Has Bitcoin Hit The Bottom?

[0:00] On July 29th, when we have the
[0:01] FOMC meeting, will
[0:02] they raise
[0:03] rates? There is a probability
[0:05] , a good probability,
[0:07] somewhere around 35-40%, that they
[0:08] will do it. If not,
[0:10] I think Bitcoin
[0:11] will be above $70,000
[0:12] when that time comes
[0:13] , because I think people
[0:15] will start to take
[0:16] that fact into account. How are you
[0:17] guys? Today we have a
[0:18] great conversation with
[0:19] Jordi Visser. We
[0:20] talk about what's
[0:21] happening with the
[0:22] AI ​​slowdown in
[0:22] the middle of the cycle, and
[0:23] why he thinks it's
[0:24] over. Why is he
[0:24] so passionate about
[0:25] Bitcoin? He believes
[0:26] that a bottom is forming.
[0:28] Plus, we
[0:28] talk about
[0:29] robotics,
[0:29] commodities, and
[0:30] what public
[0:31] stocks
[0:32] Jordy is interested in. This conversation is
[0:33] full of nuance and
[0:34] we cover some
[0:35] big stories like
[0:36] Apple and what's
[0:37] happening with Meta or
[0:39] XAI and Grok. So you'll
[0:40] like it. Here is my
[0:41] latest conversation with
[0:42] Jordi Visser.
[0:43] Okay, Geordie. This
[0:45] week, Grok and Facebook
[0:46] made very loud
[0:47] announcements about their
[0:48] ambitious plans for AI
[0:49] . It looks like we have a
[0:51] price war on our hands, as
[0:52] both have started
[0:52] talking about the
[0:53] affordability of what
[0:54] they
[0:55] can actually produce. What are
[0:56] your thoughts on
[0:57] these two
[0:58] trying to make a
[0:59] comeback and surpass
[1:01] OpenAI and Anthropic?
[1:03] Uh, so...Oh, I'm
[1:06] speechless,
[1:07] trying to think of the
[1:09] best way to
[1:10] put it. I really
[1:14] believe that we need to
[1:15] start separating everything
[1:16] that happens in AI
[1:18] and models into two
[1:19] categories. I don't think
[1:24] anyone can
[1:25] catch up with Anthropic and
[1:26] OpenAI in their plans now.
[1:29] Now let me
[1:30] put this in a little
[1:31] context. I spent
[1:33] a lot more time
[1:34] on consumer agents.
[1:35] I spent a lot
[1:36] more time on what
[1:39] these models are trying to do, and what
[1:40] everyone is striving for. They are
[1:41] all trying to
[1:42] get some
[1:43] income, absolutely all of them.
[1:45] So when Meta talks
[1:46] about cloud ambitions,
[1:48] they need revenue
[1:49] now. And if your
[1:53] business is consumer
[1:55] agents, advertising,
[1:56] etc., and personal
[1:57] AI assistants, which is where I
[1:59] now officially
[2:00] include Meta, Apple, Google,
[2:01] Amazon. Again, not
[2:06] completely, but they
[2:07] all have that piece of the
[2:08] puzzle. Amazon, Google, and Microsoft
[2:11] obviously not
[2:12] only have a personal
[2:13] assistant, a "
[2:15] copilot" line, they also
[2:16] have a cloud business.
[2:18] Obviously, we
[2:19] have a cloud business now where
[2:20] we rent
[2:21] computing
[2:22] power to SpaceX.
[2:23] So, I think instead of
[2:25] focusing on
[2:26] models that, again
[2:28] , will become a
[2:28] commodity. Well
[2:30] , what matters is where
[2:32] you plan to
[2:32] get your
[2:33] income from. And it's clear that
[2:35] OpenAI and Anthropic are waging a
[2:37] war, uh, on a
[2:37] corporate level
[2:39] . And everyone else, I think,
[2:40] is trying
[2:41] to make sure that they are
[2:43] also spending
[2:44] money in other
[2:44] areas that they need.
[2:46] And of course, SpaceX, you
[2:47] know, Tesla ultimately,
[2:48] they have different
[2:49] ambitions for how they're going to
[2:51] use
[2:52] AI. So I think I
[2:54] don't see these price
[2:55] wars as anything more
[2:57] than, first of all, the media
[2:58] just blowing
[2:59] it up: "Oh, the token index is
[3:00] going down." This means
[3:02] we will have problems
[3:03] . I think everyone is
[3:04] just overestimating
[3:05] where their
[3:06] ROIC is going to come from and
[3:07] strategically
[3:07] directing their
[3:09] decision-making process in that
[3:10] direction.
[3:10] I like it
[3:11] when we, uh, maybe
[3:12] don't argue, but
[3:13] have slightly different
[3:14] views on it. Um,
[3:15] building Sylvia, I think you and I have
[3:17] been talking in recent weeks about,
[3:19] you know, that "mandate from
[3:20] heaven" 18 months ago
[3:21] . Everyone should
[3:22] use AI.
[3:23] Now suddenly people
[3:24] started looking at the uh- the
[3:25] token prices or,
[3:26] you know, the monthly
[3:28] computing costs were
[3:29] skyrocketing. They all
[3:30] started
[3:31] cutting it down. But
[3:32] a lot of what
[3:33] people were doing—both what
[3:34] we were doing at Sylvia and what
[3:35] I was talking about with
[3:36] other companies—
[3:37] they were essentially
[3:38] changing
[3:39] the architecture of how
[3:39] models were used
[3:41] . So they didn't
[3:41] change the
[3:42] model itself. It's just, hey,
[3:43] we don't want to
[3:44] approach the model
[3:45] if you ask
[3:46] what the date is or
[3:47] something like that. Now I
[3:48] think we've
[3:49] outgrown that stage, and what
[3:50] we're actively
[3:51] working on, and I think
[3:52] a lot of other
[3:53] people are too, is how to
[3:54] start directing
[3:56] queries to different
[3:57] models depending on the
[3:58] complexity or
[3:59] computational
[4:00] need of that query.
[4:01] And so it seems like this is an
[4:03] area where you're going
[4:04] to see a significant
[4:05] reduction in the
[4:05] computing
[4:06] costs of these companies in what
[4:07] they're
[4:08] spending. And the reason
[4:09] why this becomes interesting is because
[4:10] it
[4:12] opens up this whole
[4:12] open source story
[4:13] . So
[4:14] to me,
[4:15] Facebook and XAI's announcement about spending
[4:17] is actually an attempt
[4:18] to position themselves
[4:19] as, "Hey, maybe
[4:20] you don't need to
[4:20] go to
[4:21] Chinese
[4:22] open source or
[4:23] look for American
[4:24] open source. What
[4:25] if you could
[4:26] use one
[4:27] of these
[4:27] closed-source models, but for
[4:28] a lot less? It's
[4:29] almost like an acknowledgement of
[4:30] the fact that all these
[4:31] companies
[4:32] are now becoming very
[4:33] frugal in this
[4:34] regard. I don't know if
[4:35] this will work. I mean, I do
[4:37] n't know if this really
[4:38] drives wider
[4:39] adoption, but
[4:40] it seems strange that
[4:41] two different companies,
[4:42] who probably don't
[4:43] share plans
[4:44] with each other, are
[4:45] coming out in the same week
[4:46] saying, "Hey, this is a cost-
[4:47] effective
[4:48] way to get
[4:49] superintelligence through a
[4:51] model," you know?
[4:52] Yes. So, I, I, I
[4:53] disagree with you,
[4:54] so we'll move on. Um,
[4:57] I agree with the
[4:58] fact that, I made
[5:00] the visualization
[5:00] because many
[5:01] people approached
[5:03] me with this question from
[5:04] different perspectives. And
[5:07] then there were people
[5:08] I've known in the
[5:08] tech space for a long time,
[5:09] and they're all
[5:11] pessimistic, saying,
[5:12] "Oh, everyone's going to go to the
[5:12] cheaper model," and I'm
[5:13] like, "We
[5:14] need to start
[5:15] understanding where we
[5:16] are." So I'll
[5:17] put it to you
[5:17] differently. You probably have
[5:19] n't heard it in
[5:19] this format before. Um, and
[5:21] you're looking at this from
[5:22] Sylvia's perspective,
[5:23] which I understand. In my
[5:24] opinion, having experience
[5:25] managing hundreds of
[5:26] people over the years.
[5:28] Um, you get
[5:29] a resume, and you don't
[5:30] have to hire
[5:31] the most expensive specialists
[5:33] in your firm for
[5:34] every job.
[5:36] Tasks are now
[5:37] finally being delegated to
[5:38] models where it makes
[5:39] sense. Keeping in mind
[5:41] that 100 IQ units might
[5:42] be enough for
[5:43] something, as long as you don't
[5:44] need to do it
[5:45] very quickly. So, that
[5:47] brings me back to how
[5:48] I look at
[5:49] total
[5:50] payroll: it's
[5:51] hours worked
[5:52] and hourly pay.
[5:53] So, the number of
[5:54] people you
[5:55] hire is your
[5:56] total
[5:57] payroll. And I
[5:58] think every
[5:59] work team
[6:00] now realizes these
[6:03] top-
[6:04] level models have
[6:05] super-high IQs and they're
[6:06] very cheap. Dylan
[6:08] Patel gave an interview
[6:09] that I listened to today
[6:10] on Wisdom Tree. This is a great
[6:12] podcast. Highly
[6:13] recommend. He
[6:14] talked about his
[6:15] expenses. So this is
[6:16] Dylan Patel from SemiAnalysis
[6:18] . Very experienced
[6:19] specialist. His
[6:21] expenses have risen from
[6:22] $100,000 total for the
[6:24] firm as of
[6:25] last December
[6:26] . Now it's 11
[6:31] million, as he
[6:32] said, annual
[6:33] revenue costs. What
[6:35] he talked about as the
[6:36] importance of this—
[6:37] he's not going to
[6:39] change that. He doesn't
[6:40] try to
[6:40] give it up
[6:41] . He said that
[6:42] in some cases,
[6:43] due to increased
[6:44] productivity, he is
[6:45] able to generate
[6:46] more revenue. I
[6:47] think every
[6:48] business will be
[6:49] unique. I think
[6:51] businesses that did
[6:52] n't use them at all before
[6:54] are now figuring out
[6:55] where to direct their
[6:56] costs, and in any
[6:57] case they will have
[6:59] workflows that
[7:00] use
[7:01] cheaper models,
[7:02] perhaps open
[7:03] source. Then you
[7:06] will have more expensive
[7:07] models with high
[7:08] levels of intelligence for
[7:09] decision-making, where
[7:10] they are very efficient
[7:11] and cost-effective because
[7:12] they do the work
[7:13] ten or a hundred times
[7:14] faster than conventional
[7:15] processes. So it all
[7:16] depends on the
[7:17] specific situation. The
[7:19] end result
[7:20] will be a mixture of all these
[7:21] different models. I
[7:21] completely agree, we
[7:22] always had to come to this
[7:24] . But I think
[7:25] one of the challenging
[7:26] things is that since
[7:27] Opus 4.5, the intelligence of the
[7:29] models for the agents
[7:30] writing the code has been
[7:31] constantly
[7:32] improving.
[7:33] However, the implementation
[7:34] at enterprises has
[7:35] only just begun, and for
[7:36] the first six months
[7:37] they
[7:38] were actually learning to understand
[7:39] what the costs would be.
[7:41] They are now beginning to
[7:42] go through a phase of
[7:43] minor
[7:44] adjustment. Palantir is
[7:45] obviously promoting
[7:46] the fact that you don't...
[7:47] you know, they're going at it from a
[7:48] different angle. Oh, you
[7:50] don't want to have that.
[7:51] You don't want to be
[7:52] vulnerable. They
[7:53] list all the
[7:53] negative aspects
[7:54] of creating your own
[7:55] models. I think we are
[7:55] now at the stage of
[7:56] determining the price for
[7:57] everything. Models
[7:58] will become increasingly
[7:59] efficient. They
[8:00] will become
[8:00] cheaper and cheaper. they
[8:01] will become more and more
[8:02] intelligent. I think there
[8:03] will be a combination of both
[8:04] factors, but I don't
[8:05] think it will affect the
[8:06] overall costs.
[8:07] I think the level of
[8:08] implementation and
[8:09] usage is
[8:10] growing significantly and rapidly
[8:11] every day. Yes. And perhaps
[8:14] as a nuance, I actually
[8:15] believe that the number of
[8:16] tokens used
[8:18] will continue
[8:19] to grow. So I agree with that
[8:21] . But I think
[8:22] companies
[8:23] are trying to
[8:23] figure out, at least
[8:24] we're trying
[8:25] to figure out, how to
[8:26] consume more
[8:27] tokens but
[8:27] spend less
[8:28] money? And I think we're
[8:29] starting to see that
[8:31] if you're going
[8:32] to 4.7 or 4.8 or
[8:34] whatever, and someone
[8:35] asks what
[8:37] date it is, not only is there a
[8:38] cost, which
[8:39] is an obvious pain for the
[8:41] company, but there's also a
[8:42] big delay. And that's
[8:44] actually a bad
[8:46] user experience, whereas
[8:47] if you can
[8:48] use a
[8:49] lightning-fast
[8:50] open
[8:51] source model, make sure it's
[8:52] accurate, etc., it
[8:54] becomes quite interesting.
[8:55] Um, but some
[8:56] companies don't care,
[8:57] right? They are not
[8:58] concerned about costs.
[8:59] Maybe Dylan
[9:00] Patellis says, "
[9:01] Look, I'll spend
[9:02] $50 million
[9:02] because for every
[9:03] dollar I spend, I
[9:04] get $10 in
[9:05] profit." Then,
[9:06] you know, it doesn't
[9:07] matter anymore, right?
[9:08] Yes. And this is exactly what filters out
[9:10] unsuccessful
[9:11] companies. And
[9:13] his literal quote seems to have
[9:14] been: If you don't
[9:16] spend
[9:17] more money every year, you'll
[9:18] go bankrupt. That is,
[9:19] if you try to
[9:21] optimize costs
[9:22] based on the fact that you
[9:23] don't want to spend a
[9:24] certain amount, and you
[9:25] cut costs
[9:26] on AI—you're bankrupt. And I,
[9:28] by the way, agree with that.
[9:30] This means your
[9:31] AI budget should
[9:32] increase every year
[9:33] . No questions asked.
[9:35] It absolutely
[9:36] must grow. And that's the
[9:37] only way to survive
[9:39] as a business. This gives
[9:40] startups a huge
[9:41] advantage. That's why
[9:43] SemiAnalysis is very
[9:44] different,
[9:45] let's say, from others:
[9:46] they cover data
[9:48] centers,
[9:48] semiconductors, and the
[9:49] entire
[9:51] AI infrastructure. If you're Goldman Sachs
[9:52] or Morgan Stanley, you have a
[9:53] huge staff and
[9:54] big expenses, so
[9:56] your approach will be different
[9:57] than a company that's
[9:58] just starting out and was
[10:01] only spending $100,000
[10:02] a year on AI in December.
[10:02] Yes, that makes sense. Um,
[10:04] what about Apple? Apple,
[10:05] of course, made an announcement
[10:07] about Siri—the integration of
[10:08] AI into Siri—and also
[10:09] announced
[10:10] price increases due to
[10:12] memory requirements. What do
[10:13] you think is
[10:14] happening to this
[10:15] business? Actually,
[10:17] I...Last week,
[10:22] because I was spending
[10:24] a lot of time on Meta, I
[10:25] listened to more
[10:26] podcasts featuring
[10:27] Meta. This week I
[10:30] listened to an episode with
[10:31] Alex Kananteritz
[10:33] , he was the
[10:33] CTO of
[10:35] Meta, and I'll be covering
[10:36] it on YouTube. When
[10:37] a story comes up
[10:38] , I try
[10:39] to understand what's behind it
[10:40] . They are entering the
[10:41] cloud
[10:42] technology market. They say the
[10:43] agent world is not
[10:44] where they
[10:45] expected
[10:46] to see it. Well, then
[10:47] I had to
[10:48] figure it out. So I
[10:49] started doing a bunch of
[10:50] research on
[10:51] consumer agents. And that's how
[10:52] I want to
[10:52] lay everything out
[10:53] before I
[10:54] move on to
[10:54] discussing Apple. So far,
[10:56] we have only worked
[10:57] with agents to
[10:57] write code. And I
[10:58] want people to
[10:59] understand that this is
[11:01] essentially the language of AI. We are
[11:02] using more and
[11:03] more agents for
[11:04] coding, and in this
[11:05] agent-based direction,
[11:06] workflows
[11:07] are becoming longer and
[11:08] more complex, so they
[11:09] need some kind of
[11:10] memory. We see the
[11:13] use of tokens
[11:14] growing
[11:14] exponentially.
[11:16] Consumer agents
[11:17] will drive another
[11:18] round of growth.
[11:21] Listening to Meta and other
[11:23] people talking
[11:24] about this, it becomes
[11:25] clear the
[11:25] complexity of the work of
[11:26] consumer agents, the
[11:28] commercial side,
[11:29] for example: "book
[11:30] me a trip to
[11:31] Italy." That's what
[11:34] worries me. This is much more
[11:35] complicated than "
[11:35] create me this
[11:36] table" or "build
[11:37] this model." You
[11:38] have to take
[11:39] many more
[11:40] steps. You should
[11:41] visit many
[11:41] more sites. This
[11:42] requires significantly
[11:43] more
[11:43] computing
[11:44] resources. So
[11:45] computing
[11:46] power
[11:46] will grow
[11:47] exponentially
[11:48] once they can
[11:49] solve this. Why the
[11:51] Apple case is interesting is the same as
[11:53] with Meta, because Google and
[11:54] Amazon have their own solutions:
[11:56] Alexa at Amazon,
[11:56] personal
[11:58] intelligence at Google. Apple
[12:00] is trying to make
[12:02] Siri work, and Meta
[12:03] is looking to monetize
[12:04] Facebook, Instagram, and everything else
[12:06] with
[12:07] advertising and
[12:08] artificial intelligence, and
[12:09] they all
[12:11] have their own
[12:11] hardware. Again
[12:13] , Alexa, phone—
[12:14] if you go through
[12:15] the list, each of this
[12:17] group has its own
[12:18] hardware. So
[12:19] the question is
[12:20] how difficult is it,
[12:22] given Opus 4.8.
[12:24] We have Claude 3.5, but Siri
[12:26] still doesn't work. Alexa
[12:28] still doesn't work. How is this
[12:29] possible? Well, the
[12:31] complexity of what
[12:32] they are trying
[12:33] to solve is
[12:34] incredible.
[12:35] I
[12:36] used Grok again this week in
[12:37] my Tesla. Grok at Tesla
[12:39] is working phenomenally.
[12:41] The delay is very small.
[12:43] Why can't they
[12:45] do this with Siri? Well,
[12:46] the machine is always
[12:47] on. I'm just
[12:48] riding in it. If your
[12:49] phone was always
[12:51] on, you
[12:52] could chat with ChatGPT.
[12:53] That would work. So I
[12:55] think the
[12:56] most interesting thing about Apple is
[12:57] my usual perspective.
[12:58] The stock fell, despite
[12:59] hitting new all-time
[13:01] highs, when they
[13:02] announced Siri Personal
[13:03] Intelligence and showed off its
[13:04] capabilities. Someone
[13:05] said it was terrible.
[13:07] Many people said it was
[13:08] amazingly
[13:09] good. Then there was the
[13:11] problem with
[13:12] memory prices. Okay,
[13:13] memory prices are going up.
[13:14] We will have to
[13:15] raise prices. Stocks
[13:16] suffered on both
[13:17] days. On the day of
[13:19] Siri's announcement, they reached
[13:20] all-time
[13:20] highs and then fell.
[13:21] Then they fell
[13:22] on the news about
[13:23] the memory, falling
[13:25] more than 10% from
[13:26] the highs, and then,
[13:27] without any news,
[13:28] returned back to the
[13:29] peaks. This is not a cheap
[13:31] promotion. This shows that
[13:33] the market is
[13:34] bullish even before
[13:35] bad news. And I
[13:36] like to
[13:37] figure things out.
[13:38] I think the market, and
[13:39] especially very
[13:40] smart people,
[13:41] are starting to look
[13:43] 6 months ahead.
[13:44] So, this is the first
[13:45] warning I
[13:46] want to give people.
[13:47] Infrastructure
[13:48] investments were a
[13:49] great thing for
[13:50] half a year. I still
[13:51] think you will make
[13:52] a lot of money
[13:53] investing in
[13:54] the beneficiaries of
[13:55] infrastructure
[13:56] development, but it will
[13:57] not be that easy. So I
[13:59] think the
[14:01] AI ​​slowdown is over, because
[14:02] right now everyone is a little
[14:03] too skeptical
[14:04] about AI development and
[14:05] worried about the
[14:07] computing
[14:08] power issues. I think
[14:09] all of this is shifting
[14:11] towards two main
[14:12] themes. First, with
[14:13] the emergence of consumer
[14:14] agents,
[14:16] Apple, Meta, and
[14:17] other stocks need to be revalued, as
[14:18] ROIC will start
[14:20] to increase. And I think
[14:21] people are becoming more
[14:22] optimistic about
[14:23] these companies,
[14:24] realizing that it's
[14:25] inevitable. The second
[14:26] aspect is the moment
[14:27] when you start
[14:28] moving into
[14:28] the cryptosphere through
[14:29] agent commerce.
[14:30] They are
[14:32] interconnected. So I expect
[14:33] that in a year the main
[14:34] topic will be: "
[14:36] How did we miss Apple?"
[14:38] I'm not sure
[14:39] about Meta yet, as it doesn't
[14:40] have the same technical
[14:41] picture, but I'll be keeping a closer eye
[14:42] on
[14:43] hyperscalers as
[14:44] we approach a point where
[14:46] the surprise
[14:47] move to consumer
[14:48] agents rather than
[14:49] programmatic ones could be.
[14:51] What do you do with Grok in
[14:52] your Tesla?
[14:54] It was something more
[14:55] like a live conversation
[14:56] . The irony is that I
[14:58] wanted to talk
[14:59] about the very thing I just
[15:01] discussed. When I'm
[15:03] driving for,
[15:04] say, an hour in
[15:05] Maine, and I've just
[15:06] listened to a podcast, I
[15:08] pause it.
[15:10] I turn to Grok to
[15:12] develop this conversation.
[15:14] It works
[15:16] much better in a car than
[15:17] when you're walking with a
[15:18] chatbot. And all because
[15:20] it is turned on
[15:21] all the time, and the signal in the
[15:22] car is much
[15:24] better. So to me,
[15:26] Grok
[15:27] seems more
[15:28] human when talking. And again
[15:29] , it's that feeling of
[15:30] talking to a very
[15:31] smart person: you
[15:32] just say, "Hey, I
[15:34] want to talk about
[15:35] Apple, Meta, and consumer
[15:36] agents." I will cite just
[15:38] one fact that
[15:39] emerged from this. Goldman
[15:41] Sachs published a report
[15:43] stating that
[15:44] consumer agents
[15:45] will consume
[15:46] approximately 30 times
[15:47] more
[15:47] computing
[15:48] resources than
[15:49] programming agents.
[15:51] That's why there is no
[15:52] final solution
[15:53] regarding computing
[15:54] power. If you
[15:55] think there is a
[15:56] surplus of
[15:57] computing
[15:58] power, you are
[15:59] deeply mistaken.
[16:00] Ladies and gentlemen, I
[16:01] specialize in
[16:02] software, but
[16:03] sometimes I
[16:03] come across a piece of hardware that
[16:04] I simply
[16:05] cannot live without. And for
[16:06] me it's Plaud. This thing is
[16:08] my insurance policy
[16:09] against
[16:10] memory problems, because every day I
[16:11] communicate with
[16:13] company founders,
[16:14] investors, and
[16:14] a lot of other people.
[16:16] There are three things I
[16:17] always want to take away from the
[16:18] conversation. What
[16:19] did I learn, what are the
[16:20] next steps or
[16:21] tasks, and what
[16:22] notes will I
[16:22] need in the
[16:23] future? You
[16:24] can even
[16:25] mark important
[16:26] points during a
[16:26] discussion to
[16:27] quickly find them
[16:28] later. Whether it's
[16:29] a meeting with
[16:30] a founder, a call to
[16:31] an investor, an interview
[16:32] for a podcast, or
[16:33] even a conversation with
[16:34] your wife. I can
[16:35] instantly review
[16:36] everything and find the key
[16:37] ideas, solutions, and
[16:38] tasks that
[16:39] were discussed in the conversation. Over
[16:40] time, every conversation
[16:41] suddenly becomes an asset,
[16:42] not a forgotten moment
[16:44] . In a world where
[16:44] information is abundant,
[16:45] the ability to retain
[16:46] and use
[16:47] knowledge is a true
[16:48] competitive
[16:49] advantage. That's why
[16:50] Plaud Note Pro recently
[16:51] caught my attention.
[16:52] This is a
[16:53] credit card sized notepad.
[16:54] Slim, portable,
[16:55] integrates perfectly with my
[16:56] phone. And what
[16:57] I like most is that
[16:58] it automatically
[16:59] records conversations. It
[17:00] turns them into
[17:01] presentation-ready summaries
[17:02] . You get a
[17:02] structured
[17:03] task list and
[17:04] flawless
[17:04] transcription. It's
[17:05] truly magic and you
[17:06] should try it. So
[17:07] get to know them.
[17:08] This device
[17:09] has completely changed my
[17:10] life, and I think you
[17:11] will find it
[17:12] very useful too. If
[17:13] you want to
[17:13] use Plaud and
[17:15] benefit from it as much as
[17:16] I do,
[17:17] visit plaud.ai/pmp. You
[17:18] can go to
[17:19] plaud.ai/pmp and get
[17:20] 15% off if
[17:21] you use the code PUMP,
[17:22] or just follow the
[17:24] link in the description
[17:25] . So visit
[17:26] plaud.ai/pmp today.
[17:28] When we look at the
[17:29] problem of
[17:30] computing
[17:30] power, this is,
[17:31] of course, one of the
[17:32] main issues. People
[17:33] saw Meta's statement and
[17:34] started to
[17:35] get worried about it. We
[17:37] continue to
[17:37] see
[17:38] volatility in
[17:39] many of these
[17:40] companies. Regarding Bloom Energy,
[17:42] many
[17:43] reports from
[17:44] "bearish" sellers are now starting to appear.
[17:46] It seems that
[17:46] pessimists are simply
[17:47] waiting for the right
[17:49] moment, believing that
[17:50] all this is overrated.
[17:52] Can short sellers
[17:54] make money in
[17:55] such a market? How do you
[17:56] feel about the
[17:56] opposing opinion?
[17:58] Yes? You are
[17:58] optimistic about
[17:59] this. I am
[18:00] optimistic about
[18:01] this. Many people
[18:01] are buying these assets. What
[18:03] about the other side? Do
[18:04] they have any valid
[18:05] points that
[18:06] you think are worth
[18:07] paying attention to?
[18:09] Well, of course. Listen,
[18:11] let's go back to how
[18:12] I sold Micron shares,
[18:13] but I still believe that the
[18:15] trend in the
[18:16] memory market will continue.
[18:17] Actually, to give
[18:18] people a little
[18:19] announcement. Part of the
[18:23] reason I
[18:24] left Micron was because
[18:25] speculation
[18:26] was becoming apparent everywhere. And
[18:30] again, it was
[18:31] leverage. So when I
[18:35] see
[18:37] leverage forming, which is evident
[18:38] in Korea, I see
[18:39] leverage in
[18:40] macro strategies and in
[18:41] fundamental
[18:42] "long-short" strategies, where
[18:43] half a year ago they were
[18:44] telling me that memory is
[18:46] already expensive, and now they are
[18:47] bragging that this
[18:48] trend will continue for another 5
[18:49] years. I'm just
[18:52] starting to de-
[18:53] risk and believe there are
[18:54] better places to
[18:55] invest my money. And
[18:57] now I have to accept the
[18:58] fact that in order to
[18:59] buy back
[19:00] Micron, I will have to
[19:01] buy more than
[19:03] I sold. There is no
[19:04] doubt about it. I have no
[19:05] problem with this
[19:06] , because I have
[19:07] been a trader all my life. But I
[19:08] really think there are
[19:10] ways Micron could
[19:11] have gone up to 1300. So
[19:12] if you sold at 1250
[19:14] and bought back at 900, you
[19:16] can trade, you
[19:17] know, bubbles,
[19:19] parabolas and all the things
[19:20] I was talking about,
[19:22] you're going to have
[19:24] 30-40% corrections. And in many
[19:25] of these companies, we
[19:26] have seen some of the
[19:27] largest corrections
[19:28] in the history of their stocks
[19:29] . And this, you know, is
[19:30] serious. Samsung fell 20% in a 10-
[19:33] day period
[19:35] . And this is despite the fact that
[19:38] they earn
[19:39] enough, in fact,
[19:41] more than in
[19:42] their entire 40-year history.
[19:44] That is what is expected
[19:45] this year. So prices
[19:46] may fall. You can
[19:48] open a short
[19:49] position on Bloom Energy. I
[19:50] just think people
[19:51] make a mistake if they
[19:52] look at it in the
[19:53] long
[19:54] term. My
[19:55] general rule as a
[19:56] macro analyst is: I
[19:57] try to understand
[19:58] where the river is flowing, and either
[20:00] swim upstream
[20:01] or don't trade at all
[20:02] . That's exactly how
[20:04] I thought about
[20:05] Micron. I still believe
[20:06] we'll see 4000-5000 at
[20:08] Micron before it's all
[20:09] over. I think it
[20:10] will be harder
[20:12] to go from 1000 to
[20:13] 5000 than from 100 to 500. And
[20:15] that's my general
[20:16] approach: I will
[20:16] trade them
[20:17] more often. I will go out
[20:18] . I'll go in
[20:19] and look for
[20:20] moments. But I believe
[20:21] those who want to play
[20:22] for relegation
[20:23] need to have a
[20:24] shorter
[20:25] planning horizon.
[20:25] You mentioned Samsung.
[20:27] Tell us a little about
[20:28] them. They
[20:28] reported
[20:29] super-high profits,
[20:30] but the stock fell. What's
[20:31] going on there?
[20:33] So, like I said,
[20:34] the numbers are just
[20:35] staggering, and for
[20:36] those who aren't already
[20:37] familiar with the
[20:38] story:
[20:42] they're expected to make
[20:45] $217 billion
[20:46] this year. I
[20:48] mean, with the letter "M" -
[20:49] this is the amount they
[20:50] have to earn, it's not
[20:52] income, it's
[20:53] profit this year.
[20:55] Over the past 40 years,
[20:57] they have
[20:58] lost a total of $217 billion
[20:59] . So it's worth
[21:02] considering the context
[21:03] and understanding: first,
[21:04] when people say it's a
[21:06] "bubble," it's
[21:07] not, because the profits are
[21:08] colossal. Second,
[21:10] the volumes of chips
[21:11] they sell and
[21:12] how far into the
[21:13] future they
[21:14] will sell them.
[21:15] Creating capacity
[21:16] for memory production is
[21:18] not easy at all. Therefore, Samsung,
[21:19] along with SK Hynix and Micron,
[21:21] have benefited significantly and
[21:22] will continue to
[21:24] benefit significantly.
[21:26] But, as I said,
[21:27] if you look back over the
[21:29] last 25 years, the
[21:30] only time they
[21:31] had a correction of more than 20
[21:33] % in 10 days seems to have
[21:35] been when the world
[21:36] shut down because of a virus
[21:38] that was supposed to kill everyone. And
[21:39] it was very similar to
[21:40] what happened immediately
[21:42] after the collapse of Lehman Brothers.
[21:43] These are the only two cases
[21:44] in 25 years. And yet
[21:46] they
[21:46] remain near
[21:47] historic
[21:48] highs, even
[21:49] despite a 20% drop. And
[21:50] it's still parabolic
[21:51] growth. Here's what I'll
[21:52] say. I saw a
[21:54] lot of people
[21:55] react very negatively
[21:56] because we like to say, "
[21:57] Oh, they fell for the
[21:58] good news."
[22:01] The problem is that at
[22:03] that point they had already
[22:04] fallen 15% from their
[22:05] highs. So I'm
[22:07] much
[22:08] less concerned about it. In these
[22:09] "selling
[22:10] on the news" situations, I look for
[22:11] moments when a stock is
[22:12] at an all-time
[22:13] high,
[22:14] opens higher,
[22:15] then falls, catching
[22:16] everyone by surprise—
[22:17] this usually
[22:17] results in
[22:18] a correction of about
[22:19] two months. These
[22:20] assets have been adjusting for
[22:22] about 4-6 weeks now.
[22:23] And I'm talking about 4-6
[22:25] weeks. Even though
[22:26] they reached
[22:27] new all-time
[22:27] highs,
[22:28] weakness began
[22:28] to show in
[22:29] late May. And that's when
[22:30] I started to
[22:31] make the case
[22:32] that, okay, I really do
[22:33] think that the
[22:34] AI ​​slowdown is already here. And all
[22:36] we have done at this
[22:37] point is
[22:37] eliminate
[22:39] leverage in three
[22:39] segments.
[22:40] Systematic
[22:41] quantitative
[22:42] MA strategies were forced
[22:43] to reduce their
[22:44] leverage. We
[22:45] also saw
[22:46] Korean investors
[22:47] being forced
[22:47] to reduce positions.
[22:49] We saw an ETF, a
[22:50] dual ETF in
[22:51] Hong Kong, fall 62%.
[22:54] In my opinion, we
[22:54] have seen enough
[22:55] signs that the market
[22:57] has become more
[22:57] balanced. And if
[22:59] you look at what
[23:00] SK Hynix did with its
[23:01] ADR issue, and
[23:02] how high the
[23:03] demand was during the IPO, and when
[23:04] you read that
[23:05] most of the buyers
[23:07] were long-term
[23:08] investors,
[23:09] you understand: the
[23:10] buy-only managers
[23:11] are showing their cards,
[23:13] namely—they
[23:13] missed this
[23:14] growth. moving
[23:16] too fast for the
[23:16] pace at which they are
[23:17] investing. They
[23:18] will be on the lookout
[23:19] to buy assets.
[23:20] So I'm pretty
[23:21] sure we've
[23:22] done enough
[23:23] damage. We may
[23:24] see some more
[23:24] new lows
[23:25] during the reporting season,
[23:26] especially for some
[23:27] stocks, but
[23:28] volatility is
[23:29] too high in
[23:29] momentum assets.
[23:31] It will decrease, and this
[23:32] will happen due to
[23:33] less correlation between
[23:34] the names.
[23:34] I know that as the
[23:36] AI ​​sector is perhaps
[23:37] preparing to emerge from
[23:38] this cyclical
[23:39] slowdown, you
[23:40] have become more
[23:41] bullish on
[23:42] Bitcoin. What's
[23:43] going on there? Well,
[23:46] actually, we've already
[23:47] talked about this, but I'll
[23:49] just highlight
[23:50] a few things and then
[23:51] move on to the larger
[23:52] topic that
[23:53] I'm starting to explain to
[23:55] managers and
[23:56] why I'm
[23:57] launching a crypto YouTube channel in the first place
[23:58] ,
[23:59] planned for
[24:00] September. As people
[24:02] will find out, I am still a
[24:03] trader
[24:04] and always will be. I will
[24:06] always look for the
[24:06] best moment to
[24:08] enter. But when you
[24:09] run a business, you
[24:10] want to launch
[24:11] things when
[24:12] people start
[24:13] looking for them. And I think
[24:15] this is a very important
[24:16] story. So, first off:
[24:19] I'm always looking for
[24:19] technical signals, and I
[24:21] finally got my
[24:22] first RSI divergence
[24:23] since the peak
[24:24] late last
[24:26] year. This divergence
[24:29] used
[24:31] essentially the 4-hour RSI and I
[24:32] look for points where price
[24:34] sets a new
[24:35] low, which happened
[24:36] when we recently
[24:38] broke 60,000, but the RSI is
[24:39] higher than it was at the
[24:40] previous
[24:41] low. Everything is simple
[24:43] and clear. Well, this
[24:44] was the first time
[24:45] we got it. So,
[24:46] as a trader, I think, "
[24:47] Well, now I can
[24:48] buy something when we
[24:49] get back above 60, and
[24:50] below the lows I'll
[24:51] just close
[24:52] the position." That's how a trader trades
[24:53] , and that's how
[24:54] I've always
[24:54] thought. This is how I was
[24:56] raised. I am a
[24:56] believer in
[24:57] Elliott Wave Theory and still
[24:58] believe that we are entering
[24:59] something very, very
[25:00] big. I believe that
[25:01] over the next
[25:02] year we will be
[25:02] close to the lower
[25:03] end of
[25:04] Bitcoin's range.
[25:05] Notice how I said:
[25:06] close. Can we
[25:07] go to 50? Yes. Can
[25:08] we go to 45?
[25:10] Yes. Do I believe that
[25:11] in a year we will be
[25:12] above 100? Yes. So what
[25:13] difference does it make to me whether
[25:14] I buy something for 60 or
[25:15] at a different price? Yes,
[25:17] my percentage
[25:18] return will be better,
[25:19] but if we're above 100,
[25:20] I don't care. This is the
[25:21] first. Second—I was
[25:24] making the point that we
[25:26] probably all
[25:26] underestimated—I definitely
[25:28] underestimated—the
[25:29] impact that the
[25:30] capital outflow into AI had on
[25:31] everything else. This
[25:34] applies not only to
[25:35] Bitcoin, but also to
[25:36] hyperscalers. This is
[25:37] essentially my
[25:38] benchmark arbitrage of
[25:39] capital
[25:40] flowing into the
[25:42] fastest growing sectors. Micron grew
[25:44] 20 times.
[25:46] Twenty-fold
[25:47] growth. This doesn't
[25:48] happen with big
[25:49] companies, and this is a
[25:50] big company. So for
[25:51] this to happen, you didn't
[25:52] need to
[25:53] buy Bitcoin. You did
[25:54] n't need to
[25:55] buy crypto. You did
[25:56] n't have to
[25:57] buy any of
[25:58] this. Everyone in
[25:59] Silicon Valley
[26:00] was talking about startups:
[26:01] if they added
[26:02] AI, no one wanted
[26:03] crypto. Ultimately, people
[26:05] want to invest in what
[26:06] works. And so,
[26:08] when the price started
[26:09] to decline, which
[26:10] coincided with
[26:11] the release of Opus 4.5 in October, it
[26:13] also coincided with
[26:14] another strong
[26:15] long-term
[26:16] trend that had been in place until
[26:17] last week—
[26:18] lower
[26:19] interest rates.
[26:21] As of October or
[26:23] late September of
[26:24] last year, the market was
[26:25] still pricing in 150
[26:26] basis points of
[26:27] rate cuts. Then
[26:29] it peaked.
[26:30] Some of that
[26:31] expectation disappeared
[26:32] because of the actual
[26:32] decline, but a significant
[26:34] portion disappeared because we
[26:35] started to hedge for
[26:36] an increase. And now
[26:38] we have a full decline
[26:40] plus a percentage,
[26:40] sorry, increase,
[26:42] plus a high
[26:42] probability of another
[26:43] one by the end of the year. So here's what
[26:47] I
[26:48] think will happen. Um,
[26:49] all of that is going to change, or
[26:51] at least there
[26:51] will be a change of
[26:52] pace. And I'm
[26:53] an expert at changing
[26:54] pace. I don't think
[26:55] they want to
[26:56] raise the stakes. I
[26:57] prepared a piece
[26:58] on the Fed this week and
[27:00] made sure everyone
[27:01] noticed: Fed Chairman
[27:02] Walsh made a
[27:03] statement. He believes
[27:05] AI will cause a
[27:06] productivity boom, and there
[27:08] may be some initial
[27:09] inflationary pressure
[27:10] before we move
[27:12] into deflation, but he is
[27:13] least keen to
[27:14] trigger a vicious
[27:15] cycle of
[27:16] rate hikes due to
[27:16] inflation. We
[27:17] will get a negative
[27:19] CPI. The question
[27:20] is, will
[27:22] they raise rates at the
[27:23] FOMC meeting on July 29?
[27:25] There is a probability, a fairly
[27:26] high probability,
[27:28] about 35-40%, that this
[27:29] will happen. If not,
[27:31] I think
[27:32] Bitcoin will be
[27:33] above 70,000 by then as
[27:34] people start
[27:35] to price in the fact
[27:36] that they might
[27:37] not raise rates at all
[27:39] this year before the
[27:40] midterm elections.
[27:41] Does he really want to
[27:42] raise the stakes
[27:43] before the election,
[27:43] considering he was
[27:44] appointed by Trump? So
[27:45] I think it might
[27:46] be postponed for a
[27:47] while. And that's not
[27:48] so important because
[27:50] it means they're not
[27:51] lowering rates. But
[27:52] since an increase
[27:53] is expected, this is
[27:54] positive against the background of
[27:55] forecasts. This is how I
[27:57] look at this
[27:57] situation. But I want to
[27:59] leave one more thing
[28:00] for people, and that's
[28:01] about the bigger
[28:02] picture. Scott Besson
[28:04] gave a speech
[28:06] at the Economic Club of
[28:07] New York, and on
[28:08] Friday I
[28:09] published in X what
[28:10] Matt Hogan had put out. I
[28:11] really think
[28:12] people should
[28:13] read this speech.
[28:14] I think they should also
[28:15] read
[28:16] Muhammad Alarian's article on
[28:17] this subject. It's a
[28:18] fact, guys, and we're
[28:20] talking about it here.
[28:22] The administration is very
[28:24] focused on changing
[28:25] how the US does
[28:26] business in the world, and
[28:27] part of that change is making
[28:30] sure it
[28:31] doesn't lose control
[28:32] of the financial
[28:33] system or
[28:34] remain a leader in
[28:35] setting
[28:36] financial frameworks. He
[28:39] specifically noted
[28:41] that digital assets,
[28:42] tokenization, and
[28:43] stablecoins are
[28:44] part of
[28:45] the administration’s focus on
[28:46] this new
[28:47] economic order
[28:48] . You must understand
[28:51] what this means. They are
[28:52] focused on making
[28:53] cryptoassets
[28:54] part of the
[28:55] financial framework, and
[28:56] this coincides with the
[28:57] development of AI. So I
[28:59] called it "AI Macro
[29:00] Nexus," which is why I run
[29:01] my YouTube, and it's
[29:02] about 90% a
[29:03] mix of AI's impact
[29:05] on the macro world
[29:06] we live in today, and
[29:08] 10% about crypto and its
[29:09] significance. We're at a
[29:11] point now where
[29:12] the reason I'm going to be
[29:14] running a YouTube show—which
[29:15] will be about 70%
[29:16] about crypto and
[29:17] then showing the
[29:18] relationship between
[29:19] the macro world and AI—is
[29:20] because we're about to
[29:21] enter the
[29:22] era of AI-agent
[29:23] commerce, which
[29:24] means the velocity of
[29:26] money is going
[29:26] to increase. Dormant
[29:28] assets in the form of
[29:29] housing and real estate
[29:31] around the world
[29:31] will be liquidated
[29:33] and converted into
[29:34] real money
[29:34] over the next
[29:35] five years through
[29:37] tokenization. These are all
[29:38] big positive
[29:39] developments, and that is why
[29:41] I believe you
[29:42] are at
[29:42] the beginning of a
[29:43] multi-year Bitcoin
[29:45] bull market. And whether
[29:47] it starts in October, December or
[29:49] right now, I think
[29:50] we're at the
[29:51] lower end of the
[29:52] range for
[29:52] next year. I
[29:54] recently saw
[29:55] statistics that
[29:56] showed that stablecoin volumes
[29:58] increase on
[29:59] weekends, which makes perfect
[30:00] sense. Banks
[30:01] are closed, so people
[30:02] want to
[30:03] access them. We also
[30:05] saw the announcement of an
[30:06] open
[30:07] stablecoin consortium.
[30:09] It seems there is Mastercard,
[30:11] Visa, and Stripe. A lot of people are
[30:13] getting involved in this, and it
[30:14] feels like
[30:15] some people are just thinking, “
[30:17] How do I get into the
[30:18] stablecoin game?”
[30:19] » But largely it's about the fact
[30:20] that
[30:21] agents need
[30:22] money, and what
[30:23] money they're going to
[30:24] use, how
[30:25] they're going to
[30:26] use it, and so on. Are
[30:28] you seeing anything on
[30:29] that front, or have
[30:30] you heard anything from fund
[30:31] managers that
[30:32] suggests
[30:34] they are starting to
[30:34] shift some of
[30:36] their funds into
[30:37] stablecoins or
[30:38] Bitcoin? Do you
[30:39] think we're still at the very
[30:40] beginning of this
[30:42] reversal, and you're
[30:43] starting to buy
[30:44] now because you
[30:45] think that's what
[30:46] people will be doing
[30:47] in 3 months? So,
[30:49] price is ahead of
[30:50] narrative. The first thing that
[30:51] always happens at
[30:52] the bottom of the market is people start
[30:53] closing out short
[30:54] positions because people are
[30:55] frustrated that
[30:56] it's been so long. I
[30:56] think Bitcoin was a
[30:58] source of
[30:58] funding for
[31:00] AI-related trade. I think
[31:01] it started when the
[31:02] software industry
[31:04] crashed and all the
[31:05] software companies were
[31:06] hit hard. And
[31:07] Bitcoin was part of
[31:08] this process. I would
[31:09] say that anything
[31:10] built on code
[31:11] could not avoid
[31:12] correlation. So if
[31:13] you wanted to hedge
[31:14] your semiconductors
[31:15] , which now had a
[31:16] high beta, you
[31:17] needed to find
[31:18] shorts with high beta
[31:19] . And these
[31:21] high-beta shorts were
[31:22] Salesforce.com, Adobe, Workday, and
[31:24] Bitcoin. And I think
[31:26] all of these things have become
[31:27] part of impulse
[31:28] trading. So it's no
[31:29] coincidence that when
[31:30] momentum plummeted,
[31:32] Bitcoin gained
[31:33] demand. So for those
[31:34] who want to play for
[31:34] the decline: well, it's just
[31:36] closing short
[31:36] positions. All year-long rallies
[31:39] begin with
[31:39] closing short
[31:40] positions. Each of them,
[31:42] because it means that
[31:43] the narrative that existed
[31:44] before trading on
[31:45] AI infrastructure is
[31:46] no longer as strong
[31:48] as it will be, and it
[31:48] will take time for
[31:49] this
[31:50] re-crediting. And
[31:51] as we go
[31:52] through the
[31:53] earnings reports, as I said,
[31:54] I think you're going to have some
[31:55] disappointments, even though
[31:57] every name used to work. I don't
[31:58] think this will
[31:59] continue to be the case in
[32:00] AI trading. I also don't think
[32:01] you'll be able to
[32:01] get
[32:02] five or six times the profits
[32:03] over the next
[32:04] year. I think a lot of
[32:05] these things are already well
[32:06] appreciated. Dylan Patel
[32:07] talked about this
[32:08] today and said that
[32:09] you're going to see surprises
[32:10] on the CPU, and the surprise
[32:11] is that we went
[32:12] a little too far
[32:13] on the optics. We went
[32:14] a little too far.
[32:15] You will...he
[32:16] talked about it. I think
[32:18] all of this means that this
[32:19] sector is valued in such a way
[32:20] that if you can
[32:22] get 40% over the
[32:23] next year, that's
[32:24] great. I think
[32:25] Bitcoin could yield
[32:26] more than 40%,
[32:27] given the fact
[32:28] that it is about doubling to return to
[32:29] historical
[32:30] highs
[32:31] . And a
[32:32] return to
[32:32] historical
[32:33] highs, in my opinion
[32:33] , will leave us with the
[32:34] narrative I
[32:35] just talked about. So, the
[32:36] first stage is
[32:37] closing short
[32:38] positions. The second stage
[32:39] is a change in momentum once
[32:40] we
[32:40] rise above the 200-
[32:41] day moving
[32:42] average, which, as I
[32:43] said, is a critical
[32:44] line. We are still
[32:45] well below it
[32:46] now. We're still, I
[32:47] think, somewhere around
[32:49] 76-77 thousand. So
[32:50] we still have a long way to go. We
[32:51] need 20% before
[32:52] we get there.
[32:54] So, this will be a rally on
[32:54] closing short
[32:55] positions. Let's see
[32:56] how it goes. And like I
[32:57] said, for those
[32:59] who have heard this before and are still
[33:00] asking me: are
[33:01] we still following
[33:02] Doge? Yes, I still follow
[33:04] Doge for the
[33:05] energy side of
[33:06] crypto. But I really
[33:09] think there will be some
[33:10] strange news about
[33:11] what you said
[33:12] over the course of the year, when
[33:13] people start
[33:14] to realize that the
[33:15] 40-name index that
[33:17] I created, covering
[33:18] eight separate
[33:19] verticals or
[33:20] sectors, is a direct
[33:21] overlay on
[33:22] Bitcoin, that Bitcoin
[33:23] just represents
[33:24] the ecosystem, and I think the
[33:26] ecosystem benefits
[33:27] from stablecoins, from the
[33:29] agent side, and
[33:30] that will also flow into
[33:31] Bitcoin. Michael Saylor
[33:34] . The first
[33:36] Bitcoin sale was like,
[33:37] you know, an easy entry into the
[33:38] water. The second was
[33:39] like jumping with a
[33:40] bomb into the depths.
[33:42] The first time it
[33:43] sold out.
[33:44] The second time he
[33:45] sold a little, but the
[33:46] market seemed to
[33:47] not care. And they really didn't care
[33:49] . So, you know, maybe we
[33:50] should give him
[33:51] credit for how he
[33:52] did it. Uh, and the impact
[33:54] on the market from this
[33:55] big sale is
[33:56] minimal. But are
[33:57] you paying attention to
[33:59] this sale or the
[34:00] market reaction? In a
[34:03] year, it will all be
[34:04] just noise. I'm
[34:06] even a little ashamed
[34:07] that I'm being associated
[34:08] with something we
[34:09] even have to
[34:10] talk about in such a
[34:11] meaningless way. Me,
[34:14] the story with Michael
[34:15] Saylor for me—
[34:16] people even
[34:17] blame him for
[34:18] Bitcoin not
[34:19] going higher. In
[34:22] Bitcoin, as we have already
[34:23] said, there are corrections in every
[34:24] asset
[34:25] . I don't believe in
[34:27] four-year cycles. I'm
[34:29] not interested in all this
[34:31] , but the reality is what
[34:32] he did and what he was
[34:33] able to achieve. I'm
[34:36] more focused
[34:37] on the quantum side of
[34:39] crypto and
[34:40] emerging solutions, and
[34:41] topics like, "Should
[34:42] we even be
[34:43] working with these
[34:44] dormant tokens
[34:45] ?" and "Is it worth it?" This
[34:49] person owns a
[34:50] huge amount of
[34:52] asset that I
[34:53] think the world would recognize as
[34:54] real collateral. And
[34:57] as we move into
[34:58] tokenization and
[34:59] get into the topic of
[35:00] dormant assets,
[35:01] I think people
[35:02] will start to understand
[35:03] the value of collateral again, like what
[35:04] he
[35:05] always talked about. So I
[35:07] never got around to it.
[35:08] I agree that the
[35:09] most important thing is that
[35:10] the rate is actually
[35:11] higher than it was during
[35:13] its biggest
[35:14] sell-off. This would be a
[35:15] sign that if
[35:16] you said to people, "Hey
[35:17] , what happens if
[35:19] a week after
[35:20] he makes the
[35:21] biggest sale ever -- the
[35:22] unthinkable, which he
[35:23] promised never to
[35:24] do publicly, although
[35:25] it's for a different reason,
[35:26] so it doesn't count, but
[35:28] he does it -- what
[35:29] happens?" Everyone would
[35:30] say that the price would
[35:31] drop significantly. So
[35:33] the fact that it is higher
[35:35] is, I think,
[35:36] Apple-level news, when
[35:37] the market does not collapse after such an event
[35:38] , and this is
[35:39] more of a positive
[35:41] signal than a
[35:41] negative one. So I don't
[35:43] think this will ever be
[35:44] seriously discussed again
[35:45] .
[35:46] Today's issue
[35:47] was created with the
[35:48] support of Uphold. Are you
[35:49] tired of
[35:49] using a bunch of
[35:50] different apps to
[35:51] trade, earn,
[35:52] stake, or spend
[35:53] your cryptocurrency?
[35:54] Then listen carefully,
[35:55] because Uphold does it all
[35:56] on one single
[35:57] platform. You have
[35:58] access to over 300
[35:59] cryptocurrencies and
[36:00] fiat currencies with an
[36:01] interface that is
[36:03] suitable for both beginners
[36:04] and experts. Uphold
[36:05] also offers
[36:06] instant exchange of
[36:07] anything for anything:
[36:08] cryptocurrency to fiat,
[36:10] fiat to metals, and
[36:11] tokens to tokens
[36:12] right in the app. If
[36:14] you prefer
[36:15] self-
[36:16] storage,
[36:17] Uphold storage provides
[36:18] multi-signature,
[36:19] key recovery,
[36:20] and convenient trading.
[36:22] Uphold has 100% redundancy with no
[36:24] partial
[36:24] redundancy, and
[36:26] proof of redundancy
[36:27] is updated every 30
[36:28] seconds for your
[36:30] peace of mind.
[36:31] US users can
[36:32] earn profits
[36:33] in dollars through a
[36:34] USD savings account.
[36:35] No fees or
[36:36] minimums, and
[36:38] funds are insured
[36:39] for up to
[36:39] $2.5 million
[36:41] through the Atomic Cash
[36:42] Suite program. If you want
[36:43] one app for
[36:44] your entire
[36:44] digital life,
[36:45] try Uphold
[36:46] today. Visit
[36:48] uphold.com to learn
[36:49] more. Uphold.com.
[36:51] Check them out
[36:52] today.
[36:53] Today's issue
[36:54] was created with the
[36:54] support of Mogul. 4
[36:55] trillion dollars.
[36:57] That's how much
[36:57] real estate, according to
[36:58] Deloitte, will move
[36:59] to blockchain over the
[37:00] next 10 years. This is
[37:01] good news for us,
[37:02] because this year the
[37:03] VIX index,
[37:04] Wall Street's fear gauge,
[37:05] has exceeded 30. The
[37:06] S&P has gone negative, and
[37:07] stocks are revalued
[37:08] every time politicians
[37:09] post
[37:10] online.
[37:11] This doesn't happen with real estate.
[37:12] Tokenization is the key
[37:13] to giving
[37:13] everyone access to
[37:14] real estate. Here's what
[37:15] most people don't
[37:16] understand. Thanks to Mogul,
[37:17] you can
[37:18] invest in
[37:18] tokenized residential
[37:19] real estate
[37:20] now. Mogul's founders
[37:21] came from Goldman Sachs, where
[37:22] they invested
[37:23] billions in
[37:23] real estate for
[37:24] institutional
[37:25] clients. Now they
[37:26] choose
[37:27] the objects themselves and
[37:28] manage them for you.
[37:29] Historically, the average
[37:30] annual internal
[37:30] rate of return
[37:31] is 18%—three
[37:32] times better than
[37:33] competitors. You
[37:34] get a target
[37:35] return of over 8%,
[37:36] monthly payments,
[37:37] tax breaks, and
[37:39] no calls from
[37:40] tenants at 3am.
[37:42] More than 100 objects have already been launched on the Mogul platform
[37:43] . Over 30,000
[37:45] people
[37:45] have already registered.
[37:46] If you want to
[37:47] join and
[37:48] get a special
[37:49] offer,
[37:50] go to mogul.club/p.
[37:52] Once again, this is mogul.club/p. You
[37:55] can view
[37:57] important information in
[37:58] the description and follow the
[37:59] link mogul.club/p.
[38:02] When you look at the
[38:04] market right now, we've been
[38:04] talking a lot about
[38:06] AI and about
[38:06] cryptocurrency. Are there
[38:08] any areas that
[38:09] really excite you
[38:10] or that you
[38:11] are starting to invest in
[38:12] , besides AI and
[38:13] cryptocurrencies? Do you
[38:15] think these are
[38:16] supposedly the two main
[38:17] options?
[38:17] So, I study the
[38:18] insurance sector. I am
[38:20] considering the direction of
[38:21] regional banks.
[38:22] I've talked a lot about
[38:23] Eli Lilly, which will continue to be
[38:25] Eli Lilly.
[38:27] I don't think even
[38:29] if it
[38:30] tripled from here, I would
[38:31] sell it. I am a big believer in what is
[38:36] happening in
[38:37] healthcare and
[38:38] drug development, and I
[38:39] want to be associated
[38:41] with a group that has already
[38:42] created sovereign
[38:43] AI. And I want people to
[38:47] hear all the things that
[38:47] others are saying:
[38:49] you should have your
[38:50] own models, you
[38:51] should go through this
[38:52] . Remember what
[38:53] Anthony and I were talking about
[38:55] regarding Lily Pod and the
[38:56] fact that there
[38:57] are a thousand
[38:58] Blackwell chips on their campus? They have their
[39:00] own sovereign
[39:02] AI. They
[39:03] use it
[39:04] in conjunction with
[39:04] biotech
[39:05] companies. So I
[39:06] will continue to
[39:07] argue that
[39:08] healthcare is a
[39:09] big deal. In
[39:10] my video
[39:10] last week, I
[39:11] looked at a few
[39:12] insurance companies
[39:13] that are now starting to
[39:14] see this. I think we're
[39:15] really moving
[39:16] into how
[39:17] important agents are
[39:18] for companies that have already
[39:20] used AI,
[39:21] which means a lot of
[39:22] banks are going to be
[39:23] successful, and you've seen a
[39:24] lot of banks
[39:25] hit
[39:26] all-time highs.
[39:28] Regional banks
[39:29] have renewed their
[39:29] historical highs.
[39:31] Regional banks
[39:32] fit into a very
[39:34] interesting situation: it is very
[39:35] difficult to buy a bank and
[39:36] take it
[39:37] private because of all the
[39:38] regulatory
[39:39] restrictions, which means
[39:40] that takeovers of
[39:41] mid-sized banks are
[39:42] likely
[39:43] to happen,
[39:44] because
[39:45] by implementing AI into
[39:46] their models, companies
[39:47] can buy other
[39:48] firms, integrate
[39:50] AI into them, and reduce
[39:51] staff, which immediately
[39:52] becomes very profitable.
[39:54] So I think we're
[39:55] going to see a lot of
[39:56] consolidation in the
[39:57] financial sector,
[39:58] and the
[39:59] healthcare and
[40:00] insurance sectors are already
[40:00] spending money on this
[40:01] and have already
[40:02] moved forward;
[40:03] they will start seeing
[40:04] results within the
[40:05] next year.
[40:06] What about markets outside of the
[40:07] public? Are
[40:09] there any products or
[40:10] anything else that
[40:12] you're starting to get excited about?
[40:14] Well, if I'm right about
[40:15] Bitcoin, silver will perform
[40:16] well
[40:17] , as will
[40:18] gold. This won't be a
[40:21] resurgence of
[40:22] currency devaluation trading,
[40:25] but it will be in the spirit of:
[40:27] hey, these things
[40:28] are more affected
[40:29] than we thought because of
[40:30] AI sucking
[40:31] dollars out of everything. And
[40:33] secondly, the reality is that
[40:34] we
[40:35] have
[40:36] rate hikes planned.
[40:38] So the idea of ​​"warming up
[40:39] " turned into: "
[40:41] hey, we
[40:42] don't need this at all.
[40:43] " I wrote about this. I
[40:45] talked about it. We
[40:47] still have a huge
[40:48] budget deficit. And this is
[40:51] important because I don't think the
[40:52] Fed will be able to
[40:53] raise rates.
[40:54] So, another angle is they are
[40:56] raising
[40:57] rates in July. This
[40:59] could be a
[41:00] Michael Saylor-style moment
[41:02] because in my opinion,
[41:03] if they raise
[41:04] rates, it
[41:05] would definitely be a one-time
[41:06] move for the sake of credibility
[41:08] in the Fed. Long-term
[41:09] rates would fall, meaning
[41:10] yields
[41:11] would fall, and that would be the
[41:12] reason they are
[41:13] willing to do this.
[41:14] Every macroeconomist
[41:15] I've talked to
[41:16] and had good discussions with
[41:17] on this topic believes this
[41:18] . And me too. That if
[41:20] they unexpectedly
[41:21] tightened policy in
[41:22] July to boost
[41:23] confidence,
[41:24] long-term rates would
[41:25] fall. I think at that
[41:28] moment people
[41:29] would perceive it as a
[41:30] one-time event, but by the
[41:31] end of the year we
[41:32] still have a 50% chance of an
[41:34] increase. So it
[41:35] could
[41:36] even work as a positive
[41:37] if it's a "
[41:38] dovey" raise
[41:39] , where he says: we
[41:40] did it for the sake of
[41:41] trust. We are committed to
[41:42] this, but we believe
[41:43] that now we are in a
[41:44] situation where we can
[41:45] pause and
[41:46] wait. If that
[41:47] happens—and I think
[41:48] if they
[41:49] do, it's in
[41:51] this way—I think
[41:52] that would also be a
[41:53] positive. So I
[41:55] actually think that
[41:56] gold, silver, and
[41:57] bitcoin, this group of
[41:58] assets to protect
[41:59] against
[42:00] money depreciation, will start
[42:01] to show a bottom and a
[42:02] rise. Sometimes people
[42:06] listen to this and they have to
[42:07] pick up on
[42:08] cues from what
[42:10] we're talking about. But
[42:11] there are times when
[42:12] they should pick up on
[42:14] signals from what
[42:15] we don't talk about. And
[42:16] the war in Iran has
[42:17] subsided, then
[42:17] flared up,
[42:19] subsided again, flared up again,
[42:20] and now, it seems, has really
[42:21] subsided
[42:21] . No,
[42:23] not anymore. And last week, it
[42:25] seems like we had a "
[42:26] whiplash,"
[42:27] maybe everything
[42:28] changed twice. I don't
[42:31] know, it sounds
[42:32] crazy, but I don't
[42:34] think the
[42:34] investment
[42:35] community is
[42:36] worried about the impact of the
[42:37] Iran war on their
[42:38] investments anymore. Of course,
[42:41] the situation may be
[42:42] different for those
[42:43] living in Iran, or
[42:45] for the military and
[42:46] geopolitics, but as an
[42:47] investor, I haven't heard
[42:48] anyone talk about the
[42:49] impact of Iran on their
[42:50] portfolio in the last
[42:52] two weeks. Maybe you've
[42:54] heard, what's your opinion on
[42:55] this? Um, no.
[42:58] Nobody really
[42:59] talked about it. And I, I do
[43:03] n't think it's, uh, I
[43:04] don't think people are
[43:05] just unaware of
[43:07] it or aren't, uh, I don't
[43:08] know, scared that
[43:10] it's started. And I
[43:12] think the reality is that
[43:14] oil is not
[43:15] moving. And even
[43:16] though we had, you know,
[43:18] we, we took the
[43:19] leap. I, if, if,
[43:22] oil, and that's the beauty of the
[43:23] markets, we've heard the
[43:25] pessimists say
[43:26] it should hit 200,
[43:28] let's throw that out there
[43:29] and, hey, it should
[43:30] hit 150. The problem
[43:34] is, despite everything, and
[43:36] this is not just the front of the
[43:37] curve, this is the
[43:38] back of the curve. There are
[43:42] no problems. What
[43:44] I take away from this is
[43:45] simple and clear.
[43:47] We should have expected a
[43:48] rapid increase in
[43:49] gasoline prices. Now,
[43:50] again, the
[43:51] processing spread is large. And
[43:53] the reason why
[43:54] refining spreads are large, and
[43:55] for people to understand what it is
[43:56] , is the difference
[43:57] between the price of gasoline and
[43:59] the price of oil.
[44:00] The conversion of oil into
[44:01] gasoline remains
[44:02] high. So,
[44:04] the ability to get
[44:06] gasoline is a problem
[44:07] that still exists, but
[44:08] oil overall is not
[44:09] going up, and I think
[44:11] everyone sees that as a
[44:12] short-term
[44:13] phenomenon. They don't
[44:14] think it will ever
[44:15] end. Uh, meaning
[44:16] it's a
[44:18] flash. I think
[44:19] everyone has already built this into
[44:20] their expectations. So
[44:22] I'll tell everyone here: unless
[44:23] we see some weird
[44:24] moves in oil where it
[44:26] really goes up,
[44:27] it still acts in the
[44:28] market as something that has a
[44:29] positive-negative
[44:31] effect. And I'm sorry to
[44:32] say this, but it
[44:33] looks like there's an
[44:34] excess. It
[44:35] literally looks
[44:36] like when this is
[44:37] over the price
[44:38] will go back to 50. So
[44:39] I'm not saying that's
[44:40] going to happen. I'm just
[44:42] saying that the way
[44:43] the market looks and how it's
[44:44] acting doesn't look like
[44:45] something that's going to
[44:46] grow significantly.
[44:47] So I think we're
[44:49] stuck between 65 and 80. And
[44:51] until we
[44:51] get above 80 and
[44:53] the situation with Iran gets
[44:55] so bad that no
[44:56] ships
[44:56] are going through
[44:57] Hormuz, I just don't
[44:58] think it matters
[44:59] to the markets.
[45:00] The last thing I want to
[45:02] talk about for a few
[45:03] minutes is, uh, I don't
[45:04] know if you've seen the
[45:06] 1X roboruck demo.
[45:07] No.
[45:08] So, uh, maybe the
[45:09] best way
[45:10] to describe it is—the arm has been the
[45:11] most complex
[45:12] part
[45:12] of robotics
[45:13] for a very
[45:14] long time. For 30, 40
[45:15] years, people have been working
[45:16] on this, and most
[45:17] people
[45:18] have essentially used
[45:19] robotic
[45:20] components to
[45:20] create the hand. But
[45:22] if you think about
[45:23] your hand, it
[45:24] runs entirely on
[45:25] tendons. So, 1X
[45:27] has essentially created a
[45:28] synthetic version of a
[45:31] tendon-based robotic arm. And
[45:33] in this demonstration,
[45:34] you know, they can
[45:36] do some pretty
[45:37] nimble, you know, things
[45:38] or very delicate
[45:39] actions. But more importantly,
[45:41] she can screw in a
[45:42] light bulb. She can
[45:44] pick up a cherry without
[45:45] crushing it, or
[45:46] do things
[45:47] that I think many
[45:48] people who saw this
[45:49] demonstration—the video
[45:50] has had millions of
[45:51] views
[45:52] online—said, "
[45:53] Wow, wait
[45:54] a minute." This is a significant
[45:56] step forward. My
[46:00] conclusion after what I
[46:01] saw is that
[46:02] everything that
[46:03] fascinates you and me in
[46:04] the context of AI—I would be
[46:06] shocked if this
[46:07] robotic arm
[46:08] was not designed or
[46:09] created
[46:10] using AI. So
[46:14] now we're getting,
[46:15] you know, a second
[46:15] wave in this field of
[46:17] robotics, where
[46:18] people are using
[46:19] super-intelligent
[46:21] software to
[46:21] achieve these kinds of
[46:22] successes. And it would
[46:23] n't surprise me at all
[46:24] if we start
[46:25] seeing some pretty
[46:26] crazy things where
[46:27] these robots become
[46:28] as
[46:28] humanoid as
[46:29] we all perhaps
[46:30] feared they
[46:31] might eventually become.
[46:33] Mhm. So, I'm going to
[46:35] ...And this is a Finnish
[46:36] company, right? Or a
[46:38] Norwegian company.
[46:41] I think it was
[46:42] created in the
[46:42] United States.
[46:44] But it seems the
[46:45] founders may
[46:46] be from there. I can't
[46:47] remember. There is something that
[46:48] points to a Nordic
[46:49] connection. But I don't
[46:50] think the company is
[46:51] based there. I think
[46:52] it's actually an
[46:53] American
[46:54] company, but
[46:54] maybe the founders are
[46:55] from there.
[46:56] So I agree with what
[46:57] you said, and
[46:58] let people hear it
[46:59] . I'll start with the statement
[47:02] , and then I'll backtrack a little bit
[47:03] to
[47:05] why ROI for
[47:06] hyperscalers is
[47:07] not the same as fracking
[47:08] oil in this
[47:09] context. So, first of
[47:13] all, Anoni's opinion is
[47:15] very important, and I
[47:16] agree with her. We
[47:18] underestimate the fact that
[47:19] when we reach 100,
[47:21] we are already above 140 IQ. When
[47:23] we cross the
[47:24] 160 and 200 mark, we
[47:25] will be solving
[47:26] problems that we
[47:28] could not solve before. We will
[47:29] find approaches that
[47:31] will allow us
[47:31] to produce these things. And
[47:33] when
[47:34] humanoid robots appear
[47:36] , we will no longer
[47:36] have to
[47:37] do the physical
[47:38] work of
[47:39] testing them. They
[47:40] can build them
[47:42] for us. The hand is therefore
[47:43] crucial,
[47:44] as it is perhaps the
[47:45] most complex
[47:46] tool for
[47:47] performing the delicate
[47:48] work required for
[47:49] robots to
[47:50] build robots
[47:51] or solve these
[47:52] problems. So, if
[47:54] the intelligence is already there, we
[47:55] need hands to
[47:56] bring it to life. It
[48:00] was during Moonshots where
[48:01] Dave Blondon
[48:02] actually admitted that he
[48:03] underestimated the speed at which
[48:05] we would get
[48:06] billions of robots
[48:07] because he saw
[48:08] in Tesla, "Oh my god, robots are
[48:09] going to
[48:10] build robots someday." This is the
[48:13] aspect of accumulation:
[48:15] Opus 4.5 leads to 4.6, which
[48:16] leads to 4.7, and this whole
[48:18] process of recursive
[48:20] self-improvement.
[48:21] Now I want to
[48:22] return to the topic of
[48:23] oil. This week
[48:26] I was asked if the
[48:28] ROIC risk for
[48:29] hyperscalers is the same as
[48:30] it was with
[48:31] fracking, and many
[48:33] people on VEX have written about
[48:34] this. They really
[48:35] scared some of my
[48:37] followers. They
[48:38] were addressing me.
[48:39] Here's the reality,
[48:40] guys. Intelligence is
[48:41] not the same as oil. I do
[48:43] n't even know how to
[48:44] explain it more clearly.
[48:45] Oil is oil.
[48:47] Intelligence is everything. This
[48:50] means that the price of
[48:51] tokens is falling and
[48:52] usage is increasing
[48:54] . Oil use
[48:56] increases only linearly
[48:57] with
[48:58] nominal GDP.
[49:00] There are physical
[49:01] limitations to its
[49:02] extraction. The problem is
[49:05] that for
[49:05] intelligence,
[49:06] the usage is growing
[49:07] exponentially.
[49:09] Demand for oil never
[49:10] grows
[49:11] exponentially. The
[49:13] only thing slowing down
[49:14] AI is data centers.
[49:16] These are gas turbines. These are
[49:18] physical things that
[49:19] cannot be produced
[49:20] quickly. So, to
[49:22] Anon's point, and for
[49:23] all of you who don't
[49:24] understand how
[49:25] hoarding will help
[49:26] solve the energy
[49:27] problem, and why I
[49:29] always keep an eye on
[49:30] oil: eventually, the price
[49:31] of it will drop to zero.
[49:32] It will drop to zero
[49:33] because we will solve the
[49:34] energy problem through
[49:35] fission or
[49:36] fusion. Call it anything you
[49:38] want. We will
[49:39] offer many
[49:40] solutions in the future.
[49:42] So when you listen to
[49:43] Dario, Sam Altman,
[49:44] or Greg Brockman,
[49:46] they all say the same
[49:47] thing: intelligence
[49:48] leads us to the point where
[49:49] mathematical problems
[49:51] are solved. Once we
[49:52] solve the math,
[49:53] we can beat
[49:54] cancer. Then we can
[49:55] solve the
[49:56] energy issue. These are all
[49:57] scientific problems. These are
[49:59] all physical problems
[50:00] that lead us to the
[50:01] point where everything
[50:02] starts to accumulate
[50:03] . So just as we
[50:05] talk about the
[50:05] development of intelligence
[50:07] and models for
[50:07] coding, i.e. language
[50:09] , we will soon
[50:10] see this in
[50:10] consumer agents as well as
[50:12] in science. And this is where
[50:13] robotics and
[50:14] humanoids become
[50:15] so important to
[50:17] your investment
[50:18] decisions, and this is why
[50:19] financial
[50:20] safeguards are needed to
[50:21] cope with such
[50:22] speed.
[50:23] I think we
[50:24] can end this here.
[50:25] What will you
[50:26] talk about in
[50:27] your YouTube video
[50:28] this Sunday? I'm
[50:29] going to emphasize
[50:30] , first, that the
[50:32] slowdown in the AM cycle
[50:33] is over, given
[50:34] that sentiment has
[50:35] fallen significantly and
[50:36] volatility has increased
[50:38] enough that we
[50:39] are at the bottom of the
[50:40] consolidation. And
[50:43] now we should think not
[50:44] about “will the price go
[50:46] lower,” but about the AI ​​trend
[50:47] — which I couldn’t
[50:48] say when Micron was
[50:50] at 1300, but at 950—will
[50:51] it be higher in 6
[50:53] months? Yes, it will be.
[50:55] Once we go through two
[50:56] reporting cycles, it
[50:58] will be higher. So,
[50:59] looking ahead half a year
[51:00] : you
[51:00] are at the bottom of
[51:02] the range,
[51:03] not the top. The second is
[51:04] how important
[51:06] consumer agents are to computation. For
[51:07] all of you who are afraid of
[51:09] calculations and think there is a
[51:10] possibility of their
[51:11] excess. If you
[51:12] want to criticize
[51:13] me in the future, I will
[51:15] say it again.
[51:16] We have an insatiable
[51:17] demand for
[51:17] computing
[51:18] power. We don't have
[51:19] enough
[51:20] supply. And the
[51:21] argument I
[51:21] just made: the demand for
[51:23] computation is
[51:23] exponential,
[51:24] as
[51:25] the use of
[51:25] intelligence
[51:26] increases dramatically with each
[51:27] improvement in
[51:28] the model. But physical
[51:29] deliveries are what's holding
[51:31] us back. So it's an
[51:32] imbalance of supply and
[51:33] demand, and I'll
[51:34] explain in detail
[51:35] why it's worth
[51:35] investing
[51:36] there. Third, I will
[51:38] elaborate
[51:39] on Bessin's speech and
[51:40] re-emphasize the
[51:42] importance of Bitcoin,
[51:43] as I believe that
[51:44] in a year we will be
[51:45] talking about crypto in the same way we are talking
[51:46] about
[51:48] memory now, and I will conclude with what
[51:49] you said. You
[51:51] asked me if the
[51:53] fundamental
[51:54] players are starting...No, they're
[51:55] not focused on that right now
[51:56] . Firstly
[51:57] , because they are still
[51:58] focused on AI, both in
[51:59] terms of losses
[52:00] and what has happened in the
[52:01] last four weeks,
[52:02] and in terms of "what should
[52:03] we buy at this
[52:05] level?". Once they
[52:06] realize this,
[52:07] get through the
[52:07] Fed decision and all
[52:08] the changes, they will have to
[52:10] focus on
[52:10] crypto and the financial
[52:12] safeguards of
[52:12] the future. And in 6
[52:14] months,
[52:14] approaching
[52:15] next year,
[52:16] the crypt will become what
[52:18] memory was to me,
[52:19] what captivated me so much.
[52:21] Journey, I will
[52:22] watch; Let's talk
[52:23] again next
[52:24] week, okay?