# The AI Supercycle with Jordi Visser | Raoul Pal the Journey Man

https://www.youtube.com/watch?v=jGNmWFeCIWE

[00:00] Building a business in AI has tremendous
[00:03] margins. Tremendous. It allows you to
[00:06] grow rapidly. We've never seen this
[00:07] before, right? So Reed's law, which is
[00:09] meal squared, has never existed in
[00:12] biology, not even viruses. It doesn't
[00:15] exist. And now we're seeing it and it's
[00:17] [ __ ] everywhere, which is why it's so
[00:19] hard to understand. the new world is if
[00:21] we can't actually make all the chips we
[00:23] need and we can't actually get the power
[00:26] we need, you end up with a little bit of
[00:28] demand versus supply mismatch. The
[00:30] bottlenecks themselves may slow the
[00:31] earnings of these companies. Not because
[00:33] the demand is not there, because the
[00:34] demand is too big. Okay, that's sort of
[00:36] bananas, but that's what we might get
[00:38] to. As you know by now, I'm Ral Pal and
[00:42] welcome to my show, The Journeyman,
[00:44] where we travel to that nexus of
[00:45] understanding between macro, crypto, and
[00:48] the exponential age of technology. I've
[00:51] been bleeting on for a few years now how
[00:53] this is all coming together, macro,
[00:55] crypto, and technology. They're all the
[00:56] same thing, and everything is changing,
[00:58] and it's changing extremely fast.
[01:01] There's not many people who understand
[01:02] across all of these disciplines. Um, you
[01:06] can find experts in various fields, but
[01:08] very few people who understand at broad
[01:10] in macro terms what this all means. But
[01:13] my regular guest, good friend of mine,
[01:15] Jordi Vissa, well, he's the person I go
[01:17] to to think things through. And he does
[01:20] the same with me. So, it's not really an
[01:22] interview ever. It's just us thinking
[01:24] through what the hell is going on, what
[01:26] does it mean, how to measure it, how to
[01:28] take opportunity of it. So, I think
[01:30] you're going to love it. Here's a
[01:31] conversation with good friend Jordi
[01:33] Vissa.
[01:34] Join me Ral Pal as I go on a journey of
[01:37] discovery through the macro, crypto, and
[01:40] exponential age landscapes. In the
[01:43] journey man, I talk to the smartest
[01:45] people in the world so we can all become
[01:47] smarter together.
[01:53] Jordi Vissa, how the devil are you?
[01:56] >> How you doing, my friend?
[01:57] >> I'm good. I'm good. So, lots to talk
[02:00] about, I'm sure. I have no idea as ever
[02:02] what we're going to talk about, but
[02:03] what's on your mind?
[02:05] >> Uh, well, I'm going to start in a good
[02:06] place for you. Um, I actually listened
[02:09] to you and and Julian recently and uh
[02:13] and something you guys talked about
[02:16] throughout, but really at the very end,
[02:18] um, if you remember when we sat down in
[02:20] my offices and I talked about um, no
[02:23] more recessions.
[02:24] >> Yeah.
[02:25] >> Uh, and that kind of hit you. I the way
[02:28] you guys phrased the transition away
[02:30] from labor verse capital into compute
[02:33] verse energy. Uh you and I have talked
[02:36] about this to in some way. We're both AI
[02:39] believers, but it really hit me hard in
[02:43] the same way the recession did because
[02:44] it fits in with that whole conversation.
[02:47] Uh meaning business cycles of the past
[02:49] to grow your business, you needed
[02:52] location, you needed to borrow money,
[02:53] you needed to hire people to grow it
[02:55] every year. when we were at at
[02:57] investment banks, okay, this division is
[02:59] doing well. Let's give them more people.
[03:01] Let's go open an office in Brazil. Let's
[03:03] go do this.
[03:05] In compute verse energy, it's a very,
[03:08] very different thing. And I spent the
[03:09] time and I've been writing about um that
[03:12] the AI cycle is no longer about capital
[03:15] verse labor, but not phrasing the way
[03:17] you guys did. And not specifically
[03:19] saying compute verse energy, but
[03:21] actually saying bottlenecks and
[03:23] shortages, which are the same thing as
[03:25] when supply and demand get out of whack
[03:27] on the other side. But we're we're it
[03:31] really hit me hard that we're starting
[03:32] to see that phase and that people should
[03:34] get used to that. the new world is if we
[03:37] can't actually make all the chips we
[03:39] need and we can't actually get the power
[03:42] we need in a delayed fashion, you end up
[03:44] with a little bit of demand versus
[03:46] supply mismatch. So that's what I've
[03:48] been thinking a lot and writing about.
[03:50] >> So I got further in this thinking. So I
[03:52] I've started building out a whole
[03:54] dashboard of indicators for monitoring
[03:58] this um exponential increase in the
[04:01] output of intelligence per unit of
[04:03] energy and I built an index of it. Um,
[04:05] and I'm still not I've not published it
[04:07] fully yet. I just started writing about
[04:08] it in GMI. Really interesting. So, I
[04:10] used Moore's law beforehand
[04:13] and then it starts hooking up um really
[04:16] with AI. You've got some kind of GPU and
[04:19] other stuff that starts lifting it and
[04:21] then AI it's now gone it's a log chart
[04:24] and it's now gone exponential on a log
[04:26] chart, right? Which is what I've been
[04:27] talking about Reed's law the the
[04:29] exponential squared.
[04:31] >> Um, and what it started to think about.
[04:34] So think about these these bottlenecks
[04:36] and stuff like this. Think about the
[04:38] fact that like data centers are 30%
[04:41] built versus where they should have been
[04:43] what they stated.
[04:44] Think about the race between the US and
[04:48] China and how nobody is allowed to win
[04:49] it. Think about the fact that no single
[04:53] AI frontier house can win because it's
[04:55] all too dangerous, right, to have
[04:57] >> Y.
[04:57] >> And what I get to is there is almost no
[05:00] way for this not to be a super cycle.
[05:02] >> Mhm. And even the bottlenecks just slow
[05:04] it down, but it has to keep expend. So
[05:07] what is a bottleneck? It actually needs
[05:09] more expenditure. You need to build out
[05:11] the power stuff. You need to build out
[05:12] whatever it is. And so what you've got
[05:14] is the largest capex cycle I think
[05:16] humanity will ever see at this rate.
[05:19] Well, maybe there's another one to space
[05:20] later, but right now is this.
[05:22] >> Yeah.
[05:23] >> Um and I'm struggling to see how we
[05:26] actually get a business cycle. And you
[05:28] and I will remember the days of 1995 to
[05:31] 2001
[05:33] where the business cycle went up and
[05:35] down a bit but basically it was all
[05:37] productivity and growth.
[05:39] >> Yeah. So let let's make sure we we we we
[05:43] double click on one part and let's see
[05:45] if you agree with this. So for the
[05:47] business cycle um it is based on
[05:50] perception. It's based on surveys. It's
[05:52] based on the way people see it because
[05:54] everyone can't see everything happening.
[05:57] So when people talk about PMIs, okay,
[06:00] well those are surveys. Those are this
[06:02] is what I expect to have happen. And so
[06:04] when they're high, there's a lot of
[06:06] different components in it. There's
[06:08] prices paid, there's supply delivery
[06:11] times, there's new orders, there's
[06:12] productions, and with inside the PMI,
[06:14] you can have a high PMI number, but you
[06:16] could have bottleneck showing up in the
[06:18] supply chain in the prices paid. And you
[06:20] could see new orders drop down to 50.
[06:22] You could see the other things, and we
[06:23] saw a lot of that during COVID. and we
[06:25] have the employment numbers which are
[06:27] the component which is not going higher.
[06:28] So I think my I think everything you
[06:31] said I completely agree with. I do
[06:33] believe one of the problems and I want
[06:35] to double click on something you said if
[06:37] people haven't thought about this um how
[06:39] could we possibly have this much
[06:41] advancements in the IQ stage getting us
[06:43] up to 135 plus right now without having
[06:47] the data centers being built. How could
[06:49] that have occurred? And you and I both
[06:51] know well the algorithmic side got
[06:53] better. human feedback, reinforcement,
[06:56] learning, reasoning, everything kind of
[06:58] happen. And that's why we've been able
[07:00] at least to stay up to here. I believe
[07:02] we're at a different point. But I want
[07:03] to hear that's why we're seeing an
[07:06] exponential of an exponential because
[07:07] it's not just about the build more oil
[07:12] rigs and we have more oil, right? That's
[07:14] a standard process. So build more data
[07:16] centers and we have more energy uh more
[07:18] intelligence. Yes. But that intelligence
[07:22] becomes self-recursive learning,
[07:24] improvements in the algorithm, all of
[07:25] that. And that is where the double
[07:27] exponential comes from. Um, and we've
[07:30] never seen this before. Right? So Reed's
[07:32] law, which is law squared or to the
[07:35] power of two, has never existed in
[07:38] biology, not even, you know, viruses. It
[07:41] doesn't exist. And now we're seeing it
[07:43] and it's [ __ ] everywhere, which is
[07:45] why it's so hard to understand.
[07:47] not only hard to understand, but how
[07:50] fast it starts to move at a pace that
[07:52] parabas now become something that you
[07:54] get used to. And you and I have been
[07:56] doing this long enough. And because I
[07:58] grew up in emerging markets, I used to
[07:59] see parabas a lot. And they usually went
[08:03] up and then they'd come down in in
[08:05] almost the exact same straight line. I
[08:07] think it has confused people. I I
[08:08] literally just wrote something uh where
[08:12] if you go back to January and you take
[08:14] when Jensen Yuang got on stage at CES
[08:17] and he spoke and he talked about Vera
[08:19] Rubin and he talked about all the things
[08:21] that were needed. You know, you and I
[08:23] are in in in different um we're in the
[08:26] same world. We have the same background,
[08:28] but I have two feet or I have one foot
[08:31] fully in the Trapfi world and then one
[08:33] foot fully in the crypto world. And I
[08:35] still talk to hedge funds every single
[08:38] day. I talked to mutual funds. I talked
[08:40] to people about AI and how to invest in
[08:42] it. In January, they didn't understand
[08:44] the agentic economy had started. They
[08:46] didn't understand the rise of agents at
[08:48] that point.
[08:49] >> And then by March, you had the Morgan
[08:52] Stanley TMT event and Jensen Yuang spoke
[08:55] there and Intel spoke and Dell spoke and
[08:57] they all started and everyone started
[08:59] realizing, "Oh my god, these numbers are
[09:01] going to be huge. We got to start
[09:02] getting involved." And then you had
[09:04] computex uh over the course of you know
[09:07] this month and I think the realization
[09:10] has hit people but the problem with the
[09:12] agentic world in the way that Jensen
[09:14] described it at CES which makes it a
[09:16] little bit different than the last 3
[09:17] years with the IQ acceleration.
[09:20] It's going to be very difficult and very
[09:22] correlated amongst all of the broadening
[09:24] out that's happened to make sure that we
[09:26] can continue to keep this thing going at
[09:28] the same pace that it's been happening.
[09:30] So, I do believe there's going to be a
[09:32] slowdown to some degree, but to your
[09:34] point, for everyone watching, recursive
[09:36] self-improvement and where we're getting
[09:39] a year from now, you're going to be
[09:40] blown more blown away than you are
[09:42] today, even if for the next four months,
[09:44] the focus shifts to bottlenecks and
[09:46] shortages.
[09:46] >> And also, you know, I was um building
[09:50] something out about this is like, you
[09:52] know, it's not going to be the same
[09:53] stocks all the way through because
[09:55] there's bottlenecks and shortages or
[09:57] there's the applications layer, right?
[09:58] people still aren't focusing. Something
[09:59] you talked about a long time ago is
[10:01] like, you know, the peptides and and all
[10:03] of the genetic science breakthroughs
[10:05] that are enabled by this technology,
[10:07] right? The market can't focus. The
[10:09] market can only hold so much attention
[10:10] and so much capital at any one stage. It
[10:13] hasn't figured out the agentic economy
[10:15] yet because if not, you'd see it in
[10:16] crypto because it's massive. The TAM has
[10:19] gone to infinity, which people have
[10:21] never understood. The TAM was always
[10:23] humans, right? And now the TAM is
[10:25] infinity. And we people don't understand
[10:27] this yet. They don't understand um what
[10:30] this is about to do to human biology.
[10:33] They kind of sort of do, but it's not
[10:36] reflected in stock prices yet. Um so I
[10:39] think we'll get rotations. You know, it
[10:41] doesn't always have to be about Nvidia.
[10:43] As you say, there'll be times when it's
[10:46] going to take a while to push a new
[10:47] breakthrough from algorithmic
[10:49] compression or build these data centers
[10:51] in enough scale to show the next big
[10:53] step. you know, after Mythos, maybe it's
[10:55] going to be difficult for a while. Maybe
[10:57] not. It's not proven difficult yet.
[10:59] They've never yet hit any boundary and
[11:02] they keep saying we're not hitting
[11:03] boundaries and we thought we would and
[11:04] we haven't. But I do think there'll be
[11:06] rotations as as you say it goes through
[11:08] different component parts and kind of
[11:10] breaks them down because you know if you
[11:12] think about the bottleneck theory. So
[11:14] you find a bottleneck whatever it may be
[11:18] >> power
[11:19] what that bottleneck does is concentrate
[11:21] capital into that particular issue
[11:24] >> because if the universe is solving for
[11:26] intelligence per unit of energy it will
[11:28] clear all roadblocks to get there and
[11:30] capital is how you do it. capital and
[11:32] attention.
[11:33] >> And so two two things you said on there
[11:35] on the power side, it's very clear to
[11:37] everyone now based on the data center
[11:39] delays, based on the turbines and the
[11:41] transformers and every other part that
[11:44] it's we've got a bottleneck. So what is
[11:47] the recent announcement from Nvidia and
[11:49] Seammens? Well, they're doing something
[11:51] with Fluent specifically on the battery
[11:53] side in China. They're really focused on
[11:55] solid state batteries. Huge silver
[11:56] imports. Everyone should go look at how
[11:58] much silver is needed in solid state
[12:00] batteries instead of what lithium takes.
[12:03] If you take if we had battery innovation
[12:06] today and we were able to, we have
[12:08] enough power on the US grid to plug
[12:09] everything into it. We just can't do
[12:11] that because we reap peak capacity too
[12:14] much. But if we could store and have it
[12:16] to deal with those peak capacity days,
[12:18] then we wouldn't we'd be able to use the
[12:20] grid almost completely to get the needs
[12:22] that we have by 2030. So innovations in
[12:24] the stuff that we're talking about are
[12:25] going to happen and power is a perfect
[12:27] example. You brought something up which
[12:28] I think and I'm starting to focus my
[12:30] attention. So when you say we'll rotate,
[12:33] we'll move to other places. When Jensen
[12:34] Yuang did the five layer cake and he
[12:36] talked about the fact that at the bottom
[12:37] of the stack you have energy, you have
[12:39] chips, you have infrastructure, then you
[12:41] have the models and at the top you have
[12:42] applications. Every time I talk to
[12:44] people on applications they are focused
[12:46] on SAS for some reason. And I always sit
[12:49] there in these meetings and I go, why
[12:50] are you focused on a seatbased thing of
[12:53] the past? What I envision happening, and
[12:55] this is the first time I'll say it in
[12:57] any um podcast, even though I've talked
[12:59] about the name, the reason we're in this
[13:02] stage that the buildout can happen with
[13:04] companies like Google able to raise $85
[13:07] billion in a public equity. You and I
[13:10] have been involved a long time. There's
[13:11] never been a public equity of that size,
[13:13] but let's normalize it. It's bigger than
[13:15] the bottom 360 companies in the S&P
[13:19] 500's market cap. So, you're talking
[13:21] about a massive raise barely down uh for
[13:24] the buildout. To be in that position,
[13:26] you had to be a Warner from the time
[13:28] that the great financial crisis came
[13:30] out. They sucked in all of the capital.
[13:32] You know who's starting to suck in all
[13:34] the capital right now for the
[13:35] application layer? It's for human
[13:36] software. And it's what you brought up.
[13:38] Eli Liy is sucking in all of the capital
[13:41] through GLP1s. And I believe GLP1s are
[13:45] going to lead to them. They have a
[13:47] thousand GPU um data center now uh at
[13:50] Lilipod at their campus. They worked
[13:52] with Nvidia to launch something in
[13:54] Silicon Valley. I believe we're going to
[13:56] look back and and realize that GLP1s
[13:59] were the ability to finance the next
[14:01] stage of the human software that you're
[14:03] talking about.
[14:04] >> Yeah. It's like Elon's used cars, you
[14:06] know, you use something that generates
[14:08] cash flows, you know, and we'll look
[14:10] back at Google and look back at some of
[14:12] these people and say, well, advertising
[14:13] was just the start of actually a much
[14:15] bigger process. Just a couple of things
[14:18] to add to that as well is
[14:21] talking about efficiencies and
[14:22] bottlenecks.
[14:24] Elon's been is always very good at this
[14:26] kind of stuff. So, when the Cyber Truck
[14:29] came out, he realized that there is a
[14:31] global copper shortage, right? Every
[14:32] hedge fund in the world plays the AI
[14:34] trade via the copper shortage. And it's
[14:35] it's been okay, but not the greatest
[14:37] trade on Earth because I think
[14:38] everybody's focused on it. But what Elon
[14:40] did was change the voltage in a Cybert
[14:43] truck from 12 to 24 and ended up using
[14:46] 70% less copper.
[14:49] It's like a very simple thing. He's
[14:51] like, well, nobody done it before and it
[14:53] was just basically physics and I just
[14:55] did that and we figured out how to do it
[14:57] and then it changed everything. So, I do
[14:59] think there's a lot of efficiencies. you
[15:01] know, you talk about silver and stuff,
[15:02] we will root around it because we're
[15:04] quite intelligent. It was an old friend
[15:06] of mine um who was at Goldman who who
[15:08] was on the oil trading desk wrote a
[15:10] book, The Energy World is Flat. And I
[15:12] didn't really understand it at first um
[15:14] but basically explaining that
[15:17] that the intelligence density within the
[15:20] oil companies is vast. And if you give
[15:23] them a roadblock like we can't get
[15:25] enough oil, there's going to be peak
[15:26] oil, they come up with shale.
[15:29] And it's the same with the drug
[15:30] companies. It's the same with Elon and
[15:32] Copper. You know, I think Doomongous
[15:35] will say, "Well, it's [ __ ] It can't
[15:36] happen." And my thought will capital and
[15:40] attention will just root through it. And
[15:42] again, I I I don't know. I don't think
[15:44] you guys um said this when I listened,
[15:48] but I heard the copper point, which I
[15:50] agree with. I mean, for the amount of
[15:52] times I've heard, we're we're not going
[15:53] to have copper. Um, which again it's
[15:55] true but you could have the same you
[15:58] could have said the same thing with the
[15:59] data centers and then Vera Rubin came
[16:01] out and we moved most of the stuff to
[16:02] optical fiber to move that around. So
[16:04] it's like we keep coming up with
[16:06] solutions to problems to to to deal with
[16:08] this
[16:08] >> and AI will make the solutions faster to
[16:10] to get to as well.
[16:12] >> Well, but that that's the main point.
[16:14] And so when I look at parabas and um the
[16:17] easiest way for me to say to someone,
[16:19] okay, let's assume there was a parabola
[16:21] that was justified if at the beginning
[16:24] of January in January of 2026, instead
[16:26] of him announcing that the agentic world
[16:29] was rising, what if he actually said,
[16:31] you know, there's 7.5 billion people on
[16:33] the planet, but by the end of this year
[16:34] there'll be 15 billion. Um, okay, we'd
[16:37] have parabas everywhere. It'd be in
[16:39] food. It would be we wouldn't have
[16:40] enough things right off the bat. So what
[16:43] he did when the agentic world comes
[16:44] people haven't made the connection that
[16:46] we're talking about billions of thinkers
[16:48] entering the the world and and they
[16:52] consume only one thing compute. So
[16:54] that's why we have shortages of all
[16:56] these things cuz that's what they
[16:57] compute. The reason the business cycle
[16:59] and people looking for when the next
[17:00] housing market's going to start. You're
[17:02] missing the point that these digital
[17:03] employees will never buy a house. They
[17:05] will never send their kids to college.
[17:06] They will never consume things the way
[17:08] it is. So, you have to kind of back your
[17:10] way out and realize, forget the doom and
[17:12] gloom over the job replacement. The
[17:13] reality is, as Elon says, at some point,
[17:16] if they're really good at solving
[17:18] problems and abundance comes, you're
[17:20] going to have a choice of whether you
[17:21] actually want to work or if your life is
[17:23] fine just being in nature. And that time
[17:26] commitment of the next 50 years of
[17:28] adjusting to a world that is completely
[17:30] different, that is what happened in my
[17:32] opinion there. And to your point on any
[17:34] kind of problem we have with seven and a
[17:37] half billion agents coming in over a
[17:39] course of whatever the next year, two
[17:41] years, whatever it takes, we're
[17:43] basically doing the Manhattan project
[17:45] times whatever. So any problem you want,
[17:48] forget Elon figuring it out, he's one
[17:51] person. But if you take all of the AI
[17:53] agents and let them sit and work on any
[17:55] problem there is, we will solve
[17:57] eventually every single problem. And I
[17:58] don't think people have come into the
[18:00] context of that much IQ and what it
[18:02] means to have that much IQ solving
[18:04] problems.
[18:05] >> And also cuz people think of these
[18:06] models as a single thing, right? This
[18:10] single beast, but they're not because
[18:11] there's a 100 million users. No, it's a
[18:13] billion users, I think, now of of
[18:15] OpenAI. Each one is a different instance
[18:18] using this huge intelligence of which
[18:21] they can come with different
[18:22] breakthroughs. It's not the big model
[18:23] itself. the big model enables all these
[18:26] instances and depending what you do with
[18:28] it. So it's it's it's exponential yet
[18:31] again in the amount of intelligence you
[18:34] can drag out of this thing. Yeah, that
[18:36] that's a brilliant comment and I've
[18:38] experienced it a couple of ways. But
[18:40] let's assume that every time that I ask
[18:42] a question or I go down a rabbit hole of
[18:45] of what will happen if this or what'll
[18:48] go on when I'm writing papers and all of
[18:50] that's getting added into the way that
[18:52] things are thinking about. you're
[18:53] getting all of human intelligence put
[18:55] back in. The other direction I've gone
[18:58] is to isolate intelligence to try and
[19:00] brainstorm with certain people. So, one
[19:02] of the things that's worked extremely
[19:03] well for me this year, both investing,
[19:05] but also writing papers, has been
[19:07] creating knowledge brains. And I did it
[19:09] in a very simplistic way to start, which
[19:11] was just let me take every transcript
[19:12] and upload it into notebook LM of Jensen
[19:15] Yuong of whoever I wanted to do. I
[19:17] recently did it with David Ricks from
[19:19] Eli Liy because I want to hear what
[19:21] they're talking about and I want to have
[19:23] it in one transcript. Jensen Yuang
[19:25] speaks multiple times a week and he
[19:27] these are not short conversations. These
[19:28] are anywhere from an hour to 3 hours. If
[19:30] you upload 3 hours of him speaking, he
[19:33] doesn't have a teleprompter. He's just
[19:34] winging it. You get so much good raw
[19:38] information. He's basically telling you
[19:39] what companies to buy. And to do it, you
[19:42] can either do a notebook LM or you can
[19:44] take all those transcripts and save them
[19:45] into a notebook on your computer and
[19:47] then have co-work go in there. Then you
[19:48] can have your agents go off and run
[19:50] through it and connect it to another
[19:51] person, Andre Carpathy, whatever you
[19:53] want. So the ability of taking a human
[19:56] being's brain and making that the
[19:57] content that you're working off of as
[19:59] opposed to the internet which is every
[20:01] human being I don't think people have
[20:02] made that connection yet of how quickly
[20:04] you will do it because you can isolate
[20:06] the most individual thinkers in any
[20:08] given field and get tons of information
[20:11] particularly when you pair them with
[20:12] other people.
[20:12] >> Yeah. And I've built um I'm building a
[20:14] whole I mean so many things right now
[20:16] but one is the GMI brain. So it has
[20:19] everything I've ever written over the
[20:20] last 21 years. So, I've probably got
[20:21] more long form written content than
[20:22] almost anybody else in finance
[20:24] >> and then the video transcripts and then
[20:26] my X feed and then, you know, all of
[20:28] this stuff and then I've got a, you
[20:30] know, it's that's in a rag vector
[20:32] database. I can now query it, do all
[20:35] sorts of stuff. Okay, that's
[20:37] interesting.
[20:38] Then I built something that I'm still
[20:40] working on called the lens which uses my
[20:42] exponential age framework only and goes
[20:44] to first principles.
[20:46] um and you you give it any question
[20:49] whether it was the US election uh the
[20:51] midterm election or markets and it uses
[20:55] that particular lens to analyze stuff.
[20:58] So I'm now doing replications of parts
[21:01] of myself into different things. So I'm
[21:04] not just using Jensen yuan. The issue
[21:06] I've got is I mean I love the idea of
[21:08] what you're doing with notebook LM which
[21:09] I love as well is I am running out of
[21:12] time. I cannot I can't manage this. I
[21:15] I'm so el overwhelmed with the amount of
[21:18] things that I'm building and doing
[21:19] because you know I I can do it all
[21:22] myself now which is a dangerous thing
[21:24] because the only thing I've got is time.
[21:27] That is my energy per unit of
[21:29] intelligence. Yeah, my intelligence is
[21:30] going exponential but I'm right at the
[21:33] boundary of my fixed time.
[21:36] >> So le let's we always talk about the
[21:39] similarities between us. Um one of the
[21:41] differences is I was still at a hedge
[21:44] fund. we were off building a separate
[21:45] business. When the hedge fund that I
[21:47] worked at for 20 years closed, I had to
[21:49] make a decision and I never wanted to
[21:51] work for anyone again and I never wanted
[21:54] to manage anyone again. Those were the
[21:56] two things that I kind of said brought
[21:57] me the least amount of joy. Going out
[22:00] raising money, going to investors,
[22:01] trying to convince them to give you
[22:02] money, then staying up all night trading
[22:04] the markets and explaining why you're
[22:06] doing whatever, and then people coming
[22:08] in and not able to pay their bills or
[22:11] whatever the case was. It just it was
[22:13] very difficult for me to kind of do all
[22:15] of that and still be able to enjoy life.
[22:17] So when I started to make the decision
[22:19] what I wanted to do and I realized
[22:21] content was going to be it. Not content
[22:23] for me but realizing that 8 billion
[22:26] people on the planet are not going to
[22:27] know how to navigate this thing. Very
[22:29] few people are going to be able to. And
[22:31] if I thought about it all the time and I
[22:33] took my ability to speak in analogies
[22:36] and convert it into language that they
[22:37] could understand, I could help people
[22:39] train on doing this that I could grow a
[22:41] business which has grown rapidly. And
[22:43] the one thing I can say is in your
[22:45] position, some of your time is just
[22:47] dealt with that you're involved with a
[22:49] big company. I have no employees. I have
[22:51] one person that helps me out once we
[22:53] launch um the payw wall. It's grown
[22:56] consistently. And the reason it's grown
[22:57] consistently is because I'm using AI
[22:59] agents, but I have all of my time. They
[23:02] deal with the the the subscribers and I
[23:05] deal with just creating content and
[23:06] going through it and learning and being
[23:08] able to do it. So, I appreciate the part
[23:10] you're saying, I don't know how you do
[23:11] as much as you do because I have a lot
[23:13] more time. But I do think for people
[23:15] that are listening, um, building a
[23:17] business in AI has tremendous margins,
[23:21] tremendous. It allows you to grow
[23:23] rapidly. So, you can go from zero to
[23:27] I don't even think my business will be
[23:29] around for very long because I think
[23:31] I'll it'll be monetized because I think
[23:33] people need what I'm doing and it's
[23:35] growing in a consistent linear way, but
[23:37] I'm having so much fun and I'm learning
[23:39] so much and I feel like it's making me
[23:41] >> How about So, Jordy now just thinks, you
[23:44] know what, I need to build Hermes agents
[23:46] and then I need to think about I need a
[23:47] permanent memory layer. So, how do you
[23:49] get the time to do it's that stuff that
[23:51] takes up my time, right? because
[23:53] eventually you you free up time but you
[23:56] end up finding new things to to do with
[23:58] this technology. So that time just gets
[24:00] sucked into something else. Um but but
[24:03] how do you find the time to do that? Cuz
[24:04] I'm doing all of these things and that's
[24:06] what's I'm really struggling. I've got
[24:08] 15 things on my list of different
[24:10] variations of stuff that I'm doing and
[24:12] having to learn brand new from scratch.
[24:15] I I can't I don't know this for sure,
[24:18] but I'm going to guess that I'm able to
[24:20] use No, it has to be. I use AI more than
[24:23] you do just by the nature of not having
[24:24] to do as many interviews and not having
[24:26] to deal with the business and whatever
[24:28] else things are traveling, the events. I
[24:31] choose and pick what I'm going to based
[24:33] on whether it makes sense for me totally
[24:35] from a a time basis. So, I probably have
[24:38] more time to use it.
[24:40] >> I'm not sure. I'm not sure. I mean, I'm
[24:42] all day from 6:00 a.m. till
[24:45] >> 8:00 p.m. I'm 14 hours a day at this
[24:47] stuff, you know? I do three speaking,
[24:50] you know, where I have to go and travel
[24:52] three a year, something like that. So,
[24:54] it's not really that. Yes, there's
[24:55] running real vision, but there's people
[24:56] running. I don't know. I'm just finding
[24:58] maybe I'm just overly ambitious in all
[25:00] the things I want to do because you you
[25:02] want to get everything moved along to
[25:04] get to that foundational layer. So,
[25:06] you're building all the databases across
[25:08] everything and how that they interact
[25:09] with each other and whether you're using
[25:11] a Obsidian or whether you're using a
[25:14] nation and all of this stuff is it just
[25:16] takes time. I you know what it is and to
[25:19] be fair so let's assume we do exactly
[25:21] the same amount it is a function of
[25:23] number one we see new stuff every day
[25:26] that we wish we were doing and so it's
[25:28] moving so fast and we're seeing it so
[25:31] fast because you and I are clearly on X
[25:33] we're clearly talking to smart people
[25:35] and if someone says to me you should be
[25:37] connecting to Obsibian to do your
[25:39] knowledge brains then immediately when I
[25:40] go home I go look and I'm like well how
[25:42] much is this how much time is this going
[25:43] to take how's it going to go with Hermes
[25:45] it's very interesting because I had just
[25:48] finished my second open claw and then
[25:50] all of a sudden everyone's like, "You
[25:51] got to move to Hermes." I'm like, "I I I
[25:53] just went through all this time to use
[25:55] OpenClaw and I read what they were doing
[25:59] and I I worked with one of the LLM.
[26:01] Should I be spending time on this?" And
[26:02] the answer was no, don't bother yet.
[26:04] It'll get easier in the next month. So,
[26:06] what I've gotten good at is kind of
[26:08] asking chateep. I'm I'm using 5.5 now
[26:12] much more than Claude. Um, I it just I
[26:16] think they're both at I don't know what
[26:18] the IQ is, but they're these things are
[26:20] the smartest things I've ever had
[26:22] conversations with, but chatbt is my
[26:24] style because it's less verbose and I
[26:26] like less verbose. Um, I just like quick
[26:29] answers, move on, get the next one. And
[26:30] it has been very good about telling me
[26:32] what not to spend time on.
[26:34] >> Yeah. Yeah. And what I got to is, you
[26:36] know, cuz having had that same
[26:37] conversation, most of the time we both
[26:41] know is like either Anthropic or or um
[26:44] OpenAI will build that you don't
[26:46] actually need the Hermes agent,
[26:48] >> but the foundational database layer you
[26:50] actually do need. So I'm like, I'm just
[26:52] going to build that for now. I've got a
[26:54] Hermes agent. Don't really use it yet
[26:55] because I'm building all these other
[26:57] things from that database. Because once
[26:59] you got your foundations right, how I
[27:00] think about these databases is we're all
[27:02] going to have our own vault.
[27:04] >> Yep.
[27:05] >> Of everything, your personal stuff,
[27:07] every photo you've ever had, every phone
[27:09] call you've ever will be in your vault.
[27:12] And then you can use the vault for
[27:14] various brain aspects or monetization
[27:17] aspects or whatever. So I'm just focused
[27:19] like, okay, I need to create the vault,
[27:21] the rail operating system as I call it.
[27:23] And and as people listen to this don't
[27:26] know what you're saying. Um honestly the
[27:29] beauty of an LLM is theoretically you
[27:31] can get information on any topic you
[27:33] want to get a topic on. You can have a
[27:35] chat on it. Great. But if every file on
[27:38] every computer you've used on your phone
[27:40] on everything, every piece of
[27:41] information you've ever had is now in a
[27:43] place where it can be accessed at any
[27:45] point like in in a second. I don't think
[27:48] people realize what that means. If
[27:50] they've gone out to dinner and someone
[27:51] says, "Well, where's the receipt for
[27:52] that?" you have the receipt. Like every
[27:54] single thing is at your fingertips. And
[27:56] the easiest way for me to like explain
[27:58] to people as to how important this is is
[28:01] um when a loved one dies and you're the
[28:04] executive or the administrator, you will
[28:06] hear from every person how much of a
[28:07] nightmare it is to go from point A to
[28:09] the end. I had to go through that over
[28:12] the course of the last 18 months. And
[28:14] midway through it, I just started
[28:16] putting everything into one folder and
[28:18] connecting that folder to opus 4.5
[28:22] initially and then go. So whenever the
[28:24] probate person calls, when everyone
[28:26] calls, I just go right into it and I
[28:28] say, "Hey, what they're looking for this
[28:29] answer. What what's the answer?" I don't
[28:32] have to go look. If they say, "What
[28:33] where are we right now on the estate
[28:35] account?" I just go in. What did you do
[28:37] this asset at? Where's the document for
[28:38] that? It's all in one file. It gets
[28:40] brought up by Claude. It immediately
[28:42] sends it out. So, I think for people to
[28:44] understand your comment on on the vault,
[28:46] they really have to understand how
[28:47] amazing it will be in the future, and
[28:50] that's one of the reasons why you need
[28:51] to have your personal assistant at least
[28:53] trying it just to get through the
[28:55] experience of everything that you can do
[28:56] with it.
[28:58] >> Um, have you started using granola yet?
[29:00] >> No.
[29:01] >> Okay. Granola's a great one. it it's um
[29:03] we've all seen on Zoom and everything
[29:05] else it's there's an AI that does the
[29:08] transcripts but Granola is like it has a
[29:12] bunch of different models you can use
[29:13] whether you're using chat GPT or you
[29:14] want to use Kimmy or whatever it is and
[29:16] it will do instant transcripts instant
[29:19] summaries okay fine but it becomes the
[29:21] knowledge base for everything so it's
[29:24] like oh you're speaking to Jordi again
[29:25] here's the four things you talked about
[29:26] last time here's the things you were
[29:27] going to follow up on and it all feeds
[29:30] into my brain so then every single
[29:32] conversation I have and I can do that on
[29:34] phone calls or leave myself voice
[29:35] messages all feeds the brain and it
[29:38] never forgets. So then I can go back and
[29:40] say hey listen you know when's the you
[29:43] know whether it's a real vision meeting
[29:45] a product and development team you know
[29:46] what are the last five things we talked
[29:47] about what what's going off track can
[29:49] you analyze what it's all of it so it's
[29:52] a really helpful tool and everybody I
[29:54] know uses it starts using it a lot
[29:56] >> so wait is this replace did you use is
[29:58] this notion but in a different way
[30:02] >> what it is is it just rec just think of
[30:04] it as recording everything you speak
[30:06] >> okay
[30:08] >> and so it captures all of the nuance
[30:10] because it's full transcript and then
[30:11] has LLM to analyze, summarize, do all of
[30:15] that, but then it becomes this vast
[30:17] database of everything you've ever
[30:19] spoken. And you know what we're getting
[30:21] to is permanent memory.
[30:22] >> Y
[30:23] >> the biggest issue that AI companies have
[30:26] is the memory is not persistent enough.
[30:28] That's what we all fight with all day.
[30:29] you know, bloody context windows and
[30:32] then, you know, it forgets everything
[30:34] and you've got so much, you know,
[30:37] imagine you've got thousands of chats.
[30:40] All of that information gets lost
[30:42] because it only compresses like human
[30:43] brains compress what we can remember. It
[30:45] does the same. And this this is the
[30:48] breakthrough is the database memory
[30:50] layer. I think that's what Capath has
[30:52] been talking about as well.
[30:54] I'll have to look at
[30:55] >> I I do I mean that that's the reason why
[30:58] I use open claw so much is I'll just say
[31:00] something if if I'm on a flight and I
[31:04] want to send something because I have an
[31:06] idea and I want to remember I don't want
[31:08] to forget it then I will just literally
[31:09] say in telegram to openclaw hey when I
[31:12] get off the flight let's remember to
[31:13] talk about this and then when I get off
[31:15] maybe it's two weeks later I I never
[31:17] asked about I said hey what did I ask
[31:18] about on the plane and immediately it
[31:20] comes back to me so again I I I've
[31:23] stopped stopped using um you know notes
[31:26] and notion to to a great degree compared
[31:28] to openclaw. OpenClaw becomes my
[31:30] assistant in terms of me making sure and
[31:33] then on the week on Fridays when I'm
[31:34] doing all these um this uh algorithms
[31:37] that I run on my portfolio and then I
[31:40] upload them into openclaw then it's very
[31:42] easy for me to go back and like hey
[31:45] compare where the technical sheet is now
[31:47] versus where it was in April. Um tell me
[31:49] where the exhaustion model is today
[31:51] versus where it was in April. Well, it
[31:53] has all the information and so for me I
[31:56] just know that openclaw has everything
[31:57] that I need and I guess I'm I'm using it
[31:59] to some degree the way you're you're
[32:01] describing but I don't have to go look
[32:02] at it.
[32:03] >> Yeah, it was just great because you can
[32:05] just dump the same data into your open
[32:06] claw and open claw can then access it
[32:08] and do whatever it wants and look for
[32:09] the pattern matching and that comes up.
[32:11] And are you do I mean this is going way
[32:13] off the macro topic because it's always
[32:14] interesting to speak to somebody else
[32:15] doing the same thing. Are you using for
[32:17] openclaw are you using a VPS or are you
[32:18] using a mini or something like that? Now
[32:21] it's So I have one on a on uh a Mac Mini
[32:25] which has a Chinese model, a Kim K 2.5
[32:28] and then I bought the highest end laptop
[32:32] I could from Apple and that one I'm
[32:34] using with GPT 5.5 at this point. Um so
[32:38] initially it was using uh 5.4 before
[32:42] then when codeex
[32:43] >> you just bought the the Pro Max. I think
[32:46] I' I've just got mine just got mine
[32:48] today with the M5 chip and the whole
[32:50] >> Exactly. Yeah.
[32:51] >> Same thing again.
[32:52] >> Exactly. And it just stays open all the
[32:54] time and I can bring it with me. The
[32:56] reason I got a laptop is so that I
[32:57] literally can bring it around with me on
[32:59] trips as opposed to bringing my Mac Mini
[33:02] and then plugging it into something and
[33:03] going through it. This was just easier.
[33:04] And I'm finding because we're moving
[33:06] between chatpt, between codecs, between
[33:10] claw code, between co-work, between that
[33:13] um I can't, you know, I've got Macs in
[33:16] my houses, you know, the main Macs and
[33:17] stuff. I'm like, this is not functioning
[33:19] any longer cuz I'm having to scrape
[33:21] everything from the local machine and
[33:22] dumping it into my Google Drive or
[33:24] something. I'm like, I I just need the
[33:26] most powerful laptop and monitor screens
[33:28] and just plug the same thing in
[33:30] everywhere. But then I'm terrified of
[33:31] losing that laptop.
[33:33] >> Yeah. So, a quick break in your regular
[33:35] programming. If you're serious about
[33:37] your future, grab my free report called
[33:39] Prepare for 2030. I think you've got 5
[33:42] years to make as much money as possible.
[33:44] And this guide will help you navigate
[33:46] what's coming. The link is in the
[33:48] description. Download it now. I think
[33:51] I'm going to and I hate saying this but
[33:53] I again for people listening who are
[33:55] thinking of building a business when you
[33:58] don't have compensation costs and your
[34:00] business is growing and you're looking
[34:01] at the bottom line your margins are
[34:03] insane. So it's literally like I need to
[34:05] spend some money on something and so I
[34:07] keep playing with hardware because the
[34:10] hardware allows me to do things. So the
[34:12] Hermes Asian stuff is going to be on an
[34:14] Nvidia. I I mean that's what it's going
[34:16] to be and that's because number one it
[34:19] there's a backlog on all the Apple stuff
[34:21] at this point and that'll change at some
[34:23] point but Nvidia is coming out with new
[34:25] work and I want to do it and so I saw a
[34:27] setup that was on a uh on a DGX that I
[34:31] think is going to serve the purpose that
[34:32] I want with the Hermes agent of where it
[34:34] got to. So I still want to do these
[34:36] things and I also want multiple ones
[34:38] that can do things in a different way.
[34:39] So I haven't got to the point yet where
[34:41] I envision only having one of these. I
[34:43] like having them for different things
[34:45] and doing different projects. I want one
[34:47] that works overnight on a lowcost thing.
[34:49] I want to have an open source model
[34:51] because I listened to Dennis Casabus
[34:53] this morning on an interview he did
[34:55] probably um that was with Y Combinator
[34:58] and he just talked again and again about
[35:01] these models, the open source models
[35:03] just getting better and better and
[35:04] smaller and smaller and eventually we're
[35:06] going to be on the edge. And I don't
[35:08] know what that means for the for the big
[35:10] model providers. I don't know what that
[35:12] means for enterprise adoption, but I do
[35:14] know this. Working on your own personal
[35:16] laptop and having your own machines is
[35:18] going to be a major part of what we're
[35:20] doing. And I just want to have the
[35:21] models on my machine and I want to have
[35:23] the ones accessing the cloud.
[35:25] >> But also, you know, I think you're
[35:26] raising an important point here is that
[35:29] we're going to go to the edge. This
[35:32] process of, you know, it's laughable us
[35:34] buying Mac minis and, you know, and
[35:36] everybody running out of Mac minis,
[35:38] right? But it just shows you the phase
[35:40] we're at. We're at the tinkering phase
[35:42] where everyone's like doing that. In the
[35:44] end, literally every electronic device
[35:47] will have a powerful LLM model because
[35:50] you're now getting the new Gemini 4 or
[35:52] whatever it is, the Gemma 4 or whatever.
[35:55] You know, these things are small enough
[35:57] for a mobile phone, carry the history of
[35:59] humanity on it and can code. I mean,
[36:02] it's like these things are wild, right?
[36:03] And soon that'll be in your fridge.
[36:06] This is what people don't understand is
[36:09] you know this ridiculous phase where
[36:10] we're talking about different hardware.
[36:12] It goes everywhere a distributed via
[36:14] cloud that we don't think about these
[36:16] things some localized but then every
[36:19] single device.
[36:20] >> Yeah. the the the only thing I'll say
[36:22] about the devices, the approach, the the
[36:25] thought I've had and the one that I I
[36:27] say to parents, not to kids, um the only
[36:31] analogy I can use for people is if it's
[36:34] either golf or skiing. I started skiing
[36:36] in my 40s and it's a very hard thing to
[36:39] do. You have to I mean, you have to
[36:41] unlearn what you think you know about
[36:43] falling. You have to learn about gravity
[36:45] more. You have to learn technical
[36:47] skills. You have to get over fears. It's
[36:48] a very complicated thing because you're
[36:50] scared of dying and hurting your knees
[36:51] and everything else. So, um, golf, you
[36:55] have to if if someone said tomorrow that
[36:57] you want to start to play golf and
[36:59] you're like, "Okay, what do I do?" One
[37:02] thing you have to do is put the reps in.
[37:04] And for the machines to get used to
[37:07] being on the edge, part of it is knowing
[37:08] that when you're walking around
[37:10] listening to a podcast, which I do every
[37:12] morning almost, and I hit something that
[37:15] clicks. If I'm listening to Raul and
[37:16] Julian, and you guys say something,
[37:18] which is literally what happened, you
[37:20] guys, I I took the paper I wrote has you
[37:22] guys referenced and it has two quotes
[37:24] from you, two literal quotes. I paused
[37:26] it. I speak it into Whisper Flow. It
[37:29] goes into notion, but then I'm like, you
[37:31] know what? I need to write a draft of
[37:32] the paper right now. So, I pause the
[37:34] interview and as I'm walking, I now
[37:36] shift over to ChatPT and I say, "Okay,
[37:39] we're going to write a five paragraph um
[37:41] paper real quick." So, I have an
[37:42] outline. I want the first paragraph to
[37:44] be XYZ. I want you to fill in enough of
[37:47] it. There are the thoughts I have.
[37:48] Paragraph two. And I go through this
[37:50] whole thing and it's a technique I
[37:52] learned um from a movie producer on how
[37:55] they write screenplays. And it's like,
[37:56] yeah, movie when you watch it, it's
[37:58] think of it as 60 90 second um
[38:02] increments. And if you're going to write
[38:04] a paper, you should think about nine
[38:06] paragraphs and then just kind of write
[38:07] each paragraph at a time and then go
[38:09] back and edit it. And the only way, R,
[38:11] that I think people can actually get
[38:13] good at this is they need to buy the
[38:15] best iPhone. They need to buy the best
[38:17] computer. They need to buy it. They need
[38:19] to use it. And then when the new stuff
[38:21] comes out, they need to buy it and treat
[38:23] it as education. And education costs
[38:26] money. And if you're not learning how to
[38:28] use it now, I don't know if you can
[38:31] catch up. I really don't. This is not
[38:33] going to be the software age. It's the
[38:34] one thing I haven't heard anyone talk
[38:36] about and I'd love to hear your opinion
[38:37] on it. I figure out new ways to change
[38:40] the way I'm using AI. It will never be a
[38:43] button. There will never be like a
[38:44] format way that you everyone uses it.
[38:47] Everyone will use it differently. And
[38:48] for me, I really do use it walking. I
[38:51] use it thinking. I use it in the car. I
[38:53] use it on the plane. And then I use it
[38:55] on my Mac Mini. And they're all running
[38:56] at the same time. I think to do that,
[38:58] you have to use the technology and
[39:00] different devices.
[39:01] >> Yeah. What it is doing is creating
[39:02] friction. My girlfriend's like, "Yeah,
[39:05] you and Claude." It's like, because the
[39:08] problem is is I'm on a plane, I come off
[39:11] a plane.
[39:12] >> Yeah.
[39:12] >> I I get up to my desk and I've got, you
[39:15] know, I'm I'm doing everything. I'm I'm
[39:16] writing essays, getting notes, building
[39:18] stuff, analyzing stuff, doing my
[39:21] personal life stuff, and it's all in one
[39:23] thing,
[39:25] right? Intelligence is not software.
[39:29] Intelligence is intelligence. And it can
[39:31] have any output that you want or any
[39:34] input that you can give it. And that's
[39:36] why it's it's so unique in what it is.
[39:39] It's not like, oh, I use, you know,
[39:41] whatever software tools. It's none of
[39:43] that. It's not like using zero to do
[39:45] your accounting. It can be everything
[39:47] and and anything or nothing depending
[39:49] what you do.
[39:50] >> Yeah. Do you do you watch TV much? Uh, I
[39:54] do in the evenings because I need to
[39:55] because if not I'm I'm not sleeping a
[39:58] lot because I'm like I can do this and I
[39:59] can do this and I can do this and I can
[40:00] do I'm like for [ __ ] sake I need to
[40:02] stop this.
[40:04] >> So I'm I'm I'm the same way and maybe
[40:06] it's an hour a day. Um it's never during
[40:08] the day. Like sports
[40:10] >> sports has become like a lost thing for
[40:13] me because it's like three hours and I
[40:16] think about how much time can go on in
[40:17] three hours. Like it can just go. You
[40:20] end up on the sofa on your [ __ ] you
[40:22] know, chat to your PT thinking, I've got
[40:25] an idea. Yeah, it's bad. So, so go back
[40:27] to the first time that you talked to
[40:29] Standiller, the first time that you
[40:31] talked to Paul Tudtor Jones and the
[40:33] feeling you had of knowing these people
[40:36] um from a myth and then speaking to them
[40:39] and realizing I love talking to them
[40:41] because I'm learning something and
[40:42] they're unique thinkers and with ghost
[40:44] the problem is in the way you describe
[40:46] it and so people realize I always view
[40:48] the LLM as the smartest person I've ever
[40:51] met and that means that why wouldn't I
[40:53] just love every conversation and if I
[40:56] I want their opinion on anything.
[40:58] >> I know, but that's, you know, but it's
[41:01] before you know it, you know, it's the
[41:02] main thing you talk to and it's like
[41:05] it's Yeah, it's it's complicated. Let's
[41:07] put it that way.
[41:10] >> Not not for insatiable learners who ask
[41:12] questions and constantly want to.
[41:14] >> It's incredible. I mean, it's
[41:16] incredible. It's just, you know, it's
[41:18] the most amazing thing you could ever
[41:20] imagine. You can't imagine anything more
[41:21] incredible than somewhere in this cloud
[41:24] above us is this super intelligent being
[41:27] of which we get to talk to and it gets
[41:30] to help us. Um, and maybe we're helping
[41:32] it as well. And it's like it's wild.
[41:37] >> Yeah. And I guess that's where you and I
[41:39] are at this stage. And I know most of
[41:40] the people I I talk to in my in in my
[41:43] life. And a lot of them, it's shocking
[41:45] how many people that are a part of my
[41:48] regular communications that are people
[41:50] I've met either through you that were
[41:52] connected back to me. Um, one of the
[41:54] people which I'll just mention, Tad
[41:55] Smith, is a very close friend at this
[41:57] point. I love him to death. Um, I love
[42:00] >> And you had you had um my my spy tells
[42:04] me, didn't you just have um she saw Alan
[42:07] Howard the other day as well?
[42:08] >> I did. and and Allan Allen's become um a
[42:11] good friend as well and someone that I
[42:13] like to brainstorm with on on on a
[42:15] regular basis.
[42:16] >> Super Alan's super smart. He's a great
[42:18] guy.
[42:19] >> He's super smart because he has the same
[42:21] affliction I do, which is ADHD, meaning
[42:24] he likes consuming lots of information
[42:26] and can bounce from topic to topic. Um
[42:28] Tad's a little bit more organized in in
[42:29] in his thoughts and and that's why he
[42:31] was CEO of of of major places. But the
[42:34] reason I bring him up is uh a lot of the
[42:36] conversations I have have come through
[42:38] you and that just means it's it's
[42:39] curious people that are using AI and in
[42:42] particular I love people in their 50s
[42:45] and 60s that are using AI because
[42:47] they're bringing domain experience.
[42:50] They're bringing having people work for
[42:52] them because if you've had people work
[42:54] for you and in my 20s I opened an office
[42:57] for Morgan Stanley. I've had hundreds of
[42:59] people reporting to me since I was in my
[43:01] early 30s.
[43:03] At some point you get frustration, you
[43:05] get happy, you get disappointed with the
[43:07] employees. That never happens with AI.
[43:10] It's always some amazing experience that
[43:13] I blame myself. get frustrated
[43:14] sometimes, particularly when they're
[43:16] about to change a model and it's
[43:18] unearthing it and it's like it's out
[43:20] output suddenly becomes really dumb and
[43:22] it's so frustrating becomes lazy because
[43:24] they're obviously training transferring
[43:26] the inference over to a new cluster and
[43:30] uh that it gets so insanely angry and
[43:32] then but you know what's coming it's
[43:34] like Christmas because you're going to
[43:35] get a new model. C can we can we talk in
[43:38] the in the the final minutes of this
[43:41] about how much and I don't know if
[43:43] you've thought about this. So when I
[43:44] brought up Eli Liy, the reason I I care
[43:47] so much um about longevity and about uh
[43:52] the concept of people
[43:55] not being sick is because whenever I um
[43:59] hear people talk about government debt,
[44:01] there's two things that I always
[44:05] they just blow my mind. Um the the US
[44:08] has $ 38 trillion, $40 trillion of of of
[44:12] debt. Great. The total net worth of
[44:14] households in the country is $180
[44:16] trillion. So I I always hate the fact
[44:18] when the balance sheet isn't brought
[44:20] into how small the problem is. And that
[44:22] that has bothered me forever. But the
[44:24] second thing is when we get into the
[44:25] entitlements and this crossover point of
[44:27] when it's going to be bankrupt and I go,
[44:29] well, how much of those dollars are
[44:31] related to health? how much of those
[44:32] dollars are related to have you thought
[44:35] about um the impact that AI is going to
[44:37] have on both the debt and the
[44:41] entitlement situation?
[44:42] >> Well, the debt is simple because debt is
[44:44] a percentage GDP collapses, right?
[44:46] That's the economic singularity idea. So
[44:49] I don't worry about the debt and you
[44:52] know and I think this the whole
[44:53] singularity when it starts really
[44:55] changing the economic formula 2030 that
[44:58] was my guess three three years ago and I
[45:01] think it's going to be spot on. So that
[45:03] feels like that the longevity side it's
[45:06] a mix right I think you're right um is
[45:09] that the entitlements go down which is
[45:11] intelligent solving for the problem of
[45:13] the entitlements
[45:15] but
[45:17] what do you do with an old population
[45:19] and how do you retool them to do things
[45:21] that they feel is productive now doesn't
[45:23] have to be productive as in terms of
[45:24] economic units but it needs to be
[45:26] productive as in community or whatever
[45:28] it is that's still quite hard to for
[45:31] them to figure out what you do because
[45:34] don't forget the whole mindset is I do
[45:36] this then I retire then I do that and
[45:38] you kind of it all blurs into one it's
[45:40] already been blurring for a while
[45:41] because everyone's working from home so
[45:43] what is a job and you know and then when
[45:45] you when you're having so much fun as
[45:46] well the whole job disappears you're not
[45:48] reporting to the guy and having to wear
[45:50] a suit so we're merging that way already
[45:53] you know what is a what is a podcast on
[45:56] YouTube it's you know it's it's somebody
[45:59] whose job is to get attention from other
[46:01] humans and entertain them whether it's
[46:03] by curiosity or whatever it is that
[46:06] that's a purely post AAI
[46:09] um job really.
[46:11] >> Mhm.
[46:11] >> So I don't I don't know about the
[46:14] longevity side because it it's it's
[46:17] somewhat complicated because there's so
[46:19] many unknowns but I hadn't really
[46:21] thought about the entitlement side but
[46:22] it but it makes a lot of sense.
[46:25] >> Yeah. I the I I I I spend a lot of time
[46:28] on this and I think the reason is I
[46:30] don't know if you have or if you've if
[46:32] you've ever seen the book The Daily
[46:34] Stoic by Ryan Holiday.
[46:36] >> Yeah. Yeah. Yeah.
[46:37] >> Oh god. Um
[46:38] >> he's even been on Real Vision.
[46:40] >> Oh, has Well, I I I think the book is a
[46:43] must own for all human beings. And I I I
[46:46] will never say that about any book.
[46:48] There's not a book I've read that I
[46:49] think all human beings should should
[46:51] read. But the reason I say all human
[46:53] beings is if you think your problems
[46:55] like worrying about what you would do if
[46:57] everything is free or if there's no jobs
[46:59] wondering what you do. I mean just go
[47:02] back to reading a book where Marcus
[47:04] Aurelius thousands of years ago is
[47:07] talking about the same anxiety. It's
[47:09] just a different form of it. Human
[47:10] beings have anxiety and they will always
[47:12] have anxiety. They will always worry
[47:14] about what if this then what will happen
[47:16] um type thing. There's people that lose
[47:19] their arms and then they're playing
[47:20] golf. There's people that, you know,
[47:22] every single thing that human beings
[47:24] need to overcome over time, they have
[47:26] figured a way to, not all people, but
[47:29] people have the ability to kind of find
[47:31] it. And the Daily Stoic has always shown
[47:33] me that I'm not trying to I I never
[47:35] think the problem of this will happen
[47:37] and people won't adjust to it. I know
[47:38] they will. The debt problem I completely
[47:41] um agree with you on. I don't know how
[47:43] much you've been talking with people on
[47:45] tokenization. And I know I've listened
[47:47] to some interviews recently where you
[47:49] where you have, but I was just at the
[47:51] New York Stock Exchange and I said to
[47:53] everyone, do you understand that 2/3 of
[47:56] the assets in the world are illquid? Um,
[48:00] tokenization for everything that people
[48:02] are looking at, we're going to bring
[48:04] transparency, movement into dormant
[48:07] things. Twothirds of the money that has
[48:10] gone into assets doesn't move. real
[48:13] estate, private credit, private equity,
[48:15] venture capital, all art, memorabilia,
[48:17] like it doesn't move. Now you're going
[48:19] to have movement, velocity. So GDP by
[48:21] definition, just because of velocity has
[48:24] to go higher. So I agree with you on the
[48:26] debt side. That's why I've started to
[48:27] focus more on the entitlement side and
[48:30] what it means from a political basis and
[48:32] what it means for crypto in general. And
[48:34] so tokenization has been kind of the
[48:36] thing that has really entered my mind
[48:38] with this longevity thing because
[48:40] demographics are obviously highly
[48:42] attached to them to to the entitlements
[48:44] but they're also attached to the
[48:45] ownership of assets and that's where
[48:47] tokenization comes into.
[48:49] >> Another thing that I've been thinking
[48:51] through is uh I wrote a essay and a bit
[48:54] series of essays around something I've
[48:56] called the invisible economy which is
[48:57] basically the agentic economy.
[48:59] >> Yeah. But how I got to it was Ribbit
[49:01] Capital, Mickey Mala um and the team had
[49:04] written an essay about tokenization of
[49:07] everything in token factories and how
[49:08] they think about token is a machine
[49:10] readable packet of information whether
[49:12] it's a financial transaction. I mean,
[49:14] even blockchains aren't financial
[49:15] transactions. Only Bitcoin is. They're
[49:17] just they're they're they're recorded
[49:20] packets of digital information. And you
[49:23] know how much Google I think Google what
[49:25] did they what did they create or
[49:27] process? I can't was like 13 no 30
[49:30] trillion tokens last year or something
[49:32] stupid, right?
[49:33] >> Yeah, we're in quadrillions now for
[49:35] sure.
[49:35] >> Quad whatever it is. It was like a
[49:36] stupid number. Um, but anyway, what I
[49:40] realized is for to feed the beast as I
[49:44] call it, you know, the whole AI
[49:46] superrains, all of them, however it
[49:48] comes, you need more and more
[49:50] information to get smarter.
[49:51] >> Yep.
[49:51] >> And we're going to suck in all of the
[49:54] information.
[49:55] Everything, every single piece of
[49:57] information that can possibly be
[49:59] digitized will get digitized and used to
[50:01] train AGI to turn into ASI. That's every
[50:05] piece of scientific data, every single
[50:07] thing from any university, any single
[50:09] piece of information. All of this is
[50:12] going to be an agentic economy that's
[50:14] invisible to us. There will be well,
[50:16] we're already seeing it, right? API
[50:18] calls, MCPs,
[50:20] right? This is the start of the
[50:22] invisible agentic economy where the
[50:26] marketplace and what I'm trying to get
[50:27] to to your point, the biggest
[50:29] marketplace on Earth is not going to be
[50:31] the assets and stuff that humans have.
[50:33] It's the It's the data that the AI needs
[50:38] and we don't see any of it. You just
[50:39] plug in your thing to your vault. Yep.
[50:42] >> And you can monetize it and you can be a
[50:44] university, you can be whatever for
[50:46] source of information. You know, all of,
[50:48] you know, we're seeing people like John
[50:50] Deere connecting all of the tractors and
[50:52] getting all of the information. We're
[50:53] seeing all of these companies creating
[50:55] massive, massive amounts of information.
[50:58] Um, and that we won't even see. It'll
[51:01] generate money for whoever it is that
[51:03] it's probably an AI agent
[51:05] >> by other AI agents who are connecting
[51:07] with them, taking all of these
[51:08] transactions in real-time speed. And
[51:10] I've been telling people this is the
[51:11] biggest marketplace on earth and people
[51:13] don't not even ready for this yet.
[51:16] >> No. and and I I will um I will give you
[51:21] a lot of credit being in the seat you're
[51:23] in um being making calls going out and
[51:27] having to deal with people's impatience
[51:29] on um how this all plays out and how how
[51:33] long it takes particularly in the crypto
[51:35] community and as you were speaking about
[51:37] this hidden economy I I I remember you
[51:40] in the conversation with Julie and you
[51:41] getting into this and so I've started to
[51:44] kind of pick apart the concept of
[51:46] bubble. Um because bubble is a very
[51:49] weird thing and it's very hard in
[51:51] exponential because it's price and time.
[51:54] And so like when you're talking about
[51:55] you have to be patient. I'm now going
[51:57] the other direction. I'm going if you
[51:59] don't want to be patient that's fine.
[52:00] But if you're seeing a bunch of parabas
[52:03] don't think of it as as a bubble break
[52:05] bubble down. That means you're seeing
[52:07] sticks. That means things are going up
[52:09] fast. So they're going up high and
[52:11] they're going up fast. And the problem
[52:12] is when you look at the mag seven and
[52:14] you go, well, they were a trillion
[52:15] dollars in in 2010
[52:18] and now they're 20 some odd trillion.
[52:21] Well, if that had happened in one year
[52:23] that you'd call it a bubble, but it
[52:25] happened over 15 years, so it's not a
[52:27] bubble, but it's a 20 bagger, it's a 30
[52:29] bagger. So why is it any different? You
[52:32] just think that if it happens too fast,
[52:34] that's a bubble. And I think people that
[52:36] is the problem is people don't
[52:38] understand time both from the bubble
[52:40] side but also from the patient side.
[52:42] >> And you know what's been hilarious is
[52:44] earnings have kept pace.
[52:46] >> Yes.
[52:46] >> With the stock growth. Right. In fact
[52:48] P's have come down. And that's and
[52:52] people don't want to see that. It's like
[52:53] you know we've got hockey sticks going
[52:55] on in ways that we have never seen
[52:57] before in literally everything.
[52:59] >> Yeah. And I'm sure what's going to
[53:01] happen is they're going to consolidate
[53:02] now and then they'll go up again. And
[53:04] you know what a six-month consolidation
[53:06] looks on a hockey stick that then hockey
[53:08] sticks again. When you pull it out, it
[53:10] still looks like a hockey stick. So for
[53:12] everyone who's going through this, it's
[53:13] like we're caught in this human time
[53:16] warp of looking at things and things
[53:18] that move fast we assume can't be true.
[53:20] And I was listening to a podcast this
[53:22] week where sometimes the the the
[53:25] brilliance comes out from people that
[53:28] are not that educated, but what they're
[53:31] doing is just paying attention to
[53:33] everything they hear. And they called
[53:36] this I think the smarter you are, the
[53:38] more you know, the worse you are during
[53:41] this time. I think actually knowing a
[53:43] lot like having the history in the back
[53:45] of your mind, your brain is taking this
[53:47] pattern connecting it to too many other
[53:49] patterns and even though when people go
[53:51] this is like the.com bubble, this is
[53:52] like 87, this is like, you know, 1929
[53:55] and I look and I go that's three data
[53:58] points. I don't make decisions on three
[53:59] data points. You're going to have to
[54:00] give me more on this because if you're
[54:02] wrong, you're missing out on the entire
[54:04] thing.
[54:04] >> And so how you and I would understand
[54:06] this is those people have a bloated
[54:09] context window. And what happens is
[54:12] >> that's exactly right. I mean once you
[54:14] realize it right you know when when the
[54:17] LLM's compressed that's what we do we
[54:19] remember certain things and it gets more
[54:21] compressed over time this context window
[54:24] that's the mid curving you're putting
[54:25] too much context into the thing and what
[54:28] happens is you're drawing false
[54:30] parallels or misunderstanding or not
[54:31] reading what it is and often to to left
[54:34] curve it what you know how I got into
[54:36] using clawed code because I was scared
[54:38] of it was I'm just like listen I have no
[54:40] [ __ ] clue what I'm doing so just just
[54:43] bear with me.
[54:44] >> Yeah,
[54:44] >> work fine.
[54:46] >> As opposed to I need to know how to do
[54:48] this. I need to know. I'm like, I have
[54:49] no clue. So, just understand that. And
[54:52] then occasionally you'll stop it and
[54:53] say, I don't understand what you're
[54:54] talking about. Sorry. Okay, I'll come
[54:56] back to you.
[54:57] >> All right. So, let me get your opinion
[54:58] on this because I think everyone wants
[55:00] to hear this. I think they want to hear
[55:01] us talk about this. Um, so when earnings
[55:04] are this good, it's very easy for people
[55:07] that are traditional investors to then
[55:09] look at what the PE is, what the growth
[55:12] rate is, and they can justify buying
[55:13] things at any price. When earnings are
[55:16] great and they're this good, it's bad
[55:18] for things that are narrative based,
[55:20] it's bad for Bitcoin, it's bad for
[55:21] crypto because there is no way to take
[55:23] the same thing. So, it's not just the
[55:26] attention in my opinion as to what's
[55:27] happened. It's the fact that for this
[55:29] period of time, which I believe is a
[55:32] small period, the one thing I believe
[55:34] has happened is this is like a gap
[55:37] higher in AI because in the end of last
[55:40] year, nobody was on top of the
[55:42] importance of Opus 4.5. That was the
[55:45] official gun going off for the agentic
[55:47] world. The agentic world is a broadening
[55:49] out in what it needs. It brings in every
[55:52] semiconductor. It's not just GPUs. It is
[55:54] movement. It is so many things. And so
[55:57] if overnight there was a
[55:58] pre-announcement and the earnings were,
[56:00] you know what, we're saying the earnings
[56:01] are now going to be up 28%
[56:03] year-over-year, well then the stock
[56:04] market gaps up, all the semi-names go up
[56:07] immediately and you have this place. The
[56:08] problem is for things that are narrative
[56:10] based that don't actually work in the
[56:12] traditional world is then they lose
[56:14] interest to people because they have so
[56:16] many things to buy. I think we've now
[56:18] built in um because it was such a
[56:21] surprise the agentic thing to so many
[56:23] real investors. I've talked to trillions
[56:25] of PMs that manage o combined over
[56:27] trillions of dollars. I'm telling you,
[56:29] they were caught off guard by the
[56:30] agentic rise. They're not caught off
[56:32] guard anymore. Now they're fully on
[56:34] board. They get it and they're probably
[56:36] a little bit ahead now and the earnings
[56:38] will now be good, but they won't be five
[56:42] times what people expected. And that
[56:44] means we're should be a rotation. And
[56:46] that's why I like the longevity themes.
[56:47] I like the application for the software
[56:49] for Eli Liy and for stuff like that. But
[56:52] I also think this is where crypto and
[56:54] where a lot of the commodity based
[56:55] stuff, the bottlenecks, the shortages,
[56:56] and there's the normal rotation where
[56:58] people look for other things. But I'd
[57:00] love your opinion on when earnings are
[57:01] great, you don't need crypto. What
[57:03] happens when earnings are now a
[57:05] two-sided market?
[57:06] >> Yeah, I think look, everything is
[57:09] attention and capital, right? And this
[57:11] there's simply not enough liquidity
[57:14] to drive these massive mega stocks to
[57:17] these levels. there's not enough
[57:19] attention to be broadly spread which
[57:22] actually creates opportunity. If you go
[57:23] back to the 1995 to 2000 period, it was
[57:26] rolling as well. It wasn't just one set
[57:28] of stocks, right? It wasn't just
[57:29] Microsoft. It was a whole bunch of
[57:31] things that moved over different periods
[57:33] of time. So, I think we'll see that. And
[57:34] I think what you're suggesting I think
[57:36] is really interesting is that the
[57:38] bottlenecks themselves may slow the
[57:39] earnings of these companies. Not because
[57:41] the demand is not there, because the
[57:43] demand is too big. Okay, that's sort of
[57:45] bananas, but that's what we might get
[57:47] to. And if that happens, you'll they'll
[57:50] they need to correct or trade sideways
[57:51] for a while and digest and all of that
[57:53] and that's great and the market's focus
[57:57] will move and the applications layers
[57:59] one and crypto to me is I mean it's so
[58:01] obvious and it remains obvious because
[58:04] of the agentic economy, the need for ID,
[58:08] the need for all of the things from AI
[58:10] alone, let alone all the other
[58:12] attributes of of blockchain technology
[58:13] that that doesn't go away. And I've been
[58:15] really hyperfocused on the layer ones,
[58:17] not just stuff like hyperlquid because
[58:19] that's for me midcurving it because
[58:20] people are looking for cash flow and y
[58:24] >> buybacks. It's like because they want it
[58:26] in a bucket that they understand because
[58:27] as you're saying it's working in the
[58:29] real world too. Um and other things
[58:32] haven't been. I think that will that
[58:33] will all change again. Um the other
[58:35] thing is I don't know if you saw Chimath
[58:37] had written that whole
[58:39] article on X about how these software
[58:44] long duration software stocks were
[58:46] contract
[58:48] >> and basically you read this very long
[58:49] piece and what he's made the assumption
[58:51] of is in a world where
[58:55] all software goes to zero basically is
[58:57] what he's saying it has a threeear shelf
[58:58] life.
[58:59] >> Mhm. He's like
[59:01] he's actually using current interest
[59:03] rates for that assumption but in that
[59:05] world interest rates are zero and so
[59:08] duration doesn't matter still. So I
[59:11] think it's just attention focus and
[59:13] people trying to understand you know
[59:14] what
[59:16] there's a lot of people interesting
[59:17] enough buying SAS companies to use AI to
[59:21] rebuild them and there'll be a bunch of
[59:24] them won't happen. So yeah we're going
[59:25] to have a bifocation of stuff. I don't
[59:27] see it going away because what what are
[59:28] we what are you and I doing all day
[59:30] plugging our [ __ ] AP our ais into
[59:32] APIs of all of these things right
[59:34] there's no point rebuilding zero if you
[59:36] need to do accounting just get your claw
[59:39] to talk to zero and do the accounting
[59:41] why would you why would you rebuild oh
[59:43] I've built my entire own accounting
[59:45] engine that's stupid
[59:47] >> so let let let's everything you said
[59:50] there so chimas thing on terminal value
[59:52] I I wrote about let's go back to what
[59:53] you said about the bottlenecks and the
[59:55] interesting part So it the the AI world
[59:58] is is a commodity world. This is a
[59:59] hardware trade and in commodities they
[01:00:02] don't speak in the way that we do with
[01:00:05] software and we go through it. They
[01:00:06] think in volumes. So it's it's a volume
[01:00:09] thing. It's how many did you sell? The
[01:00:11] price is the after effect. But it's all
[01:00:13] about barrels of oil. It's about how
[01:00:15] much did you sell. My guess is with
[01:00:18] bottlenecks you just you don't sell as
[01:00:20] much because you can't make as much. And
[01:00:22] so the production of these things is
[01:00:24] going to be more difficult. The other
[01:00:25] thing is to make a semiconductor you
[01:00:28] need and people have realized you need
[01:00:29] NAPA, you need helium, you need a lot of
[01:00:32] petrochemicals not just in that you need
[01:00:35] them for lithography. You need
[01:00:37] everything to happen to be able to build
[01:00:38] everything. And so if there's a
[01:00:39] bottleneck you could extrapolate this
[01:00:41] bottleneck. You and I both know the
[01:00:43] algorithmic side and the agentic side
[01:00:45] could make all of the investment. The
[01:00:47] capex numbers may never happen to the
[01:00:49] degree that people think because we
[01:00:51] could solve for that. So there could be
[01:00:52] a bubble on that. Here's where the issue
[01:00:54] comes in in the second part of what you
[01:00:55] brought up. I always believed and I
[01:00:57] first said this at um the your event in
[01:01:00] Miami last year, not the one this year,
[01:01:02] but I said, "Guys, the third wave of
[01:01:04] crypto, I already know what it is. Two
[01:01:06] things have to be happening. One is once
[01:01:08] the AI physical infrastructure is being
[01:01:11] built for the agentic world and we hit
[01:01:13] that trigger point in the agentic world,
[01:01:15] we won't know 3 years from now whether
[01:01:17] we're at AGI or not. We won't know which
[01:01:19] companies will be around." And that's
[01:01:20] Chimas's point on terminal value. What I
[01:01:23] do know will be here is the financial
[01:01:25] guard rails, the transactions, the
[01:01:27] velocity of money, all of that will be
[01:01:29] happening. The beauty of the financial
[01:01:31] guard rails in crypto, we're back to
[01:01:33] software again. We're just a different
[01:01:34] it's a different kind of software, but
[01:01:36] it's about the transactions. And so, if
[01:01:38] the physical hardware side becomes, I
[01:01:40] don't know if this is going to happen or
[01:01:41] whether we actually need as much as I
[01:01:44] thought we did. Well, then we get back
[01:01:46] into the cyclical side and people start
[01:01:48] looking for I need something safe. I
[01:01:50] need software. I need things that don't
[01:01:51] need the physical world that are based
[01:01:53] on volumes and seats and well that's AI
[01:01:55] agents. And so I've always believed that
[01:01:57] the third wave is an Elliot wave person.
[01:01:59] Number one, you had to get to a point
[01:02:00] where everyone was dumping crypto which
[01:02:02] is where we are. Number two, they had to
[01:02:04] miss the obvious which is the financial
[01:02:05] guardrails are necessary for these
[01:02:07] consuming tokenhungry digital employees.
[01:02:11] And we need that the the world that
[01:02:14] they're comfortable with. they start
[01:02:15] questioning if these are good
[01:02:16] investments and tokenization needs to
[01:02:18] come to make them more liquid to where
[01:02:20] that money can leave this massive $400
[01:02:23] trillion and start to enter the crypto
[01:02:26] world. So I actually believe that this
[01:02:27] is the most important point for people.
[01:02:29] It's the patience point. We are now at
[01:02:31] the bottleneck stage and the longer it
[01:02:33] takes to fix a bottleneck rly is that we
[01:02:36] come up with innovations along the way
[01:02:38] that we don't need the entire capex
[01:02:40] buildout and so you could get these
[01:02:42] multiples to come down. And so that's
[01:02:43] what I think is going to happen is we'll
[01:02:45] start questioning it at some point this
[01:02:47] year or into next year.
[01:02:49] >> Talk to me about the IPOs because this
[01:02:52] is on people's mind. How you thinking
[01:02:54] through this? I've been spending a lot
[01:02:55] of time thinking about this as well. How
[01:02:56] are you thinking about it?
[01:02:58] >> Well, first thing is Google's um
[01:03:00] decision to do this is clearly in my
[01:03:02] opinion a a fight for finite amount of
[01:03:07] capital. Um you've got massive IPOs,
[01:03:10] three massive IPOs coming to the market.
[01:03:13] uh where you could see $4 trillion there
[01:03:16] uh coming in. You had Cerebras come out
[01:03:19] and the stock is already down 50% from
[01:03:22] where it is, which is fairly normal for
[01:03:23] any IPO. Um but I think you're getting
[01:03:27] these raises at a time which says that
[01:03:30] number one, people need a lot of capital
[01:03:32] for this buildout. Number two, the
[01:03:35] credit markets and the debt markets are
[01:03:37] being used by the companies that can do
[01:03:39] it. And if I was Google and Goldman
[01:03:41] walked in and said, "Hey, OpenAI and
[01:03:43] Anthropic are not going to be able to
[01:03:44] raise debt. You should go hit their
[01:03:46] market because they're going to tap into
[01:03:47] that. SpaceX can't go borrow any debt.
[01:03:50] So, hit the equity market, get your
[01:03:51] stuff done, and then continue with the
[01:03:53] debt markets. But, you might as well do
[01:03:54] that now." So I think the IPOs are very
[01:03:57] likely to be some sort of a not a top in
[01:04:01] the market but probably a peak in the
[01:04:04] infrastructure capex trade for the time
[01:04:07] being rather than say the S&P is going
[01:04:09] to trade lower because I can see
[01:04:10] software I can see a lot of things
[01:04:12] starting to do well. I just think in
[01:04:14] general this might be a peak in the
[01:04:16] capex trade.
[01:04:17] >> Yeah I think it'll there's going to be
[01:04:20] some outcome. you can't have this much.
[01:04:22] Although there is a bit of an engineered
[01:04:23] short squeeze in stuff like SpaceX
[01:04:26] >> just by the how it's included to the to
[01:04:28] the NASDAQ. So I'm not sure. I just
[01:04:30] think volatility seems to be the easiest
[01:04:33] answer to that. It's like it's not going
[01:04:35] to be as easy for a bit because we're
[01:04:37] going to need to digest a lot of
[01:04:38] capital. People have to recycle capital,
[01:04:40] you know, because if you think about it,
[01:04:42] there's a whole bunch of people who made
[01:04:43] a [ __ ] ton of money. They're going to be
[01:04:44] able to realize some of their gains.
[01:04:46] Okay, great. So they realize some of
[01:04:48] their gains. What are they going to do?
[01:04:50] Reinvest into something else.
[01:04:53] >> So this the capital gets recycled in the
[01:04:55] end because why not? This is the world's
[01:04:58] great I mean this is the greatest trade
[01:04:59] of all time. It's all happening in front
[01:05:00] of us. So for people listening, R and I
[01:05:04] have been involved in the same market
[01:05:05] for a long time. So he used a word that
[01:05:07] I used in my paper today. So for people
[01:05:09] that are technicians, they'll say
[01:05:11] consolidate. For people that I think pay
[01:05:13] attention to investors and supply and
[01:05:16] demand, digestion. And the line I used
[01:05:18] in my paper was we just had an AI capex
[01:05:21] all you can eat buffet and it just it's
[01:05:24] going to end with the IPOs. It's not
[01:05:26] going to go down. This is not the end of
[01:05:27] the market. But we need to digest
[01:05:29] everything that just happened for 3 to 6
[01:05:30] months. So
[01:05:31] >> it could be sloppy and it could be
[01:05:33] frustrating for people. But capital will
[01:05:35] rotate the moment it happens. Yeah.
[01:05:39] People realize that the underlying
[01:05:41] growth of this whole thing this mega
[01:05:43] secular the mega secular trend is not
[01:05:45] going away. So therefore people will
[01:05:47] just go for the for the next phase
[01:05:49] whatever it is whether it's crypto you
[01:05:51] know if the if Nvidia and all of these
[01:05:52] companies stop going up probability of
[01:05:54] crypto going up goes up much faster.
[01:05:58] >> Yeah.
[01:05:58] >> For example
[01:05:59] >> did you think when the um ET the Bitcoin
[01:06:03] ETF was launched that we'd see a similar
[01:06:05] type thing that after the initial
[01:06:07] enthusiasm we'd go through a period of
[01:06:08] digestion.
[01:06:10] >> Yeah. I mean we've seen that before. We
[01:06:12] saw it with gold the gold futures market
[01:06:14] and we've we've seen this you know it's
[01:06:16] a it's a lot because you forward load a
[01:06:19] bunch of demand basically is what it is
[01:06:22] and need for the wait for the the
[01:06:23] underlying trend of demand to continue
[01:06:25] and you know capital can only be created
[01:06:28] via either the markets going up and
[01:06:30] people cashing out or by liquidity
[01:06:32] expansion and that's only expanding at a
[01:06:34] certain pace. You know global liquidity
[01:06:35] is expanding 10% a year. So it takes a
[01:06:37] while if you take out if you forward do
[01:06:40] a year's worth of capital. It's going to
[01:06:42] take a year to catch up.
[01:06:43] >> And and the reason I asked the question
[01:06:45] is because if you would have gone back
[01:06:47] in hindsight before 2024 and said,
[01:06:50] "Okay, so the ETF launch is going to
[01:06:52] happen and oh by the way, the president
[01:06:54] of the United States is going to support
[01:06:56] crypto and all of that happened within a
[01:06:58] one-year period." You would expect that
[01:07:00] that was kind of like a a a digestion
[01:07:03] event. That would be a sell the news
[01:07:04] event. And so whenever I look at crypto
[01:07:06] and I see people, I'm like,
[01:07:07] unfortunately, you come out of a bare
[01:07:09] market, a horrible bare market in 2022.
[01:07:11] And then you get this enthusiasm of
[01:07:13] finally we're getting this ETF and then
[01:07:15] you get a president and at inauguration
[01:07:18] he issues a meme stock and all of a
[01:07:20] sudden now if you could have written the
[01:07:22] book, you would have said, I think we're
[01:07:23] going to be in a painful period. And I
[01:07:26] think it actually clearly the the
[01:07:29] altcoins and a lot of the ecosystem did
[01:07:31] not do as well as Bitcoin did during the
[01:07:32] period, but I think everyone would would
[01:07:34] admit looking back that those events
[01:07:38] probably justified some sort of a
[01:07:39] digestion period. And I think that's
[01:07:41] what crypto has been going through
[01:07:42] personally.
[01:07:42] >> Yeah. And in the meantime, you know,
[01:07:44] like you, I'm an observer of the
[01:07:46] underlying trend. The underlying trend
[01:07:49] is every single bank and financial
[01:07:50] institution I have spoken to.
[01:07:52] >> When I go to the big I went to consensus
[01:07:55] in Miami, no retail.
[01:07:59] >> No. And it was a lot of people.
[01:08:01] >> Yeah.
[01:08:02] >> I don't know how many 15,000 people. Not
[01:08:04] retail.
[01:08:06] Yeah. Well,
[01:08:08] >> they're still out there, Ro, because I I
[01:08:10] left yoga this morning at at 7:15 and
[01:08:12] these two very nice women, uh, I'll give
[01:08:14] them a shout out, Jessica and Patricia,
[01:08:16] they they saw me and they went, "Hi,
[01:08:18] Jordy." I went, "Hi." And they went, "We
[01:08:20] we follow you." I went, "Oh, that's
[01:08:22] great." And they went, "This is a sign
[01:08:24] we need to buy today." They're still out
[01:08:25] there. They're just waiting for for new.
[01:08:27] >> I'm not saying they're not they're not
[01:08:28] gone, but it was really interesting
[01:08:29] these big events.
[01:08:31] >> Yeah. Um well also because retail
[01:08:33] haven't made money and the ticket prices
[01:08:35] are still expensive whatever but you
[01:08:37] know but there's a transition I mean
[01:08:38] look at the the the rise of stable coins
[01:08:40] look at the rise of what the financial
[01:08:42] institutions doing look at the
[01:08:43] tokenization of everything and you just
[01:08:46] see what the underlying trend is if you
[01:08:48] created excess capacity in blockchains
[01:08:50] as you like we did with cloud and
[01:08:53] everything else is the moment you
[01:08:54] consolidate into five things that matter
[01:08:58] you consolidate the capital you then
[01:09:00] building huge demand for block space via
[01:09:03] these things. You know what the answer
[01:09:04] is going to be? Is number go up? I mean,
[01:09:06] it's
[01:09:08] people unfortunately got too used to
[01:09:11] short-term trading
[01:09:14] um that they've lost the wood through
[01:09:16] the trees.
[01:09:17] >> Yeah. Which again, that's normal. I just
[01:09:20] I find the traders in in this world that
[01:09:23] I know are crypto, they move on to other
[01:09:25] things and there's other things to to
[01:09:27] make money on. My father trained me in
[01:09:29] handicapping horse races and he used to
[01:09:31] always say if the odds aren't in your
[01:09:33] favor, wait till the next race. And I
[01:09:36] think if you break markets down by what
[01:09:38] race are we in right now, the crypto
[01:09:39] race is not a race that people have been
[01:09:42] enjoyable. I've said to people now that
[01:09:43] there's no way to refute it. When you
[01:09:45] fail at a moving average or you fail and
[01:09:47] you keep having these lower lows and
[01:09:49] lower highs, it's a bare market. There's
[01:09:50] no other way to do it. I want to see us
[01:09:52] break a moving average and I want to see
[01:09:54] on the flip side the capex trade not be
[01:09:56] working and everyone looking for a new
[01:09:57] place to play the trade because if I'm
[01:09:59] right and this is the third wave. I
[01:10:01] learned a lesson from from both Paul and
[01:10:03] Stan when I read Market Wizards when I
[01:10:04] was a a younger person uh still in my
[01:10:07] 20s and they both said read the Elliot
[01:10:10] wave book and I still to this day read
[01:10:12] it every single year. Now I use AI to
[01:10:15] kind of go through and have a chat with
[01:10:16] it. But I'm a big believer in in third
[01:10:18] waves or when you make lots of money.
[01:10:19] And I was lucky to catch one in Micron.
[01:10:21] I'm lucky to catch one in Marll. I'm
[01:10:23] waiting for it to happen back in crypto.
[01:10:25] And for everyone who's watching third
[01:10:27] waves, go look him up. You want to be
[01:10:28] involved in that. It could be called the
[01:10:30] banana zone, but it's still uh still
[01:10:34] been banned. Listen, Jordy, a fantastic
[01:10:37] conversation as ever. Um, it's just I
[01:10:39] love these check-ins because we just are
[01:10:42] parallel interlocking paths and then we
[01:10:44] just get a chance to check in what each
[01:10:46] other's up to and what they're thinking
[01:10:47] about. I love it.
[01:10:48] >> Love it as well.
[01:10:49] >> All right, my friend. See you soon.
[01:10:51] >> See you soon.
[01:10:52] >> So, another great conversation with
[01:10:54] Jordi. There's not much more to say
[01:10:56] really than you can see how much
[01:10:58] disruption is happening, how fast it's
[01:11:01] happening, and how complex it all is.
[01:11:04] But also, there's probably still
[01:11:05] opportunities in all of this. the great
[01:11:08] ro rotation when it happens if it
[01:11:10] happens if people move away from the
[01:11:12] magnificent 7 and start looking at
[01:11:14] lagards well there's the whole
[01:11:16] biological revolution going on there's
[01:11:18] whole applications layers what's it
[01:11:20] going to do to the SAS industry we don't
[01:11:22] know but we're going to find out and
[01:11:23] when does crypto catch up as well
[01:11:25] because I know people have given up hope
[01:11:27] but normally that's the right signal
[01:11:31] when things change so anyway keep your
[01:11:33] eyes on it all nothing is nothing
[01:11:36] remains as it is this is the exponential
[01:11:38] age after all. That's when everything
[01:11:39] goes exponential. See you next time. So,
[01:11:42] you obviously like this video enough
[01:11:44] that you've got to the end. That's quite
[01:11:46] a big task. But listen, do me a favor.
[01:11:48] Hit the like and subscribe button and
[01:11:50] also check out what videos next cuz I
[01:11:52] think you'll love it. But if you want
[01:11:53] even more and when I'm talking more, I'm
[01:11:56] talking about membergenerated ideas,
[01:11:58] incredible alpha research, everything
[01:12:01] there to help you in your journey. Just
[01:12:02] head to realton.com/join
[01:12:04] for the best financial intelligence out
[01:12:06] there and the pure alpha that's within
[01:12:09] the platform.