Why Bitcoin WINS No Matter What Happens to Inflation

[0:00] Here's the one thing I will say about
[0:01] innovation. When you want to invest in
[0:03] innovation in a deflationary world,
[0:05] you're trying to find what Jeff Bezos
[0:07] said during the last 15 years, which is
[0:09] when Bitcoin was created. And what did
[0:10] Jeff Bezos say? Your margin is my
[0:13] opportunity. There's no more margin now
[0:15] in code-based stuff. It is in a free
[0:18] fall. That is why software stocks are
[0:20] getting up. So,
[0:21] where Bitcoin benefits is it doesn't get
[0:23] hurt by that. What's going on, guys?
[0:25] Today, we got a great conversation with
[0:26] Jordi Visser. In this conversation, we
[0:27] talk about the decoupling between
[0:29] Bitcoin and software stocks. We talk
[0:30] about inflation versus deflation and
[0:32] what his expectation is. And then we get
[0:34] into the five different thematic baskets
[0:36] that he's been paying attention to, why
[0:38] they're up so much, and what exactly is
[0:39] happening in each one of them. And then,
[0:41] of course, we talk about the scarcity
[0:43] trade and why Jordi is surfting through
[0:45] all of the different markets trying to
[0:46] find things that are full of shortages
[0:48] or scarcity, and how he's thinking about
[0:50] investing his portfolio right now.
[0:52] Here's my latest conversation with Jordi
[0:53] Visser.
[0:55] All right, Jordi. It looks like software
[0:56] stocks have taken a beating into the end
[0:58] of the week, and Bitcoin has stayed
[1:00] pretty strong here. So, help me
[1:01] understand are software and Bitcoin
[1:03] decoupling, and is that a good sign for
[1:05] Bitcoin, or is that more of a bad sign
[1:07] for software stocks?
[1:09] Yeah, I think uh I think the reaction to
[1:12] Well, first of all,
[1:13] IBM reported, ServiceNow reported, and
[1:16] uh the market did not like either of the
[1:19] commentary in there. The earnings are
[1:20] still fine, but
[1:23] the overhang or the reality that I think
[1:25] uh probably started with just Silicon
[1:28] Valley recognizing that as the agentic
[1:30] world game, we have no terminal value on
[1:32] these companies,
[1:34] has set back in. So,
[1:37] I've posted about this a few times over
[1:39] the course of the last month. When we
[1:41] originally had the correlation break, it
[1:43] was a day where Oracle broke out.
[1:46] Software names were bouncing, but
[1:48] yesterday I or not yesterday, Wednesday
[1:51] I posted in X,
[1:52] and I said that if you look at what's
[1:55] performed so far this month within the
[1:57] IGV, which is the software ETF,
[2:01] almost all of the top 10 names
[2:04] were related to crypto.
[2:06] So, let's separate Bitcoin for a second.
[2:09] Uh what is happening is the miners were
[2:12] kind of the ones that led way before
[2:14] Bitcoin. And I mentioned them last week,
[2:17] uh I think here, but definitely on my
[2:18] weekly where
[2:20] I just said we've reached a point of
[2:22] compute shortage. And if you can
[2:24] scramble and get anything related to AI,
[2:27] that's where you want to go. We have
[2:28] bottlenecks throughout the world now. Um
[2:31] it was only memory 4 months ago, but now
[2:33] it's spread to CPUs. And because of
[2:36] Iran, it's spreading to chemicals and
[2:39] other places. You're going to have
[2:40] bottlenecks the rest of this year. So,
[2:42] software still has the issue of AI
[2:44] progress.
[2:46] But then the Bitcoin miners
[2:48] in recently have started to benefit from
[2:50] the reality that we don't have enough
[2:51] compute. So, the question is where does
[2:53] Bitcoin sit? Well, first of all, with
[2:54] the miners getting a bid, well, that's
[2:56] obviously good for Bitcoin and for no
[2:57] other reason with the ecosystem.
[2:59] But I think an important dynamic is
[3:01] shifting. And you and I have differed on
[3:03] inflation. There is really no doubt in
[3:06] my mind that inflation is going higher.
[3:08] Now,
[3:09] I want to separate this again because I
[3:11] think we talked about this last week.
[3:13] And so, people hear this.
[3:15] I do not think the housing market's
[3:17] going to get better. I do not think
[3:18] wages are going to get better. I do not
[3:20] think the things that traditionally are
[3:22] correlated with inflation are going to
[3:24] get better.
[3:26] But there's no way to refute what came
[3:28] out in the PMI numbers this Thursday.
[3:30] Service and manufacturing PMIs are
[3:33] moving higher.
[3:35] And most of them are at the highest
[3:36] levels since 2022.
[3:38] So, you are getting back into a world
[3:40] where inflation has peaked on the survey
[3:42] side for all of these things, which has
[3:44] a very high correlation as a leading
[3:46] indicator. And this is the thing I want
[3:48] to say to people. I don't care what the
[3:51] trueflation number is today. I care
[3:53] about what it's going to be like in 3
[3:54] months.
[3:56] And maybe their number will stay low.
[3:57] Headline CPI is the one that I have
[4:00] confidence in will be above 4% as we
[4:02] keep getting this data. And the reason
[4:04] is
[4:06] because the manufacturing bottlenecks
[4:08] are real and they are big. And I heard
[4:10] this today in listening to a podcast
[4:12] with Craig Fuller.
[4:14] And everyone should listen to this
[4:16] podcast because I like hearing people
[4:18] that talk about the economy and what's
[4:20] happening. This is a very unique
[4:22] economy. If I wanted to figure out where
[4:24] inflation was over the last 17 years, we
[4:27] were in a bear market in commodities for
[4:30] 17 years except for COVID. So, the
[4:33] reason I kind of took a shot at people
[4:35] fitting things to historical data is
[4:37] today's world is not the same as it was
[4:39] from 2010 to 2015. We are in a commodity
[4:43] bull market. We have shortages across
[4:45] the globe. And what Craig Fuller talked
[4:47] about was freight is out of control all
[4:50] because of AI. All because of the one
[4:52] big beautiful bill. All because of lower
[4:54] energy costs here and higher energy
[4:56] costs around the globe. So, natural gas
[4:58] is low. That'll keep help inflation a
[5:00] little bit down. Oil is higher. Gas is
[5:02] higher. Diesel's higher. That'll push it
[5:03] up. But the reality is memory prices,
[5:05] CPU, semiconductors, that stuff there's
[5:08] no there's no end in sight. And the
[5:09] reason I bring this up for Bitcoin for
[5:11] people, at some point Bitcoin is either
[5:14] in the abundant bucket or it's in the PI
[5:16] PMI sensitive
[5:19] bucket.
[5:20] It's a PMI sensitive. Its returns are
[5:22] there. The only thing missing, which I
[5:23] still think will happen, is
[5:25] year-over-year CPI will get above 4%.
[5:27] 3-month bills will stay below 4% and
[5:29] we'll have negative real yields. And I
[5:31] think that's what's happening is
[5:33] software's in a deflationary problem.
[5:35] So, your world
[5:37] Bitcoin's now moving into the
[5:39] inflationary world, my world, which is
[5:41] more related to commodities and
[5:42] scarcity. One thing that's interesting
[5:44] is you and I have talked in the past
[5:46] about Bitcoin being valuable both in an
[5:48] inflationary world and a deflationary
[5:50] world. And in the inflationary world, I
[5:52] think that's the one people are used to
[5:54] and saying, "Okay,
[5:55] uh if there is global liquidity
[5:57] increasing, this thing's very sensitive
[5:58] to that, and so it should be able to
[6:00] kind of sniff that out and go and and uh
[6:02] increase in uh price." Deflation, we've
[6:05] talked about this like abundance and
[6:07] scarcity and kind of value proposition
[6:09] from that standpoint. Are those the same
[6:11] buyers of Bitcoin, or is there like uh
[6:15] if you take a you know, 100% of people
[6:17] who are uh in the investment community,
[6:19] some of them are allocating capital in
[6:21] an inflationary environment, but they'll
[6:22] actually go to cash if we move to a
[6:24] different regime, or is it no, doesn't
[6:27] matter, you know, kind of who you are,
[6:28] you're going to end up at Bitcoin
[6:30] regardless of the regime, right? I
[6:31] always think about this idea of like an
[6:33] asset being different things to
[6:34] different people. And so, how do you
[6:36] think about Bitcoin's buyer in an
[6:38] inflationary versus deflationary world?
[6:40] You know, this is a great question. Um
[6:43] and I and I want to kind of use the Jeff
[6:45] Booth
[6:47] for this, cuz I think he does the best
[6:48] job of saying that there is always
[6:52] deflationary pressures in a credit-based
[6:55] fiat fiat system.
[6:57] Innovation is deflationary. We both
[6:59] agree on that.
[7:00] >> Yes.
[7:01] Ooh.
[7:03] If people are not getting jobs as
[7:05] easily, if they can't move up the
[7:07] corporate ladder, if there's
[7:08] affordability issues,
[7:10] well, that's the deflationary pressures
[7:12] that usually lead to more government
[7:14] stimulus. Well, we have government
[7:16] stimulus happening. So, when I say
[7:18] negative real rates, that may be some
[7:20] kind of economic wonk to people, and
[7:21] they're not really thinking about it.
[7:23] That is these two forces at the exact
[7:25] same time. To have negative real rates,
[7:28] you have deflationary pressures, which
[7:30] are keeping short-term rates low, and
[7:33] you have infla- inflationary pressures,
[7:35] which are keeping CPI high. Now, in the
[7:38] US, we don't have this, but in Brazil,
[7:40] they have what's called a basic basket.
[7:43] The basic basket that people need to
[7:45] survive on. Well, I think the basic
[7:47] basket for
[7:49] young educated people in the US, where
[7:51] is my housing? Can I get an apartment in
[7:53] a city I want to live in?
[7:55] Nope, not good. Um where are food
[7:58] prices? Okay, well, we have commodity
[7:59] inflation right now. Fertilizer prices
[8:01] are going up. Food inflation's not
[8:03] coming down. That's going to be an
[8:04] issue.
[8:05] Can you get a job anywhere you want?
[8:06] Yeah, you can get a job, not anywhere
[8:08] you want. You might have to take a job
[8:10] not in the education based on the
[8:12] education you have because of disruption
[8:14] happening out of nowhere in the labor
[8:15] force. And so, you might not be able to
[8:17] do this job. You might have to go do a
[8:19] different job even though you have an
[8:20] education in this job. I think we've
[8:22] taken the basic things of, let's say, an
[8:24] advanced society and made it very
[8:26] challenging. And that to me is the
[8:27] deflationary pressures on the one side
[8:29] that Bitcoin
[8:30] usually brings the anger out of people.
[8:32] This is when people would want Bitcoin.
[8:35] On the other side, when the government
[8:36] needs to have excess spending, and
[8:38] regardless of what's happening on the
[8:40] fiscal side, we have a fiscal impulse
[8:42] right now.
[8:43] Tax refunds, very high. Tax receipts,
[8:47] very low.
[8:48] Bonus depreciation for companies, which
[8:51] is another form of stimulus, is having a
[8:53] huge impact. Craig Fuller talks about
[8:55] this. I think all of these kind of half
[8:57] deflation, half inflationary pressures
[9:00] are creating an issue. Here's the one
[9:02] thing I will say about innovation. When
[9:04] you want to invest in innovation,
[9:06] in a deflationary world, you're trying
[9:08] to find
[9:10] what Jeff Bezos said during the last 15
[9:12] years, which is when Bitcoin was
[9:14] created. And what did Jeff Bezos say?
[9:16] Your margin is my opportunity.
[9:20] There's no more margin now in code-based
[9:21] stuff. It is in a
[9:24] freefall. That is why software stocks
[9:26] are getting up. So,
[9:27] where Bitcoin benefits is it doesn't get
[9:29] hurt by that. So, it's got the one side
[9:32] now, the inflationary side, which it
[9:34] will benefit from.
[9:36] And on the other side, where it benefits
[9:37] from deflation is from
[9:40] the opportunity cost that an individual
[9:42] investor has. I need growth assets and
[9:45] this is the Rick Edelman argument. How
[9:47] do I invest for the next 40 years cuz
[9:49] I'm going to live longer. What growth
[9:51] assets can I have cuz those are the ones
[9:53] that I can be in.
[9:54] Growth assets are not working.
[9:56] Hyperscalers and all of the things
[9:59] related to software, I think this is
[10:01] where Bitcoin separates itself in the
[10:03] second half of the year because of that
[10:04] dynamic playing out both inflation and
[10:06] deflation at the same time.
[10:08] What is the thought process in terms of
[10:10] Bitcoin price action? Like when you say
[10:12] breaking out, uh that could be Bitcoin
[10:13] doesn't move and stocks fall, but it
[10:16] also could mean Bitcoin goes up a lot
[10:17] and stocks stay constant or go down.
[10:20] Like how do you think of the
[10:21] relationship of these two and and how
[10:23] big maybe is the separation?
[10:25] Uh no matter what my views have been
[10:28] this year, one thing has stayed
[10:29] constant. We will have earnings growth
[10:30] and we will have nominal GDP.
[10:32] End of story.
[10:33] Uh
[10:35] we might not have real GDP to the level
[10:37] people are used to, but if nominal GDP
[10:40] seven and inflation is five, then you
[10:41] get 2% real GDP and everyone goes, "Oh,
[10:44] it's not
[10:45] Nominal GDP drives revenues. Nominal GDP
[10:47] drives earnings. Plain and simple. Um
[10:50] if companies are able to control their
[10:52] expenses via
[10:54] AI
[10:56] through labor, then that's where profit
[10:57] margins are sitting up at high level.
[10:59] So, I think what'll end up happening is
[11:02] you want to be long the stuff tied to
[11:04] the bottlenecks in the physical hardware
[11:06] world going forward. So, this is where
[11:08] I've talked about semiconductors and
[11:10] commodities and servers and anything
[11:13] physical. The software stuff I think
[11:15] will be in trouble and will remain in
[11:16] trouble. It has done such a big fall
[11:18] already that I'm not sure how much lower
[11:21] it can go overall, but I think it'll
[11:24] underperform. But when you look at
[11:26] stocks relative to nominal GDP, that's
[11:28] where I've said, if I had to guess, I
[11:30] think 10 years from now.
[11:32] Public companies
[11:35] are not going to be the ones benefiting
[11:36] from AI. It's going to be private
[11:38] startup companies that actually never go
[11:40] public. There's going to be a change in
[11:41] the capital structure. The financial
[11:43] guardrails of the world are shifting and
[11:44] they're shifting towards crypto. That
[11:46] combination of AI disrupting public
[11:48] companies cuz public companies just
[11:50] can't adopt AI because having a lot of
[11:53] people is very difficult and I think
[11:56] they're all learning that and I got to
[11:57] be honest, one of the stories that'll
[11:58] happen before the end of the year.
[12:00] I think a lot of these big companies are
[12:02] not going to see the benefit come
[12:03] through the way they think cuz it is
[12:05] just so hard and I'm sure you've got
[12:07] your own experiences. How do we
[12:08] integrate this in a way? The costs are
[12:10] going higher, the speed of it is going
[12:12] so fast and startup businesses don't
[12:14] have that friction of people inside the
[12:17] business, culture inside the business,
[12:19] legacy systems inside the business. It
[12:22] is a very difficult thing and I just got
[12:24] to tell you, it's like it's like a
[12:26] telling a 60-year-old, you know,
[12:28] athlete, "Hey, why don't you go play
[12:29] football again?"
[12:31] They might be in great shape, but
[12:32] they're not in great football shape. I
[12:33] don't think companies are in great AI
[12:35] shape.
[12:36] You have these five thematic buckets or
[12:40] baskets that you've been writing about
[12:42] to the 20 22 V platform into the
[12:44] subscribers. I want to go through each
[12:45] one of them and if you can kind of just
[12:47] help us understand because they I think
[12:49] the common theme is they are potential
[12:53] solutions or they are potential names
[12:55] where there is massive scarcity. And so
[12:57] you're really just taking the same idea
[12:58] and applying it to these different
[13:00] articles, it seems. Um maybe we can go
[13:02] backwards and start with chemical cuz I
[13:04] don't think we've talked a lot about
[13:05] chemical companies, the chemical theme.
[13:07] Explain this one a little bit.
[13:09] So, chemicals are very PMI sensitive,
[13:12] but they're also very semiconductor
[13:14] sensitive and also optical fiber. So, in
[13:18] there is optical fiber. So, one of the
[13:19] things all of these themes
[13:21] I've written about really from November
[13:23] of last year into the early part of this
[13:26] year.
[13:27] And this was a transition point. These
[13:29] are all agentic names. So, when you get
[13:32] back to pre-training and you think about
[13:34] getting AI up to this point,
[13:37] what was the major winner? It was
[13:38] Nvidia. Mhm. Um so, people can hear
[13:40] these amazing stats. The S&P 500 now,
[13:44] they've got different levels of you hear
[13:46] sectors, but then there's industry
[13:48] groups.
[13:49] Level two industry groups, the largest
[13:51] weighting in the S&P 500 is now
[13:53] semiconductors. It's now 17%.
[13:56] S&P's about $60 trillion, which means
[13:58] you're talking about $10 trillion. Of
[13:59] the $10
[14:00] Of the $10 trillion of semiconductors,
[14:03] seven and a half of the 10 are three
[14:05] companies, Nvidia,
[14:07] Broadcom, and Micron. Those were really
[14:09] the three major benef-
[14:11] They benefited the most
[14:13] in the early part. Now,
[14:15] Broadcom and Nvidia were the main
[14:17] beneficiaries from the data center
[14:19] buildout and the GPUs, and everyone
[14:20] heard that. But beginning in November
[14:22] when Opus 4.5 came out, we started to
[14:24] shift towards AI agents. Micron really
[14:27] started to benefit because of memory. I
[14:28] was talking about memory all year last
[14:30] year is we're not going to have enough
[14:31] because agents are coming.
[14:34] Once that agent world kicked,
[14:36] everything shifted to inference. So,
[14:39] Jensen Huang announced a partnership,
[14:42] almost a takeover of Grok, because
[14:45] his business is focused on the data
[14:47] center GPUs, but he realizes, well, now
[14:49] there's going to be a lot of agents, and
[14:50] this is the action world. This is not
[14:51] the thinking world. The action world
[14:53] takes a lot more east-west traffic. So,
[14:55] chemicals are necessary because of the
[14:58] Corning tubing. So, Corning has fiber.
[15:01] Okay, they need tubing. Well, chemical
[15:03] companies benefit from that. Bonding all
[15:06] the semiconductors together. So, when it
[15:07] was just GPUs, not as important, but now
[15:10] you're packaging the GPUs with memory
[15:12] and with all of these other components.
[15:14] That's why when I talk about Marvell,
[15:16] Marvell is another one. You saw Texas
[15:18] Instruments last night. What happened to
[15:19] Intel? Those That's GPU. One's power
[15:22] semi- All of the semiconductors are
[15:24] benefiting and that's because of the
[15:25] agentic side. So, in that first chemical
[15:28] basket, or not the first one, but
[15:29] chemicals, PMI sensitive,
[15:32] does really, really well with this. And
[15:34] right now, US chemical companies are
[15:36] benefiting cuz the chemicals here made
[15:37] with natural gas, and a lot of the ones
[15:39] overseas are with oil, and natural gas
[15:42] is very low right now in the US. So, US
[15:44] chemical company. So, it's the beginning
[15:46] of a bull market in chemicals. I wonder
[15:48] how much like a Coke Industries is
[15:50] benefiting from this, but it's private.
[15:52] And you know, they got a big chemical
[15:54] business, and some of these players that
[15:56] you would never think are AI-centric
[15:59] companies, but because they are the
[16:01] inputs,
[16:02] obviously, it's probably pretty good for
[16:03] them. Yeah. And the reason I want people
[16:05] to go listen to the Craig Fuller
[16:06] podcast, he specifically says in there
[16:09] that chemical shipments are through the
[16:11] roof.
[16:12] So,
[16:13] people have to understand that chemicals
[16:15] are used in almost everything, but this
[16:17] massive industrial build-out, and this
[16:19] is just the beginning cuz chemicals are
[16:21] going to be used in the auto upgrade.
[16:24] And just so people hear this cuz I've
[16:25] started to talk to more and more
[16:26] institutional investors, we may never
[16:29] sell more cars than we did this year.
[16:31] Meaning, we did 16 million. Okay, maybe
[16:33] next year we'll do 16. We've been around
[16:35] 16 million.
[16:36] We got close to 16 million in the '70s.
[16:39] So, then people go, "Well, we're never
[16:40] going to sell more, so why would this be
[16:41] important?" And it's like, "Okay, here's
[16:43] what's going to happen." And this is big
[16:45] for Qualcomm, leading-edge thing for
[16:47] people here. Qualcomm's not ready yet,
[16:49] but it's getting close. If we do 16
[16:51] million next year,
[16:53] a good portion of that 16 million, let's
[16:55] assume 1 million this year were AI
[16:58] components. Meaning, you could speak to
[17:00] the car, the car could do things. Hey,
[17:01] move the wiper, do this. So, Alexa
[17:03] within side the car.
[17:05] That takes a lot more chips. Takes a lot
[17:08] more chemicals for that. Well, next year
[17:10] maybe we do 4 million. So, instead of 0%
[17:13] growth rate of autos, we're actually
[17:15] doing 3-400%
[17:18] of AI autos.
[17:19] >> Smart smart cars. Exactly. And that's
[17:21] what's going to start happening every
[17:22] year. And so what people need to think
[17:23] about with semis is Jensen Huang saying
[17:26] it's going to be an 85 trillion-dollar
[17:28] recycling. Eventually what happens is if
[17:30] we have 100 million cars owned in this
[17:32] in this country, which is less than
[17:34] there is, but if we have 100 million
[17:36] by the time we get 5 years out, maybe 80
[17:38] of them are AI because no one's going to
[17:41] want the older models. So that process
[17:43] is a lot more dollars that goes into a
[17:45] semiconductor. This is endless, the
[17:47] upgrade cycle. And this is happening
[17:49] fast. So that's where chemicals fit in.
[17:51] Let's talk about whole rack as a theme.
[17:55] So this really gets into a combination
[17:57] of the GPUs to the rack of things that
[18:00] go into a data center now. And this
[18:03] includes servers, it includes memory, it
[18:06] includes CPUs, it includes all of that
[18:09] stuff. So if Morgan Stanley wants to
[18:12] have some of their AI come from the
[18:14] cloud, so one of the providers, but then
[18:17] they go, "You know what? We need to
[18:18] also, because of privacy issues, we need
[18:20] to house this stuff here." Well, they
[18:22] need to go back to the old world of
[18:23] servers. And so all of the Dell, the
[18:26] Hewlett-Packard, all of those names have
[18:28] just broken out. Those are all part of
[18:30] the whole rack. So it's that it's it's
[18:32] literally not just semis, it's
[18:33] everything that people would need for
[18:35] the whole rack to actually be able to do
[18:37] AI inside their own device. Today's
[18:39] episode is brought to you by Consensus
[18:41] Miami. I'm going to be speaking there on
[18:43] the main stage in a couple of weeks.
[18:45] This is the event the entire industry
[18:47] shows up for, Wall Street, the White
[18:49] House, all of crypto and Web3. This May
[18:51] 5th through May 7th, you can discover
[18:53] everything from crypto at scale to
[18:55] institutional integration and agentic
[18:57] commerce. It's the forces driving
[18:59] trillions on chain and reshaping global
[19:02] finance. 20,000 decision-makers are
[19:04] going to be in attendance. Three
[19:06] action-packed days, deals get signed
[19:08] here, funds get raised here, and the
[19:10] next cycle starts here as well. You
[19:12] can't afford to miss it. I'll be there
[19:13] speaking and you can use the code Pomp
[19:16] Liano to get 25% off your pass and join
[19:19] me there. Go in the description and look
[19:21] how to spell my name and use code Pomp
[19:23] Liano to get 25% off your pass to join
[19:25] me there.
[19:26] You know, it's funny to me is there's
[19:28] always these stories of like a huge
[19:30] inflection point Silicon Valley was when
[19:32] you no longer had to have like set up
[19:33] the server. You just could like you
[19:35] know, cloud get started whatever.
[19:37] The full circle of like now there's a
[19:38] bunch of businesses be like this cloud
[19:40] thing, we're out on that. I don't want
[19:42] you have access to my data for training
[19:44] purposes or whatever. Do you remember
[19:45] how long it took to get companies to go
[19:47] to the cloud? Oh, you mean you had a
[19:49] bunch of people who didn't trust it,
[19:51] didn't understand it, whatever. I will
[19:53] give you one interesting data point. I
[19:54] talked to one of the major model labs
[19:56] this past week and
[19:58] we were talking to them and and I just
[20:00] on offhand mentioned the data labeling
[20:03] startups and how you know, we've seen
[20:04] all these from the micro ones to the
[20:06] handshakes. They just been exploding in
[20:08] revenue etc.
[20:09] I started asking a little bit about the
[20:11] data labeling and it
[20:12] one of the interesting things they said
[20:13] is yeah, we actually are much more
[20:15] focused on post training data now than
[20:18] pre-training data. And so you know, I
[20:20] asked a couple more questions and and
[20:21] basically it sound although they didn't
[20:23] explicitly say this, it sounded like
[20:26] they already have done so much training
[20:28] that now we are entering a place where
[20:30] it's all about the fine-tuning and and
[20:32] the evals and
[20:34] it does to me at least it was like a
[20:36] reminder like we have come a far you
[20:38] know, very long way even though it feels
[20:39] like it's still the starting line.
[20:41] 5 years ago, I mean almost none of this
[20:43] was being discussed. Well, we we kind of
[20:46] talked about this before and so people
[20:47] here it for for use for us, this thing
[20:50] is already smart enough.
[20:52] What it's not smart enough, so the
[20:54] pre-training, the brain building, what
[20:55] it's not smart enough for is
[20:58] going through the physical world and
[21:00] making decisions in an automobile
[21:02] completely. So it's not there yet. So
[21:04] there needs to be a lot more training on
[21:06] interacting with the physical world. The
[21:09] memorization of all human knowledge it
[21:11] already has, but figuring out things to
[21:13] cure cancer, figure out things that we
[21:14] haven't done yet, it needs more work for
[21:17] that. So, I want people when they think
[21:19] about this to to really break it down to
[21:21] two separate things.
[21:22] There's now two tracks. There's still
[21:24] the pre-training that's necessary for
[21:26] all of the things that Elon Musk wants,
[21:29] that Demis Hassabis want for
[21:31] the world's greatest problems, energy,
[21:32] all of that. At the same time, we're now
[21:34] entering the agentic when people hear
[21:36] agentic, maybe their eyes gloss over and
[21:38] go, "I don't know what that means."
[21:39] That's the action side. So, you've got
[21:41] the thinking side and then you have the
[21:42] action side. For the action side, it's
[21:45] more east-west traffic and that's why
[21:46] with chemicals, with the whole rack,
[21:49] these are different components that are
[21:50] necessary. And for people listening,
[21:52] like the reason I put these indices
[21:54] together, I talked to the most
[21:56] sophisticated investors in the world.
[21:58] They are still behind this change. So,
[22:01] one of the reasons that we do the show
[22:02] together, one of the reasons I won't
[22:04] work for anyone again, is because I need
[22:07] the time to listen to
[22:09] Dillon Patel was on a podcast that I
[22:11] listened to today. While you were in
[22:12] your office and I was doing a thing, I
[22:14] was listening he was on a podcast today.
[22:16] >> You see him getting an edge on me? He
[22:18] literally he was camped out in the
[22:19] office next to me and he he knew I was
[22:21] on a call and I'm going to get an edge
[22:22] on him right now. I'm going to listen to
[22:23] the thing that he hasn't had time to
[22:25] listen to and I'm going to go bring it
[22:26] up on the podcast. You're in there
[22:28] having your meeting. I look at my watch
[22:30] and go, "I got 25 minutes. This is a
[22:31] 45-minute podcast." 2x
[22:33] >> Exactly, perfect. I already know. Look
[22:35] how fast you did that now. You know
[22:37] what's You know what's interesting
[22:38] though? Is I saw a tweet and I I wish I
[22:40] remembered who who tweeted so I give him
[22:42] credit for it, but they said actually
[22:44] a great way to think about modern life
[22:47] is that we are in a bandwidth crisis.
[22:50] And they were talking not about
[22:52] hardware, software, they're talking
[22:53] about human bandwidth. And I think AI is
[22:56] a perfect example of it is nearly
[22:59] impossible to keep up with everything
[23:01] that is happening. And so when you get
[23:04] the overwhelming I don't have enough
[23:06] bandwidth
[23:07] human brain goes into how do I
[23:08] prioritize? How do I 2x, 3x, you know,
[23:11] all these different things that we try
[23:12] to work around. Mhm.
[23:15] But that's just an AI.
[23:17] Then if you're somebody who also happens
[23:18] to be a sports fan or happens to be a
[23:20] pop culture fan or wants to read books
[23:22] or
[23:23] how do you keep up with everything? It's
[23:25] impossible.
[23:27] Uh this is the beauty of the decision I
[23:29] made in being able to sit there and
[23:31] listen to people. And I want to use an
[23:33] analogy of of how important these
[23:35] podcasts and and this news is. So
[23:38] if you go to work and you want to go
[23:40] find out the the final score of the Nick
[23:43] game and you go to a box score and you
[23:45] look at the final score, you look at all
[23:47] the quarters, when the game was, who
[23:50] scored the most points and you go
[23:52] through all this, you actually know the
[23:53] result.
[23:55] Well, to be honest with you, that's what
[23:56] economists do.
[23:58] So no offense to the economists
[23:59] listening, but I always wanted to build
[24:02] my own models that were based on leading
[24:03] indicators and lagging indicators
[24:05] because looking at the box score doesn't
[24:07] tell me anything about the next game.
[24:08] Did someone get hurt? Did someone leave
[24:11] limping but they're still going to play?
[24:12] Is their jump shot off?
[24:14] Let me go through and see if someone was
[24:17] seven for 10 but all you know, the seven
[24:19] makes were dunks and the three misses
[24:22] were from three points. Okay,
[24:24] person didn't have good shooting game,
[24:25] just sent him to go through it. You
[24:26] don't get the data. By listening to
[24:28] Dylan Patel and Craig Fuller. Dylan
[24:31] Patel is on semiconductors like no one
[24:34] else. So he knows all the box score but
[24:36] he was at the game. Craig Fuller is at
[24:39] the game. I'd rather talk to those
[24:41] people and listen to those than read the
[24:42] box score. But for people that have jobs
[24:45] managing people doing this, the box
[24:46] score is all they can do cuz they don't
[24:48] have the time. So they depend on the
[24:49] economists. In this day and age with
[24:52] podcast NX
[24:55] it takes a lot to go get it. But that
[24:57] information is so real time and I'll
[24:59] just give you a live example. I was
[25:01] doing a webinar this week for the
[25:02] subscribers and I had read a media
[25:06] report that was going viral through X
[25:10] around some
[25:12] issues, not issues, it's probably the
[25:15] wrong word. Nvidia really focused on
[25:18] Korean heavy power companies
[25:21] and it even extending into shipbuilding
[25:24] companies. They're kind of find ways to
[25:25] get more power. Mhm. So that fits my
[25:28] power shortage issue that's happening,
[25:29] but it also leads to power
[25:31] semiconductors. And these stocks have
[25:33] started to go through the roof. Now this
[25:35] was before Texas Instruments. So when
[25:36] people ask me yesterday, what are some
[25:38] names? And I'm like on semi, Texas
[25:40] Instruments. And why now? And I just
[25:42] said Jensen Huang is now ready to
[25:44] release this. He's looking for new
[25:47] partners on this and power semis are
[25:49] going to be really important. That is
[25:51] all from an X thing that was in the
[25:52] Korean media this week. It's crazy. It
[25:55] won't be news for for the Goldman Sachs
[25:58] and Morgan Stanley, honestly, until
[25:59] later. They can't do a report. They've
[26:01] got so many calls and meetings already
[26:03] set up. So I don't like to have
[26:05] meetings. I don't want to be there. So
[26:06] real-time news and real-time information
[26:08] from critical people.
[26:09] >> Can I admit something that pains me, but
[26:11] I got to be honest with you and with the
[26:12] audience?
[26:14] Do you know the number one activity that
[26:16] I find personal enjoyment in that has
[26:17] been sacrificed in my attempt to keep up
[26:19] with all this?
[26:21] Is reading physical books.
[26:23] I have found that over the last 6 months
[26:26] or so, I used to read about a book a
[26:27] week and I would read everything from
[26:30] things that were work related to things
[26:31] I just enjoyed and everything in
[26:32] between.
[26:34] I've noticed that I have been reading a
[26:36] lot less, but I'm probably consuming
[26:39] more information. Mhm. So it's like
[26:41] total consumption is going up, but most
[26:44] of the reading I used to do was things
[26:47] that were talking about the past. And so
[26:49] I would barbell it. I want to be as
[26:51] current as possible and I want to be as
[26:53] rooted in history as possible. I find
[26:56] myself consuming less of the what
[26:58] happened in history or the, you know,
[27:00] biographies of people, etc., because I'm
[27:02] so inundated with the like every single
[27:04] day there is something new. Mhm. And
[27:06] maybe it balances that back out at some
[27:08] point, but as I noticed that I started
[27:10] to like I don't know if that's a good
[27:12] thing or not, right? But I just
[27:13] naturally see myself going that way. So,
[27:15] let me let me give you some um
[27:18] before you and I did a podcast, I
[27:19] actually did a podcast. It's called In
[27:21] Search of Dream Marbles, okay? I
[27:23] remember. You remember that. On that
[27:25] podcast, one of the episodes was why
[27:28] books are a waste of time. Oh, god.
[27:30] >> Now, this was pre-chat GPT.
[27:33] >> things that we disagree on. Okay.
[27:34] >> So, here we go. No, no, no.
[27:37] Remember, I'm I'm a person of nuance and
[27:40] kind of, you know, I got a little George
[27:41] Carlin in me, like, read the words.
[27:44] You already admitted it's a waste of
[27:46] time. You don't have the time for the
[27:47] books. So, that means it takes a lot of
[27:49] time, so it's a waste of time. Now, if
[27:51] you ever want to play a game, pick a
[27:53] book that you haven't read yet on
[27:54] history, okay?
[27:56] And look at what it is. Take the
[28:00] the details of the book, put it into
[28:03] chat GPT, and say, "Hey, give me a
[28:06] three-paragraph write-off of the what
[28:08] this book is about, okay?
[28:10] Now,
[28:12] hit a new chat.
[28:14] Ask it the same prompt again. You know
[28:16] what you're going to get? Mhm. Two
[28:17] different versions of the same thing.
[28:20] There are There are so many problems
[28:21] with a book. It's one person's opinion.
[28:24] Mhm. If you get stuck in a world of
[28:26] trying to figure out what something is,
[28:28] when I do my AI training videos uh that
[28:32] some of you guys helped me with here,
[28:34] part of my process is to take one thing,
[28:38] which think of as like a book, and let
[28:40] five separate LLMs do a deep research
[28:42] report on them. It's like five different
[28:44] people giving you another book version,
[28:46] and then I take those five and I
[28:48] consolidate them. It does a bunch of
[28:49] things. It gets rid of hallucinations,
[28:51] but then it gives you kind of a more
[28:53] in-depth view that takes the key
[28:56] components. So, if you ask five
[28:57] eyewitnesses about a crime that happened
[28:59] outside,
[29:01] and five of them will give very
[29:02] different accounts. The Venn diagram of
[29:05] where they all agree is usually the most
[29:07] useful information. That's what the
[29:08] green marbles means. That's what I do
[29:10] with AI. So, that's why books were a
[29:12] waste of time.
[29:12] >> So,
[29:13] now you're really going to pull my
[29:16] me admitting things.
[29:17] Uh in my understanding and
[29:20] exploration of why I was not reading
[29:21] more books, I actually did something
[29:23] similar where
[29:24] um when I was in high school, my best
[29:26] friend was SparkNotes. I don't know if
[29:28] you had this when you were in school.
[29:30] So, SparkNotes was basically
[29:31] >> was in school a long time ago, my
[29:33] friend.
[29:33] >> Yeah, well, so when I was in school, you
[29:35] could read the book they told you to
[29:36] read, or if you were like me and you
[29:38] spent more time trying to not read the
[29:39] book, you would go find a SparkNotes.
[29:41] And then We were candlelit in my school.
[29:42] >> oh jeez, okay.
[29:44] Um and so, uh I asked, "Hey, can you
[29:47] create the SparkNotes for this? I want
[29:49] to be able to read in 15 minutes or
[29:50] less, right?" And it would go and it'd
[29:51] give me a thing. And um
[29:54] I realized though one of the things I
[29:55] love about some of these books is the
[29:57] anecdotes. And so, I've been trying to
[29:59] figure out, can I basically have the
[30:00] model not give me the summary, but give
[30:03] me the anecdotes? Go and find five
[30:05] stories in this thing that you know,
[30:07] would be surprising or whatever. It's
[30:09] not perfect, and there's some things
[30:11] about like don't let it do web search,
[30:13] instead have it, you know, actually get
[30:14] the uh like the Google book text etc.
[30:18] In 6 months,
[30:20] you're going to be able to do all this
[30:21] stuff in a pretty interesting way. And
[30:24] maybe the part I'm most interested in is
[30:26] and then what happens when you tell it,
[30:28] "Okay, now create a video that's 15
[30:30] minutes long that explains to me all the
[30:31] major concepts of this." And it's like
[30:33] having a you know, somebody who read the
[30:34] book and then tells you all the answers.
[30:36] And so, to me like as we head this way,
[30:39] like maybe Maybe go sell all my physical
[30:41] books. Like I'm
[30:43] I'm intellectually short
[30:46] and shouldn't be. All right. So, let let
[30:47] me since all the writers hate me now,
[30:49] let me let me let me let me give um So,
[30:52] I love I love the Steve Jobs book, okay?
[30:55] I love Walter Isaacson's book.
[30:57] >> Well, yes, Isaacson's book. And and Josh
[30:59] Waitzkin wrote The Art of Learning,
[31:00] which I love.
[31:01] >> book.
[31:01] >> So, if I use those two, the common
[31:03] thread between two of them, which AI
[31:05] cannot do,
[31:07] they're experiential. Yes. So, the Jobs
[31:10] book is both the history of Jobs, but
[31:13] it's also his experience with Steve
[31:15] Jobs.
[31:15] >> Mhm.
[31:16] I like personal experiences because I
[31:18] learn from those. So, I don't want
[31:20] people to think that like I'm not a
[31:21] fiction writer because I am I'm an
[31:23] insatiable learner,
[31:25] and there's just something that fiction
[31:27] doesn't connect with me. I think it's
[31:28] ADHD, and I'm constantly looking to
[31:30] learn.
[31:31] It just is. I I don't rest that much. I
[31:33] don't need a lot of sleep. I get my
[31:35] sleep cuz I have major RV and I want to
[31:37] stay young, but I do love the concept of
[31:40] sitting with certain books that the
[31:42] experiential side grabs me. And that's
[31:45] why when you read my Substack or you
[31:47] read my works, there's always something
[31:49] personal in there because my own
[31:51] experience of things in life becomes the
[31:54] trigger point or the dot connection to
[31:56] explore in a book and hear other people.
[31:58] So, you're you're using experiential,
[32:00] I'm using anecdotes, same thing. Like
[32:02] that's a
[32:03] I actually I'll give you a good example.
[32:05] Um
[32:06] I forget
[32:07] uh
[32:08] there's a podcast called Rainmakers uh
[32:11] my friend Ram uh put together. And in
[32:13] it, he talks about Philip Anschutz, who
[32:16] uh AEG, like they own a lot of like the
[32:18] uh sports arenas and stuff like this.
[32:22] Um
[32:22] At a time There was a time where he was
[32:24] buying up all of the like Regal Cinemas,
[32:27] etc. And the movie business was the the
[32:30] first time the movie business was
[32:31] tanking, and everyone said it was over.
[32:33] And everyone was like, "What is this guy
[32:34] doing? Why is he buying up all of the
[32:36] movie theaters? And in that
[32:39] telling of it, he realized that nobody
[32:42] gets to the movie theater late.
[32:44] Everyone shows up at least 10 minutes
[32:45] early. So, he was the first guy who
[32:47] realized have a captive audience of, you
[32:49] know, a couple hundred people. So, he
[32:50] started playing ads before the movie
[32:52] started and he completely changed the
[32:54] economics of the movie theater and, you
[32:56] know, was able to revive this thing.
[32:58] So, I was like, wow, if you have a
[33:00] captive audience, you can play an ad.
[33:01] And so, literally, if you come up to our
[33:03] conferences, you will see everyone
[33:04] tomorrow, we will play ads during the
[33:06] conference and I got it from that book.
[33:07] And so, like, that's the stuff that I
[33:08] love is the anecdote that you can pull
[33:10] out and then say, "Hey, how do I apply
[33:12] this?"
[33:12] AI is not yet there to to be able to
[33:14] grab that. No, and I I I I do not
[33:16] believe that AI
[33:18] it's going to take a long time for it to
[33:20] understand the the nuances of the of
[33:22] what drives people's decisions, the
[33:24] psychology of people. I just don't think
[33:26] it's going to be there quickly. Let's
[33:27] talk about mythos, mythos, whichever we
[33:29] want to pronounce it. Mythos, mythos?
[33:31] >> I've made the decision. I'm I'm drawing
[33:33] a line. It's mythos.
[33:35] Take that.
[33:36] Mythos, you think that this story should
[33:39] scare a lot of people. Why?
[33:42] I
[33:43] So, the story is that apparently
[33:45] unauthorized users had access to it.
[33:48] Uh again,
[33:49] >> What does that even mean?
[33:50] >> Yeah, that probably means they think the
[33:52] Chinese have them cuz there was a report
[33:53] today that whatever the case is because
[33:56] the Chinese have apparently been
[33:59] going through the distillation process
[34:01] with all the whatever the case is.
[34:03] I think it says two things. One is it is
[34:05] ridiculous for people to not think that
[34:08] people can hack into anything already.
[34:11] It's just
[34:13] it's going to happen. Um number two, it
[34:16] does mean that people should expect
[34:18] that one of the biggest dangers from AI
[34:21] is AI itself. And I do think there's
[34:23] going to be a hacking situation this
[34:25] year. I do think cryptography is going
[34:27] to become a place that is more
[34:29] respected,
[34:30] uh which will help Bitcoin as well, but
[34:31] I do think that
[34:32] >> Bitcoiners have been all over that one
[34:33] for a half a decade and a half. I And
[34:36] again, um,
[34:38] you're going to have issues for sure
[34:40] with artificial intelligence. When you
[34:41] get to this much power in it, um,
[34:45] it's just we're accelerating so fast
[34:47] that we're not ready for it. So, the
[34:48] same thing I said about enterprises,
[34:51] they don't know how to use it. Uh, there
[34:53] was a great uh, post by Aaron Levie
[34:57] at from Box. Mhm.
[34:59] He's excellent on the AI stuff. He's
[35:01] great and he his ex-posts are very
[35:03] thought out. And again, you're you're
[35:05] talking about someone who's at the game.
[35:07] Um, if you strip out I mean, software
[35:09] cares about software, but
[35:11] his whole thing about if
[35:13] the arc How how do you set up an
[35:15] architecture of AI for your company if
[35:17] the architecture you're using that you
[35:19] decide on today is obsolete in 3 months
[35:22] from now? So, think about where Open AI
[35:24] was 6 months ago. I think Gemini was
[35:26] ahead of everyone 6 months ago. And 6
[35:28] months before that it was Open AI and
[35:30] now it's Anthropic.
[35:31] Well, if Anthropic doesn't have enough
[35:33] compute and all of a sudden Open AI
[35:35] rolls out some of that excess spending
[35:36] they did in their Blackwells allow them
[35:38] to get to the agentic side,
[35:40] is moving to Anthropic the best idea? I
[35:43] think it's very difficult for people to
[35:46] figure this out. And I think
[35:47] governments,
[35:49] everyone is going to have a hard time
[35:51] with it. So, with Mythos,
[35:53] the fact that this model is such a step
[35:55] change over the last one, how do you
[35:57] deal with this? What do you deal with
[35:59] it? And that's why they've been giving
[36:00] it to banks, not just in the US, but now
[36:02] they're giving it to banks in Europe for
[36:04] the same reason. I just think it's going
[36:05] to be an issue that people should be
[36:06] ready for. It won't be the end of the
[36:08] world cuz we'll figure out the fences
[36:10] for it, but I think to get to a solution
[36:13] to the problem, we're going to have a
[36:13] lot of problems first.
[36:16] When you look at these new models and
[36:18] how fast they're coming up, um,
[36:20] there's two things I've been told over
[36:22] the last 2 weeks that I think are
[36:23] interesting. The first is a an
[36:24] investment banker who covers many of the
[36:26] large model companies explained to me
[36:28] that there is a very large divide. What
[36:30] we generalized and and thought we knew
[36:32] about Open AI is going after consumer,
[36:34] Anthropic is going after enterprise.
[36:37] He hammered home for me like, "Oh,
[36:39] whatever you thought was the focus it's
[36:41] like 10 times more of a focus." Like
[36:43] Anthropic, if you are not part of the
[36:44] enterprise, they do not care. And so
[36:45] this whole idea of like giving it to the
[36:46] banks, giving it to these people,
[36:49] they also in a weird way by giving them
[36:51] access to a model that is, you know, the
[36:53] greatest best model, whatever, it's kind
[36:55] of marketing for also you should be a
[36:57] customer of ours. And so if we can build
[36:59] this thing that you're scared of, then
[37:00] maybe you'll come to us for the
[37:01] solution. And so I do find that, you
[37:03] know, somewhat interesting is like the
[37:04] marketing component.
[37:06] The second thing is um
[37:09] do you know who is buying lots of data
[37:11] right now, which is pretty interesting,
[37:13] is Meta, from what I understand. So they
[37:16] are out trying to get tons of different
[37:18] data sets that they almost it it least
[37:21] what I'm being told, they feel like
[37:23] they're a little behind. And so how do
[37:24] they accelerate? Well, let's go get a
[37:25] bunch of this data, start training it,
[37:27] etc. Uh my understanding is somebody who
[37:29] is not buying a lot of data
[37:31] is XAI. They've been hiring kind of
[37:34] contractors to go and do data labeling,
[37:36] etc.
[37:37] But now SpaceX and Cursor.
[37:40] Yep. And so it feels like everyone had a
[37:43] different strategy.
[37:44] >> Yep. And now it's not just a land grab
[37:46] on power and infrastructure and all
[37:47] this, it is now how do we get the data
[37:49] sets we need? And then also the creative
[37:52] deal making going on in Silicon Valley,
[37:54] whether it's how do we, you know, buy
[37:55] Scale AI but not buy it, to this like
[37:58] Cursor deal where you basically get, you
[37:59] know, hey, we'll pay you $10 billion in
[38:01] the worst case, in the best case we buy
[38:02] your company for $60 billion.
[38:04] I mean, this is like the dream, right?
[38:07] Yeah, and just so people understand this
[38:08] cuz what you're saying is true, I think
[38:11] people need to simplify this into why
[38:13] these companies are doing this. Number
[38:15] one, they need more compute. Mhm.
[38:17] 100%.
[38:19] Well, they also need capital for the
[38:21] compute. And that's why the raises and
[38:24] SpaceX is going to the market because to
[38:26] build Terafab and to do and I think they
[38:28] need
[38:29] >> Does he need some money? Yeah, I mean
[38:31] they all need money and they're using
[38:32] the capital markets. So they're using
[38:34] bond issuance. Like they're going out
[38:36] there to raise a lot. OpenAI just raised
[38:38] $122 billion.
[38:40] For startup private company? Yeah, I
[38:41] mean it's bigger than the bottom 300
[38:43] companies in market cap of the S&P 500.
[38:45] That's how much money they raised for.
[38:46] So you're dealing with just an insane
[38:49] amount of money that's necessary.
[38:51] And they need revenues. So they are in a
[38:54] what do we do? Where do we get the
[38:56] revenue? So Anthropic is generating
[38:58] revenue. So the competition migrates to
[39:01] enterprise. We all need enterprise. So
[39:04] it's it is capital needed for the
[39:06] compute which is then necessary for
[39:07] this. Now Elon has a different story.
[39:11] Where's his revenue coming in from
[39:12] SpaceX? Well, it's not coming in from
[39:14] flying rocket ships in the space. It's
[39:15] Starlink. So
[39:17] Starlink is enough of a profitability at
[39:20] this point for that company to be
[39:21] judged. It's going to have a very high
[39:22] multiple at where it's being done. But
[39:26] Tesla does too. And you go through this,
[39:27] it's like okay, how is he going to raise
[39:30] the capital except from
[39:32] individuals and company we're reaching
[39:34] that point where when you get that many
[39:37] companies that need this much capital to
[39:40] generate the revenue and they believe
[39:41] the revenue is going to come from the
[39:43] enterprise. And I keep saying
[39:46] it's really weird how this goes, but
[39:49] annualized revenue rate
[39:52] is not real money guys yet. Meaning
[39:55] you're annualizing it, which means
[39:57] you're taking it on okay, we had this
[39:59] much this times 12. Okay, great. Today's
[40:02] episode is brought to you by Uphold. Are
[40:04] you someone who's tired of juggling
[40:05] multiple apps just to trade, earn,
[40:07] stake, or spend your crypto? Then listen
[40:09] up because Uphold does all of that in
[40:11] one single unified platform. You can
[40:13] access 300 plus crypto and fiat
[40:15] currencies with an interface that works
[40:17] for you, whether you're a beginner or an
[40:19] expert. Uphold also features any-to-any
[40:22] swaps, where you can swap crypto to
[40:24] fiat, fiat to metals, and tokens to
[40:27] tokens directly in the app. If you
[40:28] prefer self-custody, Uphold's vault
[40:30] gives you multi-sig security, key
[40:32] recovery, direct trading access, and
[40:35] peace of mind without giving up
[40:36] convenience. Uphold is also 100%
[40:38] reserved, meaning no fractional
[40:40] practices, and proof of reserves are
[40:42] updated every 30 seconds, so you can
[40:44] verify your assets anytime. For US
[40:46] users, you can even earn yield on
[40:48] dollars with a USD interest account. No
[40:50] fees, no minimums, and funds are insured
[40:53] up to $2.5 million
[40:54] through the Atomic Cash Sweep program.
[40:57] If you want one app for your entire
[40:59] digital asset life, check out Uphold
[41:01] today. Go to uphold.com to learn more.
[41:05] uphold.com,
[41:06] go check them out today.
[41:09] Today's episode is brought to you by
[41:10] Figure. We're living in a world of
[41:12] higher for longer rates, tighter credit,
[41:14] and fewer places to earn real yield.
[41:16] That's why Figure's interesting.
[41:17] Figure's the largest non-bank mortgage
[41:19] lender in the United States with over
[41:21] $19 billion unlocked on their lending
[41:23] platform. They're now applying the same
[41:25] institutional infrastructure to
[41:26] crypto-backed loans and real-world asset
[41:28] yield. If you're holding crypto, Figure
[41:30] lets you borrow against it at 8.91%
[41:33] interest with 50% loan-to-value. It
[41:36] gives you liquidity without selling,
[41:37] without triggering taxes, and without
[41:39] stepping out of the market. On the risk
[41:41] side, they use decentralized MPC
[41:43] custody, so assets stay in segregated
[41:45] wallets, and they've introduced
[41:47] liquidation protection to help manage
[41:49] downside volatility. And for investors
[41:51] looking for yield, Democritized Prime
[41:53] offers up to 8.5% APY, paid hourly and
[41:56] backed by real-world private credit,
[41:59] real cash flow, not speculative token
[42:01] emission. In a market where capital
[42:02] efficiency matters more than ever,
[42:04] Figure gives you real options. Check
[42:06] them out using the link in the
[42:07] description, or go to
[42:09] figuremarkets.co/pomp.
[42:12] Again, that's figuremarkets.co/pomp.
[42:16] Well, have you seen the um
[42:19] controversy with C A R R, contracted
[42:24] annual run rate?
[42:25] >> No. Okay. So, what the point you're
[42:27] bringing is um
[42:29] companies and and this is not
[42:30] necessarily a nefarious thing. They're
[42:31] clear about it, but they will say,
[42:33] "Okay, I did $100 of revenue in the
[42:35] month of March. If I times that by 12,
[42:37] that means I am doing $1,200 of annual
[42:40] run rate." Now, if you're in a
[42:41] subscription business and you can prove
[42:43] churn and stuff, that's kind of where it
[42:44] emanated from. But, now people are just
[42:46] taking like the monthly revenue and it
[42:49] could be really spiky or lumpy and they
[42:50] just annualize it. Okay. So, I think
[42:53] that's like investors are smart. They
[42:54] they understand some of the perils of
[42:56] that. But, now there's this new
[42:58] contracted annual run rate. And my
[43:01] understanding of how these deals work is
[43:03] let's say that I am a model company or
[43:06] I'm an AI company and I come to you as
[43:07] the customer and I say, "You know what?
[43:09] You want this thing that I have.
[43:11] It's really important to me that we sign
[43:12] a multi-year deal. So, let's sign a
[43:13] three-year deal.
[43:15] The contract in the third year is going
[43:17] to be $3 million.
[43:20] But, let's make sure you like it. So, in
[43:23] the first year, I'll give it to you for
[43:25] 100k.
[43:27] And then in the second year, it'll be a
[43:29] million and in the uh third year it'll
[43:31] be 3 million. Mhm.
[43:33] You also can cancel after 12 months.
[43:37] And so, you're
[43:38] contractually obligated to give me 100k
[43:40] in the first year.
[43:41] >> Yep. And you can leave after 12 months.
[43:44] But, my contracted annual run rate
[43:46] >> Yep. is 3 million because in the third
[43:49] year you're going to pay me 3 million.
[43:50] And I we're signing a $3 million a year
[43:52] contract is the way it's positioned. And
[43:54] so, I go and I say, "Well, I was doing
[43:55] 10 million. Now, I'm doing 13 million in
[43:56] contracted annual run rate."
[43:59] Unsophisticated investors Mhm. may not
[44:02] realize that really only got 100k or
[44:04] they could get out after a year. And so,
[44:07] I think people have to be very cautious
[44:09] when they see some of these headline
[44:10] numbers and it is pervasive now, not
[44:12] just in AI companies. I mean, this is
[44:13] becoming a how do you market, you know,
[44:15] how good your company is doing, but
[44:16] there's a lot of this type of stuff
[44:18] going on and I think that's where people
[44:19] are getting caught up with like, how did
[44:20] a company go from, you know, zero to
[44:22] some crazy number? It's like, was it
[44:24] like real revenue or is it, you know,
[44:26] fugazi, you know, type math? And the key
[44:29] thing out of what you're saying what
[44:30] people need to understand and this is
[44:32] why the ROIC is going to remain a risk
[44:34] no matter what. So, the reason I was
[44:36] taking a shot at the annualized run rate
[44:37] is just a fact, which is if they run out
[44:39] of compute, guess what guys? It's not
[44:41] going to go higher. Um, you've run into
[44:44] a point where exponential change has an
[44:46] impact on the on value on the way people
[44:47] think of things. I was just on a call
[44:49] and someone said, so at what level will
[44:51] you care about SaaS companies? And I
[44:53] literally said,
[44:56] in fact, I'll use something here. Books
[44:58] are waste of time. Spending time on
[45:00] software to try and figure out where
[45:02] valuation is is a waste of time.
[45:05] This stuff is moving so fast that when
[45:06] there's no terminal value, how do you
[45:09] value something? So, you're asking me
[45:11] where I see value in the future when I
[45:13] don't know if they're going to exist in
[45:14] the future. So, do the reverse with the
[45:17] contracted annualized revenue. You're
[45:19] doing the same thing, it's just a
[45:20] different version of it, which is, okay,
[45:22] I'm only going to pay you 100,000 now
[45:24] because I don't know if your technology
[45:26] is going to work and I don't know if I'm
[45:27] going to be able to change my business
[45:28] to do it and I don't know if this will
[45:30] be the best technology in a year. That's
[45:31] all because of exponential change. It's
[45:33] different than software.
[45:35] How long have we been doing this podcast
[45:37] show every week? A year now? A little
[45:39] more than a year? A little more than a
[45:40] year.
[45:41] I don't have the smartest brain, which
[45:43] is why it took me this long to come to
[45:45] this conclusion, but I just realized
[45:47] what your entire world view. You're a
[45:49] scarcity trader.
[45:51] Like that's pretty much what you were
[45:53] looking for. You were looking for
[45:54] shortages and scarcity, whether it is
[45:57] power, whether it is chemicals, whether
[45:59] it is, you know, optical, what whatever,
[46:02] whether it's Bitcoin, like you
[46:04] essentially are searching through all of
[46:06] these markets looking for areas where
[46:08] there are shortages or scarcity and then
[46:10] you're deploying capital into it. Is
[46:11] that fair?
[46:13] You left one part out. So, as I
[46:15] mentioned, I was told you I wasn't that
[46:16] smart.
[46:17] I was trained by a handicapper.
[46:20] Everything to me, regardless of whether
[46:22] it's scarcity or abundance, has to do
[46:24] with one important thing, where's the
[46:25] money being bet right now, which has to
[46:27] do with sentiment. So, the odds on the
[46:29] tote board are based on the way people
[46:31] are betting. So, one of the the great
[46:35] things about talking to institutional
[46:37] clients, mutual funds, hedge funds, the
[46:39] sharpest people that are investing
[46:42] at tens to hundreds of billions of
[46:44] dollars
[46:45] is if they tell me why Marvell, and I
[46:47] just wrote a piece on it, and I go,
[46:49] "Well, because of silicon photonics and
[46:51] they're one of the" and they go,
[46:54] "Is that going to start kicking in now?"
[46:56] When I start hearing, I'm looking at the
[46:58] tote board going, "Oh, the odds on this
[46:59] are much higher than I thought." So,
[47:01] you're right that I think you're looking
[47:05] for places where like competition could
[47:07] be scarce. So, you know, when you get
[47:10] monopolies, you don't have much
[47:13] competition, so it's hard to break into
[47:14] their model. It's all about scarcity.
[47:16] Capitalism at the end of the day and
[47:18] economics is is all based on something
[47:20] having to do with scarcity. The reason I
[47:22] say abundance is a problem is cuz if you
[47:26] get to a world where everything is free,
[47:27] there is no capitalism.
[47:29] So, in this theory of abundance, it gets
[47:32] in, but there's still an exchange of
[47:33] services, they're just free. And this is
[47:36] where, if you listen to Elon Musk, he
[47:37] eventually gets, "I don't know what
[47:39] money means in the future." I'm kind of
[47:41] in that situation too, where if I look
[47:43] 20 years out, if I believe people are
[47:45] going to live
[47:47] a lot longer than the current lifespan,
[47:49] like
[47:50] much longer than the current lifespan,
[47:53] and I don't know what jobs are going to
[47:55] exist because humanoids will be here and
[47:57] Elon's phrase that I like to run with,
[48:00] which is oh there'll be jobs. It's just
[48:02] a question of whether you want to work.
[48:04] >> Yeah. And if that's the world we live
[48:06] in, then what does fiat money mean? What
[48:09] does the S&P 500 mean? What does that?
[48:12] And if that's only 20 years ahead
[48:14] because of how fast we're going. So you
[48:15] have to think if every year is like 10
[48:18] years of innovation. So take Joseph
[48:20] Schumpeter's destruction. That means in
[48:22] 20 years you're talking about
[48:24] two centuries worth of innovation.
[48:27] If I believe in 20 years that that world
[48:29] of abundance is here, then the only
[48:31] thing that I want to have as kind of a
[48:33] thought is why I think Bitcoin in the
[48:35] crypto world, the exchange of velocity
[48:37] of services through this
[48:39] value thing is the way people I should
[48:42] be. So if I think that's a
[48:44] 30% chance of happening and in my net
[48:47] worth if I have 5% in Bitcoin or 1% in
[48:51] Bitcoin, that's the wrong number. The
[48:53] right number is 30%. It's 20% and this
[48:56] is my pitch to wealth managers out there
[48:57] and our RFAs going it's a probability on
[49:00] if this fiat system's going to be the
[49:01] way you think it is and you should have
[49:03] a certain percentage in there.
[49:04] So
[49:05] we do a very good job of avoiding any of
[49:07] the political nonsense that goes on in
[49:09] the world. But what scares me about what
[49:11] you're saying is that Elon Musk and
[49:14] Mamdani agree.
[49:17] Hassan
[49:18] uh was it And Elon agree.
[49:22] Basically it is this breaking down of
[49:24] the high trust relationship between
[49:27] humans and capital
[49:29] in a world where
[49:31] you kind of get into undiscovered
[49:33] territory of like a free-for-all.
[49:36] And people have very different view. I
[49:38] mean I don't think Elon and Mamdani
[49:39] agree in terms of like, you know,
[49:41] explicit policies etc. But it's actually
[49:44] everyone sitting around saying hey this
[49:46] future world we're headed towards I got
[49:47] a different idea of what we can do to
[49:49] prepare for, or for, help people etc.
[49:53] But the more people that agree on that
[49:54] world being where we're headed, the more
[49:56] likely it is we head to that world.
[49:58] And that I think is a little
[50:00] nerve-wracking, right?
[50:02] Okay, so two things. One, you've pretty
[50:04] much described the fourth turning.
[50:06] Correct. Um okay.
[50:08] >> I read the book. It wasn't a waste of
[50:09] time. So, but I I the reason, you know,
[50:12] I've I've talked about my grandmother
[50:13] both with you and when I've sat down
[50:15] with with Natalie Brunell, um
[50:17] arguably the most important kind of
[50:19] person in the way that I think about
[50:21] markets other than my father. And the
[50:23] reason is cuz she was born in 1920
[50:25] during the Great Depression. So, if
[50:26] everyone goes back and says, are my
[50:28] kids, so that's the fourth, you know,
[50:31] I've got my grandmother, my mother,
[50:33] myself, and my kids. Great, they're the
[50:34] four turnings. They're They're the four
[50:36] generations.
[50:38] The difference between my kids and my
[50:39] grandma My grandmother was born in the
[50:41] Great Depression. To her, any debt was
[50:43] bad.
[50:44] She lived in a trailer home, a mobile
[50:45] home, I should say now. She She lived in
[50:47] a mobile home in Florida when she died,
[50:49] or not when she died, but when she had
[50:51] to eventually move to assisted living.
[50:54] Um
[50:55] if I think of about the way she thought
[50:56] about money,
[50:58] she had enough food money if she had a
[50:59] canned good in this. She's good. Well,
[51:01] my kids don't have that attitude. So,
[51:03] Mondani has existed for 80 years. It's
[51:06] just that the voters decided now is the
[51:08] time for him to be elected.
[51:10] >> You saw what happened in Manhattan, in
[51:11] the East Village? What? Uh the stat is
[51:14] 70% of people in the East Village voted
[51:16] for him. Right? Cuz all And for those
[51:18] who don't know New York City, it's
[51:19] basically a lot of young
[51:20] >> It's kind of like Williamsburg. That's
[51:21] kind of like Williamsburg, right? Yeah,
[51:23] yeah. It's uh it's basically a lot of
[51:25] young people. They first move to New
[51:26] York City, it's a little bit cheaper to
[51:27] live there. Um a lot of fun bars, you
[51:29] know, all the whole thing. Um but now
[51:31] they want to move. Uh I think it's one
[51:33] of the mental institutes
[51:34] to that neighborhood. And they're like
[51:36] right They're like suing him, and
[51:37] they're signing petitions, and like all
[51:39] this stuff. And of course, people on the
[51:40] internet are like, "Hey, like you kind
[51:41] of get the policies that you vote for."
[51:43] type thing.
[51:44] Um but But do think I mean, this is the
[51:46] whole game, right? It's like every
[51:48] generation's got to go through and kind
[51:49] of relearn these things and um
[51:52] maybe it is good that they learn rather
[51:54] than from a book from experience. It's
[51:55] just a little bumpy, you know, it's a
[51:57] bumpy landing when when they're doing
[51:59] it.
[51:59] >> You know you know what the best part of
[52:00] what you said? So, yeah, we don't get
[52:02] into politics here. There's a reason why
[52:03] I don't get into politics.
[52:05] Because I I I'm not only an independent,
[52:08] but I I love great leaders. That's what
[52:10] I love. I I'm and I I
[52:14] I'm not saying that no president has
[52:16] been a great leader, but when people are
[52:18] trying to get votes and they're doing
[52:19] things based on what will get them
[52:21] elected, there's a conflict that I think
[52:24] in the modern day has just become kind
[52:26] of an issue, especially when you bring
[52:27] social media in because these are
[52:28] one-line things. When I listened to
[52:30] Jensen Huang speak with Dwarkesh, I
[52:33] bought the stock Nvidia the next day. Um
[52:37] and the reason was because I actually
[52:39] love the way that he handled himself. I
[52:42] love that Dwarkesh pushed him.
[52:45] Every interview he's ever on, he's just
[52:47] very nice and affable and everything is
[52:49] a nice conversation. We love you. You're
[52:51] the you know, have the best company in
[52:53] the world. And Dwarkesh kept pushing
[52:54] back and it got into China versus the US
[52:57] and all this stuff and I thought he did
[52:59] a great job. So, I think one of the
[53:00] reasons that you and I don't talk about
[53:02] politics even off the air. It's just
[53:05] it's not something of my makeup. I don't
[53:07] watch CNBC. I don't watch Fox News. I
[53:09] don't watch news. And the reason is cuz
[53:12] I want to just go through and figure out
[53:13] what I want to believe in based on what
[53:15] goes on. I am very uh I I I I I I I
[53:18] I'm in the middle on everything. I make
[53:20] decisions and I'm I'm a gambler at heart
[53:22] in terms of my my risk reward, but I
[53:25] don't get into the thing. I just believe
[53:28] the Daily Stoic. And you and I have
[53:30] talked about it, but for people who have
[53:31] never bought the book, this is a
[53:32] shout-out to Ryan Holiday. Uh I I think
[53:35] Matt has the book, too. Yeah, he's
[53:36] showing it to me right now. Um this is a
[53:38] plug for that book.
[53:41] All of the things that you're talking
[53:42] about and all the emotions people have,
[53:43] the voting they do,
[53:46] reading a book or a passage in a given
[53:48] day is about something someone wrote
[53:50] about 2,000 years ago, and it was the
[53:53] same as what's going on in the East
[53:55] Village today, that's really cool to
[53:58] know that the human brain is doing the
[54:00] same anxiety, the same fears, the same,
[54:03] I could have a better life if they did
[54:04] this. Marcus Aurelius was going through
[54:06] that with the Roman Empire a long time
[54:08] ago, so I don't want to say who said
[54:10] this because it'll taint the way people
[54:12] view this, but there's a person who is
[54:14] very well known who
[54:16] gave an interview one time, and they
[54:17] asked him,
[54:19] you know, how do you deal with stress?
[54:21] And he basically was like, it doesn't
[54:22] matter. And they're like, what do you
[54:23] mean? He was like, you can be doing the
[54:25] single most important thing in your life
[54:27] and you know, focused on it and stressed
[54:28] and this and that, whatever, and then
[54:29] all of a sudden there's an earthquake in
[54:31] India and 400,000 people die.
[54:34] And it doesn't matter. And so again,
[54:37] very like kind of pessimistic view
[54:38] almost to it to a degree, but
[54:41] what is Marcus Aurelius, right? You know
[54:42] what I mean? Like it's I mean it's a
[54:43] very similar trend all throughout
[54:45] history. I think people are going to ask
[54:46] a lot of these questions because of the
[54:48] whole AI thing and like, what do we do
[54:49] with our time? You know, what what is
[54:51] the role of a human? What is the meaning
[54:52] of life? All this like crazy stuff that
[54:54] I don't know. We're going to find out.
[54:56] Uh we've I don't know if I mentioned
[54:58] this here, but um I think 9/11, the
[55:01] aftermath of 9/11 for me, because I was
[55:04] so impacted by it, because the best man
[55:06] at my wedding died in 9/11 and my best
[55:08] friend growing up. So, this was a major
[55:11] event for me where I questioned life. I
[55:13] left Morgan Stan, made the decision to
[55:15] leave Morgan Stanley no no more than a
[55:17] month after, um because I didn't want to
[55:20] spend my life firing people. I didn't
[55:21] want to be in a job that didn't bring me
[55:23] joy. I wanted to focus my attention on a
[55:25] very Buddhist philosophy, what does it
[55:27] bring me joy? Do they bring me joy?
[55:29] Whatever it is, it needs to bring me
[55:30] joy. I I made a very conscious decision
[55:33] in my life that I didn't want to be in a
[55:35] stressful situation with people forcing
[55:37] me what to do. Now, at that point, I
[55:39] didn't think I'd ever work for anyone
[55:40] again. I set up my own business. I ended
[55:42] up in another business. I told myself I
[55:43] didn't want to manage people, and
[55:44] eventually I was managing people again.
[55:45] So, it always tends to happen. Not this
[55:47] time, people. It's not going on. But,
[55:49] the reason I bring it up is
[55:51] those events in in in in life, A, they
[55:55] go on. That one had a huge impact on
[55:57] exactly what you were saying to me. When
[55:59] I ever get worried about anything,
[56:01] there's two things I think about. I did
[56:02] the eulogy for my friend, or one of the
[56:05] people I did the eulogy for him, and I
[56:07] did the eulogy for my grandmother. So,
[56:09] someone who died at 98, and someone who
[56:11] died in their 30s who had a long life to
[56:13] live.
[56:14] And in both cases, the same message came
[56:17] out, which is you don't know how long
[56:18] your life is going to be. You don't know
[56:19] if it's going to be 98. You don't know
[56:20] if it's going to be 33. You have no idea
[56:22] how long it's going to be. You get one
[56:23] chance,
[56:24] and if you ever start to stress about
[56:26] things, just be grateful because these
[56:29] people went through the Great
[56:31] Depression, and died in the towers
[56:34] having to make a decision on what to do.
[56:36] It's just It's crazy. I agree.
[56:38] All right. Where
[56:40] Where do you want us people to go?
[56:41] YouTube?
[56:42] Search on YouTube, Jordy Visser. Go
[56:44] there. Hit the subscribe button. And as
[56:46] a digital thank you, it's like sending
[56:48] me a gift card, you know? But, this is
[56:49] not money. It costs you zero other than
[56:51] 30 seconds of your time. If you got Open
[56:53] Claw, you can have your bot go do it. If
[56:55] you got Cowork, you can have your little
[56:57] bot go do it. Or you can be old school,
[56:59] you know, like an artisan. You can go
[57:01] and actually do it yourself, you know?
[57:03] Maneuver on over. Use your little
[57:04] keypad. Go and hit the follow button or
[57:07] subscribe button.
[57:08] See you guys next week. See you.