YT YouTube2Text

YouTube2Text → Transcripts → Bitcoin Is the Toll Collector of the AI Economy

Bitcoin Is the Toll Collector of the AI Economy — Transcript

by Anthony Pompliano · 9,210 words · 1,329 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00when it starts to happen, there's an AI
0:02theme behind it. I want to make sure
0:04people realize when it comes to crypto,
0:06this is 100% for me about agents. And
0:11I'm telling you, the amount of
0:12traditional people that continue to bang
0:15on my door to talk about tokenization,
0:18to talk about how to make money on this.
0:19They're doing their homework now. I
0:21think next year is going to be the year
0:23that we see the biggest moves in crypto
0:25because of it. What's going on, guys?
0:27Today, we got a great conversation with
0:28Jordy Visser. In this conversation, we
0:30talk about what's going on with AI
0:31agents, how crypto and Wall Street are
0:34coming together. Then we talk about the
0:35idea of this going from a contrarian to
0:37a consensus trade. And then Jordy and I,
0:38we debate a little bit as to general
0:40purpose models versus specialized
0:41workflows. Where is value going to acrue
0:44in the AI industry? And then we get into
0:46the bubble talk. All these people
0:47worried about a big bubble. And Jordy
0:49helps us separate both the signal from
0:51the noise. Here's my latest conversation
0:53with Jordy Visser. All right, Jordy.
0:55This week there's a Bloomberg piece
0:57about you. The Bloomberg piece said,
0:59"Forget the AI bubble." Jordie Vesser
1:00says, "The bank apocalypse has begun."
1:03Congratulations on the uh Bloomberg
1:06coverage. You're a big deal now. Um but
1:08>> they waited for me to say something
1:09negative.
1:10>> Negative. They didn't say anything
1:10positive.
1:11>> No, no, no. The media loves a bearish
1:13thing.
1:14>> Well, bank apocalypse will definitely
1:16get clicks. Uh what is the bank
1:18apocalypse and why do you think it is
1:20coming?
1:22>> So, I posted an X that morning. I didn't
1:24see the article until someone pointed it
1:26out to me. Uh we've talked about it
1:29before, but I really did a lot of it on
1:31my my video last weekend. So rather than
1:33turn this into some like the banks are
1:35all going to go to zero thing, the
1:37software companies went through their
1:39SAS apocalypse. So number one thing,
1:40it's a play on the SAS apocalypse thing.
1:43SAS got hurt for two reasons or two two
1:48stories behind it. One is the
1:50recognition from people that AI was
1:53going to disrupt seatbased SAS software.
1:56Not all of software. Cyber's done well.
1:58Bitcoin miners have done well. They're
1:59in there as well. Microsoft's been fine.
2:02But the reality was AI hurt these
2:05companies despite the fact that they
2:07continue to print for the most part
2:09record earnings.
2:11And that was a sign if you remember in
2:13that first quarter, one of the most
2:14important signals for everyone to think
2:15about when it comes to sectors is that
2:18because of exponential innovation and AI
2:20continuing to increase progress, at some
2:23point it will be a negative for almost
2:26every single industry and especially
2:28legacy industries. And I just want to
2:30make sure because I tried to I tried to
2:32show that in the expost and I'll use
2:33some of that here. I believe the banks
2:36now are going to go through a problem
2:39which is they're going to come under
2:42multiple compression over the course of
2:44the next year because of two factors.
2:47One is the rise of AI agents which means
2:49consumers are going from let's say lazy
2:54individual small businesses have their
2:56money in banks that pay them zero yield.
2:59That's really good for the banks when
3:00the banks can take your deposits of
3:02which there's a lot pay you zero
3:05take that money invested in treasuries
3:08yielding 4% short-term not like SVB did
3:12long-term but shortterm uh you can get a
3:16spread on this so we saw the article
3:19from Torston Sllock with an agentic bank
3:22run so he was going a different angle
3:25but that's part of it and I don't want
3:26to say that there's going to be a run
3:27because that's what's going to happen.
3:28What I want to say is that agents are
3:30not going to make mistakes. They are
3:32going to search for the cheapest way to
3:34do thing and that means that as agents
3:36control more of the money decisions,
3:39banks are going to be in trouble. The
3:40second thing is that the guardrails are
3:42going to apply pressure as well. And
3:43this is where crypto fits in. So when
3:45you have the users, the consumers no
3:47longer being humans but being digital
3:49agents and at the same time you have
3:50financial guardrails which are being in
3:52the system more where there's same day
3:53settlement which means stable coins
3:55which means tokenization. We're headed
3:56towards that settlement. Long-term
3:59settlement is a good thing for banks,
4:01for Visa, for Mastercard. The longer it
4:03takes for some, somebody gets that
4:04float, somebody gets the chance to
4:07invest that money for a day, for 3 days,
4:09and if an instant settlement, it's not
4:10the same. So, this was really a a belief
4:13that banks will still make money. Um,
4:16but the spread, the edge that they had,
4:18their multiples over the future are
4:22going to be hit by terminal value again.
4:24So, it's not a like doom and gloom
4:25thing. I kind of played with the SAS
4:27apocalypse to create bank apocalypse.
4:29But that is what in in my opinion is
4:31certainly going to happen with the rise
4:33of consumer agents.
4:34>> Is there a positive to this? Like the
4:36consumer wins in this scenario?
4:41>> Yeah, the consumer wins. The people who
4:43um invest in the financial guardrails
4:45are going to see more traffic. So if you
4:47invest in crypto and it's a small asset
4:49class, you'll do well. The the only
4:51place I see getting hurt is the
4:52middlemen. And so labor verse capital
4:55has been an issue. The consumers, we
4:57have a K-shaped economy. Anything that
4:59is going to make it more powerful to
5:00bring, you know, inflation down, at
5:02least on the service side, uh, is going
5:05to be positive in the long run. So your
5:07question, everyone should be thinking
5:08about it. What's the positive side?
5:10Well, the positive side is if we can
5:12reduce the fees, you'd be shocked at how
5:15much money the banks take purely on a
5:17spread basis from everyone. Well, the
5:19reason I say that is uh we never see the
5:21headlines of the positive news. Like uh
5:25one of the chapters in the book I just
5:26wrote actually I don't know if you know
5:27this is uh uh planes that land safely
5:30never make the news right and so like
5:33same thing here is when people see
5:35disruption occurring they always cover
5:37it from the angle of who is being
5:39disrupted. They never are like consumers
5:43will make X billion dollars more because
5:46they now will have their money not
5:47parked in you know low yielding assets
5:50or whatever. So, the perspective I think
5:53matters, but it also plays into I don't
5:56know if you saw Paul Graham uh this
5:58week, uh Amazon made the decision to
6:00stop I think it was Muse and some other
6:02bots or agents from accessing Amazon.
6:05And he came out and I thought it was
6:06probably one of the most interesting
6:08perspectives uh this week was the first
6:11time in decades that he believes someone
6:14could expose or exploit a wedge to
6:18disrupt Amazon. and his point was if
6:20somebody could build an Amazon that was
6:22made for the agents
6:24>> that could be very disruptive. Now,
6:26Shopify, I think, is leaning into, hey,
6:28how do your agents come and buy things
6:30and do all this, but that feels like the
6:33banks are the same thing. Like, if
6:34somebody builds a bank that is
6:36agentfriendly, it's probably going to be
6:38pretty valuable, right? Well, let's
6:40separate the bank, meaning where you
6:43place your money. If they pay you the
6:46fair spread, meaning if you deposit
6:48money in a bank and they pay you and all
6:50you're doing is I'm getting interest on
6:52that money, then there's the other
6:54service side. They call you up. We want
6:56you to invest. You know, they all own
6:58brokerage houses. They all own they want
6:59they want to make sure they're making
7:00money. This is this is the service side
7:03of the business. Stripe is not a bank,
7:06but Stripe is competing with all of
7:09them. And that's what I included in the
7:11piece. This weekend I will go through
7:12about 15 slides of Stripe. City Bank is
7:15a $200 billion market cap company with
7:20$200,000 people.
7:23Stripe has 10,000 people. It's about $
7:26160 billion currently off the latest
7:29valuations. If you go through their M&A
7:32activity, if you go through their
7:34volumes, their revenue is up 41%
7:36year-over-year according to them. You go
7:38through every metric in their revenue
7:40side. They have already built for the
7:42crypto world. They already are set up
7:44for this. They're already set up for the
7:46agentic side. They've been talking about
7:47agentic commerce for a long time. These
7:50older banks, of which I worked for one,
7:53it's very hard to get through the
7:54bureaucracy and make a decision. They've
7:57always had the ability and this is one
7:58of the things that I think people should
8:00just really listen to when it comes to
8:02something that I feel by using AI and
8:04why I try to help people use it as much
8:06as you can. The progress is insanely
8:09fast. As you and I both know, these
8:12businesses never move fast into
8:14something. First of all, there's a
8:15calendar year, the summertime, nothing
8:17gets done. And then you come back in
8:19September and they start focusing on
8:20when they're going to pay people for
8:21bonuses. And that process, trust me,
8:24it's a while. Then they've got to do 360
8:26uh reviews. Then they have to hand out
8:29the bonuses. They have to sit and put a
8:31budget in for the next year. This is all
8:34time. AI doesn't wait for that. Stripe
8:37doesn't wait for that. They don't have
8:39the amount of people that Croup does.
8:40They don't have to bring in the managers
8:42from overseas. They don't it's a it's a
8:44completely different side. So I think
8:46the nimble companies and what Paul said
8:49about Amazon, if I remember exactly what
8:51it said, he basically said never in
8:53their history have they been in have had
8:55to do this. The reason I think this is
8:58important for people to think about when
8:59it comes to the hyperscalers, I say all
9:02the time, you have to look at how these
9:04Mag not the Mag 7, but let's well let's
9:06say the Mag 7 became so big. They don't
9:10compete with each other for the most
9:11part. Apple, Apple has its own business
9:14model. Amazon has its own business
9:16model. Google has its own. AI agents cut
9:19through all of them. They all want
9:21consumer agents. And if they start
9:23playing this game of we're not allowing
9:25Muse to come in because we're waiting on
9:27Alexa. And then Siri does the same
9:29thing. Apple also announced they're
9:31changing the iOS and they said it was
9:34because they didn't want agents
9:36penetrating. So, they use the scare
9:38tactic as opposed to no, we don't want
9:40agents to get into this. I would love it
9:42right now if my AI agent in Muse could
9:44get my Apple messages. It would make
9:47make it a lot easier. I think they're
9:49all now in trouble that they're all
9:51basically trying to defend their turf.
9:54So, they're playing for the same TAM
9:56now. And that means that their business
9:58to me is being disrupted. And I think
10:00Amazon, what Paul Graham wrote is just a
10:02sign that they don't know what to do
10:04because these companies are big
10:06bureaucracies. Now,
10:08>> uh Elon Musk tweeted uh two or three
10:10days ago saying that Grock was going to
10:12open up the backend to any model. And
10:16for those that are unfamiliar, um, if
10:18you use OpenAI, they will give you
10:22either a model that you select or if you
10:24are just talking to ChatgBT, they may
10:26route your query to an OpenAI family of
10:29models. They would never send your query
10:33on ChatgBT to Anthropic, Claude, etc.
10:37Grock historically has used their
10:40internal stuff, but now they're opening
10:42it up and they're saying, "Well, we're
10:43going to actually just route to whoever
10:44is the best model." To answer your
10:46question,
10:48if that forces the hand of the other
10:50players, to me, that signals that the
10:53models are commoditized and now the
10:56value of those models becomes much more
10:59in question than they already are. I
11:01personally think that's where we're
11:02headed. like we bet Sylvia on this of
11:05building a model router and saying hey
11:07we don't really care what the model is
11:09we just want it to be the best one to
11:10answer uh for the uh end user now Elon's
11:14doing it does that force open AI and
11:16anthropic both to open up themselves as
11:19well or do you think that they can keep
11:20these kind of closed you know kingdoms
11:22and uh similar to how Amazon seems to be
11:25kind of creating a closed kingdom like
11:26we we're playing this game of like
11:28bundling and unbundling that is like the
11:30story of you know business and finance
11:32for decades
11:33So, this is the one place where you fit
11:36in with the masses. So, when we when we
11:39talk like I disagree with where you're
11:41headed with this,
11:42>> okay,
11:43>> um, and I I've heard this repeatedly
11:45from people over the course of the next
11:46six months. This is not exactly what
11:47you're saying, but it's pretty close.
11:51>> Open AAI and so ChatGpt and Claude are
11:54the best models, plain and simple. Like,
11:56I don't care what anyone says to me as
11:58someone who uses all of them and has
11:59used open source models. The best model
12:02right now is Opus 5.5. The next best
12:04>> for general general purpose stuff.
12:06>> No, no. For the stuff that I do, so that
12:08includes images, video, like a general
12:11model that will do everything that I
12:13need. Not like Gemini makes a lot of
12:17mistakes and it comes back with answers
12:18that are ridiculous. Opus 4 Opus 5.5 is
12:22a game changer. Now, here's the thing.
12:24>> You tell Gemini it's silly.
12:26>> You know, it's the most disrespectful
12:27thing you can say to a man. It's just
12:28you're silly.
12:30I don't understand.
12:32So I I'm maybe it's because I use them
12:35so much that
12:37>> I don't get frustrated by the answers. I
12:40don't have to explain twice to them. And
12:42trust me, you have employees and you
12:44know which ones you might have to say
12:45something twice to.
12:46>> Listen, I today's episode is brought to
12:48you by Arch Public. Arch Public has just
12:51expanded its agentic trading platform
12:52beyond crypto. So pay attention. This is
12:54a big one. Now, they are automating
12:56strategies across stocks, commodities,
12:58and ETFs. And I think that this is going
13:00to be huge. You can now automatically
13:02take profits when one market hits new
13:04all-time highs and rotate that capital
13:05into other markets, showing more
13:07opportunity. Whether you're rotating
13:09capital into AI stocks, gold, if you're
13:11investing in the S&P 500, or you're
13:12accumulating Bitcoin, Arch Public brings
13:14real discipline and automation to your
13:16investment strategy. Additionally,
13:18they've launched a powerful new tax loss
13:20harvesting tool. With crypto being so
13:22volatile and its exemption from the wash
13:24sale rule, Arch Public can offset gains
13:26with losses without compromising your
13:28long-term positions. It's exactly what
13:30every serious investor does.
13:32Institutional grade automation that
13:33works across every major asset class.
13:35There's no more emotional trading, no
13:37more missing tax opportunities, just
13:39smarter, hands-free execution of your
13:41preferred strategies. Go to
13:42archpubic.com
13:44right now. Connect with their team. Set
13:46up a time. Bring your accountant if
13:47you'd like. Then you can learn what
13:49automated trading can do for you.
13:51archpub.com.
13:52>> I'm a very big believer and uh have
13:55always said I think that people
13:57underestimate the importance of their
13:59intuition because it is you know it's
14:01your human algorithm. You've been
14:02training it. Just as there are
14:04benchmarks that are these like objective
14:06open- source benchmarks that all the
14:07models try to you know game or whatever
14:10you have a Jordy benchmark and what
14:13you're describing is that opus is 5.5 is
14:16the leading model in the Jordy
14:18benchmark. Yes. And it's not just for
14:21me.
14:23How many times does Grock get announced
14:25as solving the greatest math problems
14:27ever? How many who's in the fight for
14:29that? Like it's not just me. Who finds
14:32an enzyme for crisper?
14:34Okay, that used to be Google. That's
14:38what Alph Go was. That's what Alpha Fold
14:41was. That's not what's happening
14:42anymore. The stuff that's happening is
14:44in these models. And here's the issue.
14:47when you start realizing how much
14:49compute is needed for consumer agents.
14:52So the Paul Graham thing what what you
14:54didn't see Elon
14:56also posted that and said seriously like
15:01this is we're going to have
15:02commoditization in intelligence. There's
15:04no doubt about it. The difference is and
15:06what people don't realize the amount of
15:08compute you have leads to the models
15:10being more efficient. So, if you haven't
15:12felt this yet,
15:15I'm I'm surprised the cost of the models
15:19for like Opus 5.5, they're getting more
15:21efficient because they have a ton of
15:24compute power behind it. But if it takes
15:26you
15:282 minutes to get to an answer in Opus
15:315.5 and a cheaper open-source model
15:33takes you 12 minutes. When you add up
15:36the token usage and the cost and it
15:38comes through, the time does matter. not
15:40just the answer. And I think everything
15:42in AI is about the time relationship to
15:45the probability of getting the right
15:47answer. So everyone only talks about the
15:49cost side. That's not the way the world
15:51works. We have cheap cars. We have
15:53expensive cars. Like everything in life
15:55becomes that. I just think for solving
15:57cancer, for solving the greatest
15:59problems in the world that people are
16:01going to need to get to space to do
16:03this, we're getting there's going to be
16:04highlevel things that are going to take
16:06these models. And the problem is
16:08Anthropic and OpenAI are going to own
16:10most of the compute going forward. And
16:13if the compute matters that much, which
16:15it will because I still will keep saying
16:17and I hear this on the other side,
16:20people still fight on like there's a
16:22bubble coming like we are so early in
16:24the demand side of tokens. So early the
16:26coding agents didn't even exist a year
16:29ago. So that just started the adoption
16:31that means for corporations we went
16:33through token maxing. Now we're going
16:34through okay we can't throw this much
16:36money at it. We have to be more
16:37efficient with our thoughts. Well,
16:39they're figuring out how to do that. The
16:40models keep progressing and getting
16:41better. And Ver Rubin's going to come
16:43out next year. As Ver Rubin comes for
16:45the same amount of gigawatts, you're
16:47getting more. Well, who owns most of
16:49that? The really smart groups own most
16:52of that. So, that's where OpenAI
16:54Anthropic, but also Elon Musk, he owns a
16:56lot of it, too. He's been buying as much
16:58Ver Rubin as he can. So, I think we're
17:00going to get to a point as you move down
17:02the road that the open source is fine
17:04for the majority of tasks. 90% of them
17:06will be close to zero. But for the 10%
17:10the total TAM of that 10% is going to
17:13continue to grow through adoption. And
17:14that's where Jevans paradox actually
17:16matters because the cost of it and the
17:18efficiency will keep going. And that's
17:19really important for people when they
17:21think about investing because I will
17:22continue to say there are two themes
17:24that I care about right now. Agentic
17:26compute. So the AI infrastructure and
17:29the financial guardrails. That's what I
17:31believe over the next one year will
17:33create the alpha. And if you look back
17:34over the last four years, the alpha has
17:36all been about AI.
17:38>> So I have uh somewhat of a hot take.
17:40>> Good.
17:41>> The big labs already are behind in two
17:44very important ways. What are the
17:47leading personal agents?
17:50Muse and Instinct, not the big labs. I
17:54know they launched DOTs. We'll see what
17:56happens, but Muse and Instinct are out
17:57in front leading in those two.
18:00When you look at specific applications,
18:03legal, finance, etc., the large language
18:07models are not the leading models.
18:09They're leading in general purpose, but
18:10when you start to get specific, they're
18:12not. When you look at it on a cost
18:14perspective, I can just go off of our
18:16data. If I process 500 million tokens
18:19from uh on Claude, about $5,000. If we
18:23process the same 500 million tokens on
18:24our own model and our own hardware, it's
18:2797% cheaper. So
18:30what I think is going to happen, but I
18:32think the fact that you and I don't
18:34agree on this actually is like maybe
18:36that's the story of like nobody really
18:37knows, right? Is you're going to start
18:41to see the large language model say we
18:44got to pick our battles. Where do we go
18:46and fight to win because just having the
18:48general purpose model is actually not
18:50going to be as valuable because of the
18:53commoditization. And so we should
18:56accelerate into specific use cases.
18:58Anthropic recently announced that
18:59they're going to go really heavy into
19:00bio and all the biotech stuff they want
19:03to go do. It feels like they think that
19:05is a massive opportunity. If they can
19:07win there, that is trillions of dollars
19:08of value that they can create. Open AAI
19:11is saying dots and personal assistant.
19:13They've always been more of a consumer
19:15type company than an enterprise type
19:16company. And so I don't think it's just
19:18like, hey, go all in on one specific use
19:19case, but it does feel like they're now
19:22realizing that the general purpose model
19:24can serve as like a foundation, but the
19:26monetization of this and the sustainable
19:28moat is going to be in the applications.
19:32And this goes back to the idea of in
19:34crypto, if you go back to 2017 or so,
19:38there used to be this thing called the
19:38fat protocol thesis. And the whole idea
19:41was like all the value acrrues to the
19:42protocol layer. And I could be wrong on
19:45this, but I just continue to go back to
19:48the fat protocol thesis. The equivalent
19:50in the AI is like the fat model lab
19:52thesis where all the value acrews to the
19:54model lab.
19:56What we found in crypto was there is
19:57definitely some value that acrrews to
19:59the protocol obviously bitcoin,
20:00ethereum, etc. But actually a ton of the
20:02value acrewed to either the
20:04infrastructure underlaying it in terms
20:06of miners and things like that or the
20:08application layer. It feels like that's
20:10the same thing happening here where as
20:12the technology gets built out, we have
20:14tons of infrastructure that's making
20:15money, we have tons of applications that
20:17are yet to be built, but I just find it
20:20really hard if like Anthropic and Open
20:23State as general purpose model
20:24providers. I just think that is like a
20:27cutthroat business. It's like kind of a
20:28race to zero. Who can get the, you know,
20:30highest level intelligence for the
20:31cheapest cost? And so like what's the
20:33moat? What is the the the, you know,
20:36thing that defends against everybody
20:37else? But I don't know, maybe you think
20:39differently.
20:40>> I do. Um
20:42I don't think the application So when
20:44you say Muse and these things, how h I I
20:48I don't understand how Muse is going to
20:50make money.
20:51>> Oh, I think they're going to print cash.
20:53>> Okay. So, I don't understand how they're
20:54going to
20:55>> because again, the they have a couple of
20:59decisions they need to make, but let's
21:01use Instinct as Amuse's uh uh kind of
21:04proxy. What they very clearly are trying
21:07to do is they want to get as much
21:08transaction volume going through that
21:10personal agent as they can and they
21:11think they can basically take a vig and
21:13my guess is that they will go and start
21:14pressuring whether it is the credit card
21:16companies, the payment processors, the
21:18merchants etc. and say, "Hey, I'll route
21:20stuff your way, but you got to give me
21:22something." Right?
21:23>> So, M or uh Instinct is going after a
21:26transaction-based revenue model. He has
21:28said that he would like to keep it free
21:30forever.
21:32>> Facebook, they got one business model.
21:36>> Yeah.
21:36>> And you and their target is about to get
21:38way better, right? And so, whether it is
21:41actually transactions where they get
21:43paid or the advertising is where they
21:44get paid, I worked inside that business.
21:46Trust me, if you're going to use their
21:49products and you're going to give them
21:50more data, they're going to get better
21:52at ads. And so to me, it's not Muse
21:55makes money as much as Facebook makes
21:57money. They just happen to own the Muse
22:00thing. And look at like WhatsApp like
22:02that. They've started to do this across
22:03a lot of their interfaces. Um, and I go
22:06back to when I opened up Instagram and
22:08I've got a interstitial at the top of my
22:10Instagram feed that's like download
22:11Muse. They're going to get people to
22:13download it. They're going to get the
22:15data. then they're going to target you.
22:18>> Did you listen to the Noah Shinar
22:20interview?
22:20>> Yeah.
22:21>> Okay. When you say ads, that's he called
22:24that the attention economy.
22:26>> The attention economy is going to
22:27collapse.
22:29>> Maybe
22:29>> I I I personally I I don't I mean part
22:33of what's going to happen, it is very
22:35clumsy at this point to have to use your
22:38phone for your personal agent.
22:40>> Agreed.
22:41>> Okay. So that's why they're all trying
22:42to figure out the next gadget to go
22:44through. By by the way, uh I think that
22:46Instinct, which we really haven't heard
22:48Meta talk as much about this, but
22:50Instinct, he specifically uh he wants to
22:55monetize the transactions and what he
22:58his belief, which I actually agree with
22:59him on, is the agents transacting should
23:02lead to more transactions, which
23:03therefore it's like a market expanding
23:05technology for both the merchant, him,
23:07everyone kind of wins in that because
23:08the consumer also gets served.
23:11I think in the attention to come you're
23:12talking about like yeah if the agents
23:14are perusing or looking for something
23:17serving a display ad in the feed of
23:20whatever is not going to be very
23:21valuable and so I is there going to be
23:24like agentto agent advertising here
23:28here's the point and this is where we
23:30can end this conversation um I only
23:33believe there's two things to focus your
23:34attention on the toll collectors
23:39>> that's crypto
23:40>> as far as I'm concern. I want the
23:42digital money guardrails, the toll
23:44collectors.
23:45>> I don't want I don't want to invest in
23:47Muse and then watch as seven other
23:51agents are released and that's before we
23:53have agents in Europe and agents in
23:56Asia. Like this is the thing. We have
23:58competition globally.
24:00>> I just want to collection velocity. I
24:03don't want the attention economy. These
24:05are all things to me of the last 17
24:07years. I want to bet on the things that
24:10we are always going to have demand
24:12greater than supply.
24:14>> Mhm.
24:14>> We will not have enough compute for the
24:17agents. That is the way to play what
24:19you're describing. In my opinion, the
24:21competition, it's all going to be
24:23commoditization. The reason I say I
24:25don't agree on the anthropic and open AI
24:26is because I think there's infinite
24:28amounts of money for super intelligence
24:30and they're going to have it. Open
24:32source models can never have the size
24:34intelligence they can because they can't
24:36have the compute. All of these places
24:38they don't have the compute. So they can
24:40keep copying distilling the models but
24:44at some point you need this because we
24:47won't release the best models to the
24:49market. They will go to the businesses
24:52>> and that's where it can't be
24:53commoditized. So at some point people
24:56have to realize if you have the compute
24:58you get the best model. If you have the
25:00best model and it can solve all the math
25:02problems, then every single science and
25:05physical place humanoids will want that
25:07model. You don't release it to the
25:09public. And if you don't release it to
25:10the public, how do you distill all the
25:12information if it's not a public model?
25:14>> Maybe this is like the root of where we
25:16see a different world. I don't think
25:18that uh Claude will be able to create a
25:24model that outperforms a company. Let's
25:28say that you're a huge law firm. Let's
25:30say you're a huge finance firm.
25:32>> If you take open source models and you
25:35customize them via fine-tuning to your
25:38specific architecture, your specific
25:39data, etc., you should be able to out
25:42compete them.
25:45So again, how many business are going to
25:47do that? Do they have the talent? It
25:48cost money. Like there's a lot of
25:49whatifs in there. But if that is the
25:51case, I guess that's where I get a
25:53little bit more less enthusiastic is
25:56just like you can keep iterating on the
25:58model, but it's still a general purpose
25:59model. Yep.
26:00>> The businesses that say, "Hey, we're
26:01going to build custom models for
26:02ourselves."
26:04>> You can show up with the smartest model
26:06you want, but they're like, "But it's
26:07not customized to us." And so the
26:08business is less likely to go do it.
26:10Now,
26:11>> I think to your point, that's a black
26:14and white analysis. What's probably
26:16going to happen is the business like
26:17give us all of it and then we'll figure
26:19out what we use and over time they'll
26:20kind of fall away but they'll try almost
26:22everything. And that's kind of what we
26:23saw. There's a Wall Street Journal
26:25article uh that says uh the title is
26:27spending on AI is becoming almost
26:29impossible for businesses to budget
26:32which is basically this story, right?
26:34>> And it's it's a great one. So I I we'll
26:36leave it at this. I I don't disagree at
26:38all with what you're saying. The models
26:39will not be the models behind a law
26:41firm. Those will be more open source and
26:43cheaper things. The problem is the law
26:46firm will end up being a oneperson law
26:48firm and a billion of them. So if you're
26:51like I hear you talking and and I think
26:54you think differently than I do in the
26:56endgame, which is I don't want to be a
26:58VC fund. Like I don't think you can pick
27:00the winners in the world you're
27:01describing because I believe it will
27:04always be an entrepreneur that says I'm
27:06willing to make less
27:07>> 100%.
27:08>> And I'm going to build a business and I
27:09don't I only need to make a hundred
27:10grand a year. So when people ask me how
27:13to like and this was asked to me
27:15yesterday's like okay I'm I'm I'm in my
27:1720s. I want to I want to be a part of
27:19this. I'm watching your videos like okay
27:21focus on signal verse noise. Don't
27:24listen and just keep working hard. AI is
27:26going to keep going. The alpha as you're
27:28stacking your chips building a poker
27:31chip which you're getting through your
27:32job. If you don't feel like you're
27:34growing fast enough in your job, then go
27:37build an app on your own and collect
27:4010,000 and then go build a second app,
27:41collect 10,000. Believe it or not,
27:43because of Stripe, you can do that.
27:45Like, you can create an app that
27:47competes with the world. And if you
27:48create 10 of them, they're each
27:49making$10,000 and that's $100,000 and at
27:53your job you're making $115,000. Then
27:55you can quit your job and go create
27:56these regularly and build up a stack and
27:59invest that money and it compounds. I
28:01met people at your event. That's what
28:03they basically do. They've made their
28:05financial wealth through investing by
28:08not listening to the bubble talk.
28:09>> Mhm.
28:10>> We got to do some bubble talk. It was a
28:12bubble week again.
28:13>> Well, why?
28:14>> I want to talk Bitcoin and bubble. I've
28:15had enough of of tech.
28:17>> What's going What's going on with
28:18Bitcoin?
28:20>> I think I beat him into submission
28:22there.
28:24Did you just cry uncle?
28:25>> May maybe maybe this is us. It's like us
28:27at Thanksgiving now. We we were together
28:28twice this week. Maybe it's it's
28:30>> for those that don't know, Jordan and I
28:31try not to talk during the week, so we
28:33can come in here, but we we had a
28:34midweek uh midweek session and so, you
28:37know, we're we're uh we're
28:38>> it's like Thanksgiving at the Pompiano
28:40House right now. I've been a guest for
28:42two.
28:43>> All right. What What is uh what's going
28:45on with Bitcoin?
28:46>> Uh it's consolidating. No, no, no. So, I
28:48I I do want to give because I've a bunch
28:50of people have asked and I I want to
28:52make sure everyone hears this.
28:54>> For Christ's sake, be patient,
28:57>> Jordy. Why isn't it at 100K? Yeah,
28:59exactly. Um, come on. It's October. What
29:01happened?
29:02>> Let me go through. I want everyone to go
29:04bring up a chart of Micron. You know, we
29:06talked about Micron when it was $100. We
29:08talked about when it was 150. If you go
29:10look at 2024,
29:13we started this thing. A lot of people
29:16in your office bought bought Micron.
29:18Should I stay in it at 200? Should I get
29:20out at 200? It went to 1,400. Now, if
29:24you go back and look, it broke above the
29:25200 day moving average, just like
29:27Bitcoin, just like Ethereum, and then it
29:30retraced back to it. It tested it. I am
29:33not sure that's going to happen with
29:34Bitcoin and Ethereum and the rest of the
29:36crowd. But you have to go look at why
29:40that happened. When it starts to happen,
29:43there's an AI theme behind it. I want to
29:45make sure people realize when it comes
29:47to crypto, this is 100% for me about
29:50agents. And I'm telling you, the amount
29:54of traditional people that continue to
29:56bang on my door to talk about
29:58tokenization, to talk about how to make
30:00money on this, they're doing their
30:02homework now. I think next year is going
30:04to be the year that we see the biggest
30:06moves in crypto because of agents. And
30:08so I just want people to realize that
30:10I've watched this now a lot, but in
30:13particular with the semiconductors. It
30:15took a long time when it was obvious to
30:17me because of AI, because we were going
30:20to need an enormous amount of memory,
30:23but very smart people were not on it.
30:25And right now, there are a lot of smart
30:28people running around saying we're in a
30:29bubble. That prevents people from
30:31jumping into anything. They see the
30:33rates going higher. Everyone should be
30:35grateful that with rates going up as
30:37much as they are, somehow another crypto
30:40went higher. with gold going down as
30:42much as it has gone down, crypto somehow
30:44has been distancing itself and going
30:46higher. So, when you go through this
30:48whole thing, I I I really do believe
30:50people have to be patient with this. Um,
30:52it's a bull market and I feel
30:54comfortable in saying that because I
30:55said when it was a bare market, we're
30:56above the 200 day moving average, the
30:58200 day moving average is pointed
30:59upward. Just be patient. These things
31:01can never go as fast as you want and the
31:03thing that will accelerate it the most
31:06is when the traditional finance crowd
31:08comes into it. And for that, we need the
31:11tokenization volumes to go higher. We
31:13need the tokenization asset size to go
31:15higher. We need the stable coin volumes
31:17to go higher. That's why I talk so much
31:19about Stripe cuz Stripe is talking about
31:22this daily. It's happening. Every person
31:25listening to this show, you probably own
31:26Bitcoin and you got it the same way. You
31:28paid full price on an exchange. But
31:30there's a smaller group of business
31:31owners, family offices, and investors
31:33with sharp CPAs. And they're quietly
31:35getting theirs a very different way.
31:37They own the machines that make it
31:38mining. Here's how it works. You buy the
31:40miners and Simple Mining hosts and runs
31:42them at their facilities in Iowa on
31:44industrial power rates that you can
31:45never get on your own. The Bitcoin they
31:47mine goes straight from your machine to
31:48your wallet. There's two reasons I find
31:50this interesting right now. One, the
31:51economics. Cheap power let you convert
31:54electricity into Bitcoin at a discount
31:56to spot. You're acquiring the asset
31:57below the sticker price that everyone
31:59else pays. But second, and listen
32:01closely, if you've got a big income
32:02year, miners are classified as computer
32:05equipment. Equipment means depreciation.
32:07Qualifying buyers can write off the
32:09entire purchase in year 1. It's the same
32:11playbook real estate investors have run
32:12for decades pointed at Bitcoin. Your CPA
32:15will want to see this before the end of
32:16the year. And so with attractive
32:18financing available, you can deploy
32:20these assets without selling any of your
32:22assets to do it. This is not some garage
32:24operation. Simple Mining runs over
32:2640,000 machines and was just ranked the
32:28fastest growing company in Iowa on the
32:30Inc. 5000. They put together a valuable
32:32resource called the 2026 Bitcoin Mining
32:34Blueprint. It covers the five mistakes
32:36investors make when allocating to mining
32:38and how to avoid them before deploying
32:40capital. You can access it for free at
32:42simplemining.io/p.
32:44That's simplemining.io/p.
32:47Go check it out today. Today's episode
32:49is brought to you by Abra, the secure
32:51way to grow crypto wealth. Abra offers
32:54individuals and institutions end-to-end
32:56crypto wealth management solution,
32:57including custody, trading, yield, and
32:59cryptobacked loans through a unique,
33:01separately managed account structure,
33:03where title is retained by you, the
33:05client. In other words, your assets stay
33:07your assets. Abra's been partnering with
33:09crypto holders for 8 years and has
33:11processed over $2.5 billion in loans to
33:13date. With Abber's loan product, you can
33:15access up to 50% of the current value of
33:18your collateral. There's no minimum or
33:20maximum loan sizes and Abberra can
33:21handle 9 figure plus loans. Rates are
33:24extremely competitive in the 4.5 to 6.5%
33:27APY range. Loans are open term, meaning
33:30you can keep borrowing against your
33:31collateral without closing a loan as the
33:33price appreciates. Buy a house, take a
33:35trip, make a large purchase, or
33:36diversify your investments without
33:38selling your Bitcoin. All deposits are
33:40held in secure MPC wallets, and there is
33:43no rehypocation. Find out more at
33:46abra.com and tell them Pomp sent you.
33:48Again, that's abra.com and tell them Big
33:51Pomp sent you. So, in the last two
33:55weeks, I have seen uh Black Rockck. I
33:58have seen Bloomberg and also uh Citrini.
34:03They put out a research piece with a
34:05nice uh image that had DeFi on Wall
34:08Street on it. And uh if I did not know
34:12Catrini wrote it and I read through the
34:14the notes of this, it could have been a
34:17Jordy Visser piece in terms of the the
34:19main message of there's this
34:21intersection of crypto and Wall Street.
34:23I think that some people will read some
34:26of these other pieces, whether it's
34:27Black Rockck, whatever, and be like, "Oh
34:29my god, Jordy Viser's ideas are
34:31infiltrating whatever." You said
34:33something to me interesting though,
34:34which is this is good because it means
34:37that like we're transitioning from
34:39almost like a contrarian view to
34:40consensus view and you kind of need
34:42that. Can you explain a little bit as to
34:45as people see this idea show up in other
34:47places why that's actually a positive
34:49sign
34:50>> it when you when you talk so the way
34:55people view you
34:57uh
34:57>> oh
34:59you've been positive on Bitcoin for how
35:01many decades now?
35:04>> Well Satoshi got the idea for me didn't
35:05he here?
35:07>> So I don't bring that up as as as as a
35:11dig. I bring that up as once you kind of
35:14have this story that you say over and
35:16over again, and this works for me too,
35:19your voice doesn't appeal to people who
35:22don't believe in you because they go,
35:23"Ah, he's a perma bull on crypto, so it
35:26doesn't matter." So, in fairness, the
35:28reason I'm saying this is if Jordy gets
35:31viewed as someone who will only be
35:34positive on Bitcoin, which isn't true.
35:37>> No, you were bearish last year.
35:38>> No, but bearish was
35:41saying I I I was never bearish. What I
35:43said is we're in a bare market like
35:45acknowledging the facts. You can't fight
35:47the market. Like the market is the
35:49market. I believe the market knows
35:51everything.
35:52>> What we're trying to do is forecast what
35:54the market will think in the future. And
35:56people have to realize the only way you
35:58make money in anything that you invest
36:00in is not because you were right. It's
36:04because other people that didn't agree
36:06with you at the time came in and pushed
36:08that asset higher. At the end of the
36:09day, you are trying to bet where the
36:11money is going to go, not where the
36:13asset is going to go. So that is the
36:16problem that happens. We I mean
36:18Salesforce.com is down yet their
36:20earnings are good. So fundamentals don't
36:22matter. What's driving it? It's because
36:23people were overweight SAS because they
36:25had a view that it was a stable business
36:27because private credit has a lot of
36:29software. There's a bunch of reasons. So
36:30for me, we need Catrini. We need
36:33BlackRock. We need other voices to speak
36:35up. In fact, I'm going to say this right
36:37now. Jordy wants to go to as many
36:39conferences next year. That makes sense.
36:41Now, I'm saying this, you're going to
36:43help me. I've already reached out to
36:45Natalie. Like, I'm gonna I'm gonna get
36:47out there because I do want to sit and
36:49make sure that people hear why I'm
36:51positive on it.
36:52>> Jordy Vistler World Tour.
36:54>> Yeah. I The one thing I don't
36:56>> T-shirts made.
36:58>> The one thing I don't hear enough about
37:00is the connection that Catrini did,
37:03which is agents to crypto is really
37:06important. And it gets back to the Mark
37:08Andre piece. And that's why I don't want
37:10this to be about the basement. I don't
37:11want, oh, the bonds are going down. The
37:13government's going to have to step in. I
37:15believe a year from now, it is very
37:17likely that 10-year yields will be about
37:19the same level a year from now as they
37:21are today.
37:22>> The only way I'll say they'll be lower
37:25is because if the jobs market gets
37:29noticeably weaker, which I don't think
37:31it will over the next year, I definitely
37:33think it's going to hover around zero.
37:35But if it got weaker then the debasement
37:36thing will become a bigger thing. But
37:38other than that, it's because we're
37:39growing so fast. What that me means is
37:42the only thing that's driving growth is
37:44AI. It's the only thing.
37:46>> And so the bet becomes about AI. And I
37:49have full belief that demand is greater
37:52than supply over the course of the next
37:53few years. So the buildout is going to
37:55happen. The spending is going to happen,
37:58which means nominal GDP is going to
38:00happen. I don't think the government can
38:02move rates down lower. And so if we're
38:04going to run this thing hot, that means
38:07the agentic world will continue to
38:08accelerate and that means crypto is
38:10going to do well. So I think the pieces
38:11by Catrini, by Black Rockck, by everyone
38:13is very important. I think where my edge
38:15is in this whole thing is really
38:17spending the time every single day of
38:19connecting consumer agents importance to
38:22what's happening right now. And when you
38:24look at that chart that I've said so
38:26many times from page 71 of situational
38:27awareness, everyone should just
38:30recognize that the chaos is going to be
38:32worse next year than it is this year.
38:34What we saw this year, all of the chaos,
38:37the daily, oh, we're in a bubble, we're
38:39this, it's only going to get worse
38:40because people are overwhelmed with the
38:42speed of change and we're going to
38:43accelerate next year in speed of change.
38:46>> How do they prepare and navigate that?
38:49I it
38:52signal verse noise is really important.
38:54Um but how do you do it?
38:56>> Okay, so this week a very famous hedge
39:00fund manager Ray Dalio
39:03was quoted as saying
39:06bond yields may and I and I emphasize
39:08the word may prick the AI bubble.
39:13So this gets a lot of attention. People
39:15send it to me and they go, "What do you
39:16think?"
39:17>> Bubble talk.
39:18>> Now, if if this was said by, you know,
39:22Joe Joe Smith on the corner,
39:24>> by me.
39:24>> Yeah. If Joe Smith got interviewed by
39:26Bloomberg and said, "I think it may
39:29prick the AI bubble, it wouldn't even
39:31Nobody would care." But because Ray
39:33Dalio said it, they care. But here's the
39:35thing, and I'll show this this weekend.
39:38Ray Ray Dalio. If you just go into any
39:40LLM and say, "How often has Ray Dalio
39:43said there's a bubble over the last 15
39:45years?" It's a lot. And it's not to say
39:48that he's not a smart guy. What it is to
39:50say is he has now made a career, his
39:53like post hedge fund career. He's
39:56writing books about the history of the
39:58world. He's a very smart guy in
40:00understanding the history of the world.
40:02So, he's saying we're going to end up in
40:04the exact same thing that we've had. And
40:05maybe we will. maybe they'll maybe
40:07something will go on, but I think he has
40:10a um a thesis that he has to stick by
40:14and this is he's just going to keep
40:15doing it. And so before it was a bubble
40:18on something else and as I sent you
40:19something and I'm going to show this
40:21weekend,
40:22if there was truly a bubble,
40:26I think if we had a pandemic, the whole
40:28market would collapse. Oh, check. We
40:31already had that. It didn't do that. We
40:32were back at new all-time highs and up
40:34for the year. Okay. Well, if the Fed did
40:37the fastest rate hike in the history of
40:39the United States in an un unexpected
40:42manner, we would certainly blow up this
40:44bubble of 40 trillion in debt and the
40:45whole world would collapse. We had that
40:48one. Didn't do it. Inflation hit almost
40:5010%. Didn't do it. Okay. Maybe after
40:53they were done though, we'd see the the
40:55after effects and we'd have the biggest,
40:57fastest bank run in the history of the
40:59country. Oh, check. We had that with
41:00Silicon Valley Bank plus all the little
41:02banks that fell too. We had deposits
41:04flowing out. And guess what? New
41:06all-time highs.
41:08Maybe if we decided to go back to Smoot
41:11Holly and we announced massive tariffs
41:14around the globe, that would do it.
41:17Okay, we fell 20%. Then we rallied back
41:20to new all-time highs. Finally, this one
41:23there's no way you can stop because
41:25those were all kind of easy ones for the
41:27Fed to print money out of. So, you could
41:28think that they were going to print
41:30money. What about if we shut down the
41:32straight of Hormuse?
41:34Every oil would go to 200, 300, 400, 500
41:37and we would well we had that this year
41:39can't print money and yet somehow or
41:41another the market found a way to go
41:43higher. So now the new one is rates are
41:45going to do it. So I think people have
41:48to just understand the way markets work.
41:52We've had new all-time highs many, many
41:54times during my career and there's been
41:55many, many bad things that have
41:57happened. And there may be another 20,
41:5930% correction next year. And it might
42:01be because rates get to such a high
42:03level that we end up with, I don't know,
42:06a bunch of businesses that we didn't
42:08know had a lot of debt in Europe, in
42:10wherever, that end up collapsing. I
42:12don't know. But I also know that if that
42:14happens, the governments have
42:16financialized everything. So they will
42:18come in just like they did after Silicon
42:19Valley Bank and make sure that they put
42:21money into the system to make it not go
42:23down. A pandemic took trillions and they
42:26did that. I don't know what other people
42:28want to do. So I think when you listen
42:30to these bubble talks particularly when
42:32it comes to AI the way to navigate watch
42:35my video. I am showing facts about the
42:39demand verse supply. I am going to show
42:41why Ray Dalio's conversation although
42:44may be true
42:46it's just the headline that he's put out
42:48there the alpha which names are working
42:51I'm literally going to go through and
42:52say okay almost nothing has worked
42:56nothing over the course of the last
42:59decade very few companies have made
43:02money if you go back to chat GBT only
43:0424% of the companies in the S&P 500 1500
43:08so 1500 companies only 25% % have
43:10produced a return greater than the index
43:12as a whole 25%. So it's very
43:14concentrated. Guess what? We've been
43:16concentrated for a long time with the
43:17MAG 7. If you're trying to make money,
43:19you need to know the things that are
43:21working. The only thing that I think you
43:22can make money on with the highest
43:24probability of certainty
43:26is things related to building out the
43:28compute because we don't have enough of
43:30it and we're going to continue to build
43:31it out. On the other side, I think you
43:34need crypto. So the alpha I'm trying to
43:36give. And then finally, the best way to
43:39feel empowered is to learn how to use
43:40AI. If that's what's causing the
43:42problems and that's where the biggest
43:44fear is, if you want to make sure that
43:46you know more than Ray Dalio, trust me,
43:47he doesn't sit on an LLM all day. So go
43:50build something, upload it to GitHub,
43:51you have something over Ray Dalio that
43:53he doesn't know about today's world. So
43:55the agency side, the empowering side is
43:56very important, too.
43:57>> I could not agree more. What are you
43:59going to do your video on this week?
44:01>> I just kind of did it. It's a bunch of
44:03that stuff. Uh we we just talked about
44:05it, but in more detail and seriously I
44:08uh I enjoy doing this. I had a great
44:10time with the group that I saw and I'm
44:13not kidding for people that in the
44:14crypto community that are having big
44:17conferences where you have a lot of
44:19people where you think I can add some
44:20value. I'm going to go to a lot of them
44:22next year and before my calendar fills
44:24up with things for you and for other
44:26people that have first dibs, I would
44:28like to be in a position where I can go
44:31to places and talk. I think the most
44:33important time to get back to your thing
44:35about Catrini, Black Rockck, Joseph
44:36Chalum, William Magguire, whoever you
44:38listen to right now, the voices are
44:40getting louder and louder on crypto. I
44:43think making sure that people have a
44:45chance to hear this and understand it
44:47and learn from it is really important.
44:49So, I think the next 12 months is is
44:51something I regret with the Micron thing
44:53of not really explaining to people, but
44:55I didn't have a big enough audience.
44:56Now, I have a big enough audience. So, I
44:58want to make sure that as many people as
44:59possible hear our show, but they also
45:01learn about the importance of crypto.
45:03>> The Jordy Visser World Tour begins
45:07January 1st, 2027.
45:10Book them now before his availability is
45:13out. By the way, did you know that I'll
45:15never ask you to go anywhere other than
45:16in New York City? Cuz I don't go
45:18anywhere for any event. Like, if we do
45:20an event, it has to be in New York City
45:21because I don't want to travel. Um,
45:23>> I wish I had a video for everyone to see
45:26that you and I passed each other in DC
45:28where you didn't tell me you were going
45:30to be there and it was literally not a
45:31joke. You were going out of turn style.
45:33I was coming in a turn style.
45:35>> High fived.
45:36>> Yeah.
45:37>> And by the way, he was going and Jordy
45:39Jordi is a amateur when it comes to
45:41conference attendance. He was there for
45:42like two days. All I went down on the
45:45train in the morning, did my 15-minute
45:46talk and went right back. I had the best
45:49time I've had at an event with those
45:52people in DC.
45:53>> They were they were great. I I uh had to
45:55come back for other reasons. But I have
45:57one last thing, a surprise for you. Uh
46:00did I hear that you read some of the
46:02chapters of my new book?
46:03>> Yes, I did.
46:03>> The book is so good that I made Jordy
46:05read. Just think about that. That's the
46:07only ad you need. It
46:08>> it got Jordy Visser to read a book.
46:10>> It it it's it's really not a book. It's
46:13a it's it's it's very similar to a book
46:17that I love, The Daily Stoic, because
46:19it's just passages. So, I can read like
46:21a page I can read a page a week. I'll
46:24finish it by the end of the year.
46:25>> They're they're letters, so they're like
46:26two pages long. Uh and there's 50 of
46:29them. But I just I figured out what's
46:30the best ad that I could make. Just the
46:33book's so good, Jordy Visser read it.
46:35>> And I love reading about spinach, so
46:37it's fine.
46:38>> The devil doesn't tempt you with
46:39spinach. All right. Thank you all for
46:41watching. We'll see you guys next week.

About this transcript

This page contains the full transcript of Bitcoin Is the Toll Collector of the AI Economy by Anthony Pompliano, generated from the public captions YouTube serves with the video. The transcript has 9,210 words across 1,329 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.