ability to provide for their families. Just like they're not going to just— there will be a ton of, you know, civil, social type of unrest before it gets to the point where Citrini's, you know, thesis fully plays out, in my opinion. And then to Jan's point, like, if you go back to the Industrial Revolution, like steam engines and shit, like, yeah, somebody who rides a horse or, you know, does like a horse and buggy might not have been able to like fathom what like a car and like a highway system might be, but like, it's really not that big of a retraining, right? Like you're still driving something. Like it's still like in the same universe of, of things. Yeah. Whereas like now, like, I don't know, maybe it just goes like the way of like, you know, what trading floors used to be. You had like dozens or hundreds of people on trading floors with phones and shit. Nowadays you have like a couple people with a lot of computers. And maybe it's just— yeah, the problem is when you have that everywhere, right?

It's one thing if it's in one area and they can do something else, right? Yeah. Now you have quants taking over that role, but it's another if it happens in multiple places simultaneously very quickly. Yeah. I'm like, yeah. And you've seen Elon talk about kind of, you know, you'll have UBI. I think he, that's a narrative he has to shill obviously because he needs to be glass half full to really accomplish what he wants to do. But I'm pretty bearish on when all that plays out, when, when just like, no, you know, people just can't become artists because they have nothing to do and that's going to make them happy.

Oh man, I've, I've saw, I've seen this video. Instagram is so bad for like AI slop these days, but I saw this video and they were interviewing like Musk, Altman, Bezos, and they're all kind of like these fat guys and it's like, the theme of this is like, you You know, when everybody lost their jobs, people started like not working, not being active, they didn't know what to do. Uh, and so now we developed blah blah blah, and it's like all these people really fit on bikes and they're just like powering the AI. It's like, this is what the normal people are like looking at and thinking about, right?

I gotta jump, guys. Sorry, I'm already late to a call. But the one thing I've been— I got— I went to Egypt the other year and Cairo, there's like 23 million people in Cairo. And that's one thing I often think about when I'm like, what are all the people going to do after AGI? I'm like, what are they doing now in Cairo? You know, there's like 23 million people there. There's no real industry. There's no real like, uh, like big companies that have headquarters there. And so it's like, they're mostly doing like small, as far as I can tell, like informal jobs and like selling on stuff on the streets and, and all this kind of stuff, but it seems to just like work. People are, people are living, you know. Um, so I mean, obviously you wouldn't want the, the whole world in a sense.

How often are you going to Cairo?

Yeah, the, the New York, the New York slums. You think the New York slums are coming?

Everyone just pivots to tourism.

I looked out my window trying to picture Cairo.

I'm not sure that'll fly over here.

It's very hard to go back in standard of living.

The Cairoization of the world.

Yeah.

But yeah, guys, this is, this was fun. Yeah.

We'll just wrap it in general. Um, yeah, this was a good episode. Um, I feel, I don't know, I feel like we're a bit more optimistic. Uh, the, the pro Bitcoin was actually flat during this podcast. That's like a major win.

Damn, that might be the most bullish thing.

Yeah. It's the most bullish. It might actually be up like $200. So, uh, yeah, thanks everybody for joining, uh, and we'll be back in a couple weeks, hopefully with, uh, $80K+ Bitcoin. Thanks everybody.

Why so bearish?