Elon Reveals GROK 4.20... and it's getting scary good

All right, everybody. Welcome to the live stream. I'm g I'm just going to

live stream. I'm g I'm just going to assume we're live. If we're not, nobody will know. But if we're live, you know,

will know. But if we're live, you know, welcome. We're live.

welcome. We're live.

>> Yeah. So, um I guess first and foremost, well, let's let's give a few minutes for people to kind of pile in here.

>> Um >> right, build some tension because you've got quite the story for everybody today.

>> Oh boy. Yeah, this was kind of insane.

I'm still not 100% sure. Well, it's

real, but I I I feel like it took me a while to figure out like, is this happening for real or or or not.

>> I thought the same thing. I was like, "No." I was like, "No, must be one of

"No." I was like, "No, must be one of the parody accounts or something." But

>> that's exactly what I thought at first and then I was like, "No, okay, that's not a parody account. Is he messing around? Is is what is this?" Uh,

around? Is is what is this?" Uh,

actually, we got confirmation this morning, so everything is 100% uh people are saying it's live. Yes, thank you.

Okay. Thank you everybody for being here. Um, how's the sound? Uh, sound

here. Um, how's the sound? Uh, sound

check. Testing, testing. I think we we said we're gonna start sounding good.

>> Two minutes in. Okay. Yeah. And then you sound great. Uh, at least on on my side.

sound great. Uh, at least on on my side.

All right. Let's uh let's dive in. So,

first and foremost, uh, we've had the Alpha Arena backed by the N of One team, I guess you could say, company, and they've been running a few benchmarks.

They had season one, season two.

Yes, Elon is cooking. Um, and so what happened was it was very interesting because these models, large language models from OpenAI, Google, uh, they had

Quen and Deepseek and Claude, etc. Grock was in there. They were doing pretty well trading crypto. And then later

season 1.5 started and, you know, they were trading stocks. So, Nvidia, Google, you know, Palanteer, kind of those stocks, tech stocks basically.

And they did have a mystery model, which we weren't sure exactly what it was. I

had a few guesses, but apparently they turned out to be incorrect. Um, but what happened was I posted kind of like my what I thought the model was. So I went

I looked at a paper that the company published where they created their own model for trading stocks in the in the stock market which is it's very interesting how they did it. So that's

like a whole separate discussion that's very interesting and I kind of thought you know I connected the dots this this is the model that's the mystery model

that's doing so well in the markets but I was very wrong because as I'm going to sleep last night I see Elon

responds to that message going actually it's Grock 420 or 4.2 2 um an experimental version of it. So

it's not released. It's an experimental version of it. Um and the thing about that model is it's I mean it made a lot of money. [snorts] Um had a pretty good

of money. [snorts] Um had a pretty good ROI in a very short period of time and there were four sort of constraints under which it ran and it made money in

every single one of those which is weird. There's a lot of Yeah.

So I mean I guess let me pause there.

like what's your whole take on this?

>> Well, yeah, we have I feel like there's so many different ways we can talk about this, but the the thought is there is models out there that are that are just not public and they're doing all sorts of interesting things and all sorts of

interesting fields. I always constantly

interesting fields. I always constantly wonder which ones are out there trading because it's just one of those dead obvious, you know, billions, maybe even

maybe even trillions of dollars are like built around these investment firms that are actually trying to beat the market and they always have, you know, frontier technology doing things. And to see that

this one was actually one of Elon's models means that it was trained in one of those, you know, servers that he has, which are top-notch. And we have, you know, he's got a data team behind him

that's very interesting. And Twitter has a lot of stock integration, which is really interesting. So, just seeing it

really interesting. So, just seeing it all come together and actually out there in the real world just kind of shocked me this morning.

>> Yeah. And there's there's so many different things that I guess we could talk about. I mean first and foremost

talk about. I mean first and foremost um I mean a lot of people are wondering you know so I guess we don't know how well this model is going to perform in

the long run but I mean let's imagine for a second that it continues sort of beating the market in a very spectacular way like what if um you know what implications does that

have sort of for the world that's going to be interesting uh interesting to to note Um, number two, like one of the big

reasons I like this benchmark is it's very hard to game, right? Because, you

know, first of all, nobody's going to lose. You you can't, you know, if you if

lose. You you can't, you know, if you if you win, yeah, to do well, you have to do well in the stock market. Like, it's

a very hard thing to beat. Um, it's very competitive. It's a zero sum. You know,

competitive. It's a zero sum. You know,

you can't really game it. It's happening

in real time. There's no we don't know the answers until you know the future happens and we know what stock goes up goes down. Um the stock market didn't do

goes down. Um the stock market didn't do that well in during you know the period that they were testing these models and video was even down a little bit. Um and

yet this thing was just you know I think over 12% return plus 12% [clears throat] in those in that twoe period.

I don't know it's there's a lot. Yeah.

>> Yeah. So I'll I'll just throw in a few thoughts. Now I I have followed a few

thoughts. Now I I have followed a few places before where people try to track the stock market using Chad GBT and normally what it does better than other models is sort of sentiment analysis.

It's the better tool right now for seeing what's trending, what people are talking about, what's kind of a hot topic, which does, you know, push the market up and down. like fundamentals

are part of it, but certainly hype is part of it and you hope that fundamentals drive it in the long run and hype and news cycles kind of run it in the short run. So LLM have, you know, traditionally been pretty good at that,

but also there's a speed element. they

can um read quarterly reports instantly from every company at the same time and sort of make connections about who's working with who and like how things are

growing and then kind of give you give us human feedback to then sort of look through. So, um I have no idea like

through. So, um I have no idea like where it's going to go, but I don't even I don't even know if the stock market survives forever, you know, or maybe it

does, but it's this world where not everybody has access to a frontier model and they're sucking up most of the resources. I don't know. Where do you

resources. I don't know. Where do you think it goes?

I have a very hard time envisioning it.

Some people uh sometimes in the comments they'll say um well yeah it means the market's going to be more efficient so it's going to be harder to sort like everything's going to be priced more accurately. There's going to be less

accurately. There's going to be less whatever alpha to be had or whatever. I

mean that's assuming that everybody or enough people have access to these models and they're all at the same level. But it doesn't seem to be it

level. But it doesn't seem to be it seems to be like there's always going to be a better model or if there's going to be a model like this there's going to be

a time during which only a few people have access to it. Um and

>> you know >> I mean like we so to be clear like me and you can't go use that model right now right but you know somebody inside of Elon's orbit can. And do you think

they are or do you think they're just playing as a test?

I mean, they're, you know, if this is not a fluke, they're obviously testing it. I would,

they would have to be testing it, uh, in other places. It's not like they

other places. It's not like they randomly decided to set it live in this benchmark. They would have to have

benchmark. They would have to have tested it somehow beforehand.

Um, >> yeah, think about that data. So, that

data source is so interesting like Twitter actually is the the fastest or, you know, X is the fastest place to see what's happening. And um you know I

what's happening. And um you know I actually have a story like I had a a friend that was um working at Twitter at the time when it was was purchased. And

>> there um there is actually uh for years an initiative to make it so that when you typed in a stock the uh number would uh the stock ticker would like come up

in the feed, right? And when Elon took over, he saw that that was sitting in the feed and nobody had like actually finished it. There was just too much red

finished it. There was just too much red tape. And that was something that got

tape. And that was something that got solved pretty quick when he was there.

And we've heard him talk multiple times about how one of the ways to profitability for X is to kind of turn it into something like a PayPal, which he's done in the past, and and think of

it more as a financial kind of uh money exchange network. Um certainly crypto

exchange network. Um certainly crypto seems to be part of the future in the way that he's kind of envisioned it. So

yeah, predicting how the market is going to go might be one of those big in inside insight initiatives at X.AI that they might be uniquely position through

data and computation and architecture to actually do.

>> Yeah, that I mean it definitely seems like there's a lot there. [snorts] Um,

so in this, as far as I can tell, with with this approach, how they were doing it with the alpha bench or alpha, what's it called? Alpha arena. Um,

>> they were feeding all the news and stuff to the models in a prompt. So, all of them had the same exact information.

Um, but definitely it seems like the model like why would it be so good at it? It

seems like they they figured out how to train it on some specific data. Probably

exactly what you're saying, all the stuff that they have available from X and all the data they have from, you know, the data in the stock market. I

mean, there's something something interesting uh happening there, right?

>> Yeah. I mean, imagine like one of those data sources from like Robin Hood or investment trading firm that whoever is first to get that data maybe could just get that much better of a signal. Um, I

don't if you ever heard a book called The Flash Boys. It was I remember reading it quite a while ago, but it was just about how intense uh the the fight was to get real estate that was just inches away from the New York Stock

Exchange so that you could make, you know, models act faster. And of course, this was pre-AI. There was a few evolutionary algorithms. There was a few just handcoded algorithms. And then

there was just a lot of sort of knowledge and kind of skill that went along with the people who got a feel for how the market works and where the exploits are. But a general

exploits are. But a general intelligence, >> never seen anything like it really.

>> Exactly. It wasn't it like milliseconds that they were milliseconds they were trying to shave off in terms of how fast the wires were. It's insane. I think for one company it was like worth the

multi-billion dollar investment to run a cable under the ocean from like London to New York so they could just talk to each other faster and like skim off some amount of money. And I mean there's a

whole another kind of question there. Is

it like really a you know is the invisible hand of the free market even is that really what it what's happening at this point? But you know maybe kind of maybe kind of not.

>> Yeah. Uh I

I I want this news to sort of break so people um that are experts in the field maybe can talk a little bit um about it.

I guess let me show some of the stuff um that that I've been seeing by sharing my screen.

If everybody just um let's see haven't used Riverside in a minute.

How do you share screens? Oh, boy.

Probably should have >> Oh, that's fine.

>> Thought about this beforehand. Um,

>> while you're do while you're doing it, I can I can tell a quick story if you want.

>> Yes, please entertain people.

>> And then you bring uh you go you go bring up those things. So, um I have a friend uh that did some kind of interesting investing through Chad GPT a little while ago, too. and he he was

kind of curious to ask it about stocks and sort of who the owners were and the ethics, you know, in the in the chain of

um how things were built. And I've never quite seen anybody be able to invest their money ethically like that before.

So, one of the cool things that I've been thinking about recently is that, you know, with a Chad GBT or a Gemini, you could actually say like I, you know,

I care about countries um that that are like poor and what stocks are actually like helping build out their infrastructure. You can say like what

infrastructure. You can say like what kind of um companies can I invest in where the leaders donate like the largest percentage of their profits or have philanthropic foundations that

actually do things and you can kind of check into it. So, one of the craziest things is that if you do a deep research query and you actually have a reason that you want to vote with your money on

how the world can be different, we have a tool and it's chat GBT and not anybody really thinks about it as much as just like how to make money, but if we all put our money like ethically where we

think the world could go, it could make a big difference and we have that tool now.

>> Yeah. and or or even at the very least even if you just wanted to make sure it's not going somewhere you don't want it to right like even if you're like you know I want returns but I really don't

want to support anything that has to do with these particular industries um so whether you're like trying to specifically you know do do sort of like

positive change or you just there's some things that you really don't want to support you can kind of like it's whatever you want you can kind of um target your research in that direction.

>> Yeah, this person was actually an anti- like Elon person. He didn't want to support Elon Musk anymore. But of

course, he wanted all sorts of things and technologies, but he wasn't sure which ones were connected to his um world or to Sam Alman's world or to Google's world. And if you, you know,

Google's world. And if you, you know, don't believe in one of those people, which he didn't, he said, you know, how do I invest money and I don't invest specifically in this person, but still in the technology sector? and it found a

whole list of like 20 stocks that are like very far away from Elon's orbit.

>> Um, so yeah, it just shows, yeah, if it is hopefully it's more in the positive direction, but yeah, it can be about avoiding and yeah, avoiding with your capital certain people and certain ethics.

>> Yes, somebody's mentioning the EU was going to uh find Elon 120 million. Um,

if what we're looking at right now is for real, then that might not be an issue for him, but um, well, I guess it's not an issue either way, but okay.

So, this is what we're talking about.

So, this is Alpha Arena. This is at nof1.ai.

nof1.ai.

N of one is the company. They got some pretty interesting people behind it.

Some like serious people. Um, one is an ex Citadel person with a machine learning PhD candidate, I guess. Um, so

the Citad Do you know a lot about Citadel? So, it's a massive hedge fund.

Citadel? So, it's a massive hedge fund.

I think in that it was in that movie.

Um, do you know what I'm talking about?

Deep Value.

>> Short.

>> Oh. Uh, more recent.

>> Citadel is in the Big Short. I don't

know. Yeah.

>> Oh, it was the more recent one.

>> No, I Somebody help us out.

>> Movie.

>> The Citadel.

>> Citadel is gigantic. Here, I'll check.

>> Yeah. So, it's about deep effing value. You know, the Reddit

effing value. You know, the Reddit traders, the uh I I guarantee you chat is going to know what what what we're talking about. What movie was that? I I

talking about. What movie was that? I I

you somebody will they know exactly what I'm talking about. I'm 100% sure of it.

But, um mystery model. So, um it was broken up. So, this competition was

broken up. So, this competition was broken up into four different sort of I want to use the word tanches here just like they did in the big short, but that's probably not the right word. Um,

so you have the new baseline where it's trading equities, news, and sentiment data is getting fed to it, I believe, every six minutes, and we'll we'll kind of verify that in a second, but um so

this is kind of the the regular, but uh Dumb Money, thank you. Thank you, Hagar.

Yeah. Yeah.

>> Everybody else, thank you so much. Yeah.

So, yeah. Uh, Dumb Money and the, uh, it was that that was the movie pretty good.

Pretty I I enjoyed it. Um, and it was just about the the Reddit, the GME trade, you know, all that stuff. So, um,

and I guess the Citadel was the company that was one of the main sort of companies that they were looking at as this these people shorting these

stocks. I'm not going to go into the

stocks. I'm not going to go into the story, but um so notice this mystery model here is up to uh $11,812.

Everything else is is negative ROI, right? So everything else lost money.

right? So everything else lost money.

It's up under sort of normal conditions.

Monk mode is emphasis on capital preservation, survival, stronger risk management policies.

Um, let me see if I can Yeah, so it's still Oh, you know, it just slightly dipped below 10,000, but I mean it was the

competition was supposed to be for two weeks. So during that time it was

weeks. So during that time it was positive in every single category. Oh,

and looks like it's it's up, right?

Yeah. So it, you know, it kind of varies, but I mean it didn't lose money on any of these. um in the aggregate it did not lose money nor did it lose money

on any one of these sort of approaches.

The third one was situational awareness.

So this is where the models are aware of their competition, their current rank and the uh profit and loss statements of the other models etc.

This is insane to me because it's up to well past $15,000.

So, uh, so, so, so somebody's saying not, uh, a normal market. Um, so that's an interesting point because, so I'm not sure if you're referring to kind of like the the market during the two weeks that

this was taking place, but for example, Nvidia was down a little bit, some portion of 1%. Uh, like the NASDAQ was

only up, I want to say 3%. Right? Mhm.

>> So, it's not like the stock market surged up and this model just rode the wave up. That's not what happened here.

wave up. That's not what happened here.

Here's the max leverage. So, this is the only one where it's not the number one model. GPT 5.1 is, but notice it's still

model. GPT 5.1 is, but notice it's still it's still positive, right? Um,

and if you think about it, so it's a plus 55% return, right? So you give it $10,000 and comes back to you with 15,000

plus in two weeks, >> right? And imagine it's a million

>> right? And imagine it's a million dollars and you're like, "Okay, I'll just get an extra house out of that couple weeks and I'll just reinvest it and take dividends."

>> I mean, this is absolutely nuts.

>> Absolutely. Absolutely crazy. Yeah,

>> imagine imagine if this man if they have a reliable way to to just beat the stock market. I mean,

just I don't know. I've always wondered if there was ever going to be anybody that sort of did that with AI, but maybe maybe there will be. And maybe when that technology gets out, there'll be some competitors, but there might be a brief

window here where there's just a ton of money that it can handle.

I I kept waiting for this. I kept

looking for papers that would show something like this. Again,

it's why it's so bizarre. Just

everything about how this happened is just weird to me. Um just I think it's this it's called um Elon's razor or Musk

razor is like the most entertaining outcome is the most likely. This was to be the most entertaining outcome possible. So of course this is exactly

possible. So of course this is exactly what happens. Um, so for people that

what happens. Um, so for people that maybe don't know the story, I go on this sort of deep research because we didn't know what the mystery

model was. And yes, for people asking,

model was. And yes, for people asking, well, hang on. Some people are asking, is the mystery model Grock 4.2?

No spoilers. Okay, hang on. Hang on.

[laughter] This gets good because like I have this whole, you know, that uh it's always sunny in Philadelphia when he's like pointing at the map and he's so like crazy. He's like connecting all the

like crazy. He's like connecting all the dots. That was me like trying to figure

dots. That was me like trying to figure out what this model was because the people behind you know N of one they published a paper that was very

interesting. They took the same sort of

interesting. They took the same sort of approach that the Darwin girdle machine did. The same approach that kind of like

did. The same approach that kind of like the alpha evolve from Google deep mind did um you know Nvidia's Eureka or or their their Voyager models. So it kind

of had some of those ideas behind it and with those ideas and a large language model they built a

um a a a model they called profit or or profit. So it's if you can if people can

profit. So it's if you can if people can see that you know profit profit whatever program search for financial trading. So

they use this evolutionary search for automated discovery of this trading algorithm. Now, to me, I obviously put into two together. I'm

like, obviously this is the model because they just published this a couple days ago and it's there's a mystery model. I'm like, of course, this

mystery model. I'm like, of course, this is it. Um, so I I posted all my results

is it. Um, so I I posted all my results and I I I was feeling pretty pretty pretty clever. I thought I nailed it.

pretty clever. I thought I nailed it.

And then Elon comes in and goes, [laughter] "The mystery AI model is an experimental version of Grock 4.2 420."

I guess there's nothing like watching somebody else connect the dots differently than you know they are and then needing to be like fine I'll just go chime in here. I got to correct the record on Wes here.

>> But you know what [laughter] I mean just I look I um I was joking that uh you know what there's this law online. It's

like the the fastest way to get an answer on the internet isn't to ask a question. It's to you know make the

question. It's to you know make the wrong answer and somebody's going to show up and correct you. And that's

exactly what I did here. you're watching

it play out live. I said [laughter] I stated something incorrectly and Elon comes out. He's like, "Nope, actually

comes out. He's like, "Nope, actually it's Grog 4.2."

Um, so I mean, what do you think about this whole thing in terms of I mean, do you have any any thoughts? I mean, you've

had a few thoughts, but I mean like anything that's that's like weird or or or jumps out at you about this? I mean,

my first thought was, is this true? By

the way, the person behind So, just to confirm for people, the people behind this did confirm that this is true, >> right? Yes. You did some investigative

>> right? Yes. You did some investigative journalism, you know, you went and said, "Hey, like, is this really the model?"

You can see Elon said that it was, they said it was. So, yeah. I mean, I don't I don't doubt that it's true. So, I live in a world where facts are facts. So if

this model outperforms today, is it evidence that it will always outperform or that this is going to continue for months? Not exactly, but

it's a sign that maybe. And it's a possibility that when you put this much data into a neural network, it can learn something. Whether it's the speed that

something. Whether it's the speed that it can think or whether it's the way that it thinks in a that is different than any human and any model has before.

there is like un unlimited money to be pulled out of the stock market if you know how to move things up up and down.

I've done some videos on dynamic pricing before about how um you can have different AI models that are trying to op optimize the price of an airline ticket for example and they'll slowly

start to sync together right without sharing any information technically not doing anything illegal but if they're always moving prices up and down like take Delta and Southwest or something and they're they've got a both have

different AI models different architectures they're both trying to optimize money they can learn to not go too low and to like basically say, "Hey, when I go lower, then you go lower." And

then the free market brings it too much down and we can stay up here together and make money without ever talking to each other. So, there's interesting

each other. So, there's interesting knock-on effects that might happen if we live in a world where there's hyper intelligent AI that's so good that it can kind of figure out the market, but

also its place in there with other people and other agents and it's not just humans kind of doing a selfish thing, but it's more of a coordinated effort that's being coordinated just

through the information from the trades.

Like, there's I mean, it's a Pandora's box of what the future looks like, it seems like.

So I've heard also a similar thing to what you're describing with actually stock trading bots. Um something over time like with reinforcement learning

over time they also learn to somehow coordinate their moves somehow you know again without communicating without but somehow they figured out okay if we like

I can see somebody on the other side of these trades doing this. So it's what is it like the p prisoners dilemma or something like that. Um, it's

fascinating uh how reinforcement learning will push them to discover these things that I don't know. You'd

think you'd have to have some sort of a communication to be able to what is that? Collusion to collude.

>> Yeah. Collusion. Yeah. Yeah. I mean,

there are some examples like, you know, if you're walking your dog every day and you realize, you know, you kind of keep bumping into someone at the same time and the dogs sort of want to wake up at the same time to see the other dogs.

Like, these things do happen. But it's

just it was pretty fascinating how people have sort of a limitation and for the most part the stock market is too complicated and the rules that we build to make money selfishly tend to not

figure this out and it does kind of seem like we're entering a world where there's an intelligence you know behind these systems that kind of is learning

the coordination problem the collusion kind of solution to making money.

So, uh, somebody's asking dynamic watcher, would it buy? Did it succeed only because it bought Tesla before an upswing? Can we see what the actuals

upswing? Can we see what the actuals actual actions are?

>> Yes. And so much more. Um, I think you're going to be kind of blown away because first and foremost, this is running live >> and this is season 1.5. We already had

season 1 and uh, we're going to have season two. So, not only is everything

season two. So, not only is everything logged and you're able to access it in real time in the previous season, actually.

So, this is Tyler. Wait, you're not seeing this. Hang on.

seeing this. Hang on.

This is Tyler Storm. So, he's at XAI.

Um, speedrunning XAI as he puts it here.

Um, so he's uh he had a company. I guess

it got sold to Ripple XRP for people that follow that stuff. Now he's at XAI.

So um so he mentioned here that on the model page you can find a link to all the trades that were made.

So uh for people that you know I guess it's um I assume this is for I I assume this is for um crypto only.

But basically, if you want to see the crypto trades, they're actually on the blockchain. Verifiable

blockchain. Verifiable transactions. Um,

transactions. Um, oh no. Oh, shoot. Okay. No, this is I'm

oh no. Oh, shoot. Okay. No, this is I'm sorry. I I I did not pay attention. No,

sorry. I I I did not pay attention. No,

these are the actual uh trades that are recorded live on the blockchain. So, so

for for for people asking um so uh so this person of XAI who worked

with in the crypto so he had a he had some sort of a startup with AI that got sold to Ripple if I understand correctly I'm not >> too familiar with his background but

everything is logged right so you can see these um happening live and getting logged on the blockchain so this is this can't be falsifiable It can't be backdated if that makes sense.

>> Yeah, it was more It seems like it's more just a question like did he I I doubt it but yeah, there could have been the thought that like oh he just bought a bunch of his stock before an announcement or something. You kind of

see that like with different companies, a little bit of insider trading, but no, it seems it seems like this probably isn't that case.

>> Oh, no. The the biggest jump for example was Nvidia.

Um, and I'll show you kind of because they they analyzed in real time. So,

this is uh Jay Azang.

Um, oh, so yeah, I I kind of made the joke that, oh yeah, is he just trading Tesla?

Is it trading Tesla based on um, you know, somebody posts something negative about it, they just buy into Tesla or whatever? Uh, I think the biggest

whatever? Uh, I think the biggest actually some of the biggest gains came from Nvidia. So, for example, this was

from Nvidia. So, for example, this was one of the trades that they've talked about um holding Nvidia with 10x leverage, right? And they they they they

leverage, right? And they they they they walk through their their reasoning.

>> Um >> and as far as you know, the the model is they're all all of these models are getting the same information to make a decision off of, right? Or do they also have different ways they're connected to

the internet?

So, as far as we know, every single one of these is getting the same information because we're able to see the the prompt. So, for example, for any of

prompt. So, for example, for any of these, I'm able to look at this and say, okay, I can see the user prompt. The

user prompt is this massive thing. By

the way, this is what I've seen with other companies, other research projects like out of Nvidia. This is exactly how they do it. each turn which I if I understand correctly happens every six

minutes it it gets fed everything so it gets fed you know NASDAQ wears everything you know all the technical signals it gets fed its holdings right

so it kind of gets an update about its own holdings it also gets fed all of the narratives and all of the news right so

Nvidia narrative new Malaysia data center whatever All right.

>> Wow.

>> And if I understand correctly, this is the same um user prompt for every single one of these. So it's not like one of them has access to better data than another one.

>> Yeah. So it's not just about front running. It's not about a different data

running. It's not about a different data source. It's just about making a more

source. It's just about making a more intelligent decision with the same data >> that was provided. That's crazy.

>> So no, not just crypto.

not just crypto. And in some of these cases, I think they were going like the max leverage one.

Uh I mean it's max leverage. They're not

even like Yeah, they're going like full auto. Um you know, um

auto. Um you know, um by the way, so you know, the reason why I want people to poke holes in it is because as far as I can tell, there's no

there's no holes to poke. like this is as as sort of like I I haven't seen anybody say how this could be faked or gamed in any way.

So if anybody has any idea, let us know.

But as you can see, everything gets recorded on Hyper Liquid. Right? So you

can see these things.

We see the um the user prompts. We see

their reasoning. So for example, here's the chain of thought. So here's how this model reasons, right? It has its justifications. It has its stop-loss. It

justifications. It has its stop-loss. It

has kind of the invalidation conditions.

So it's saying, "Okay, if it really swings this way, then I was wrong. My

thesis was wrong." Whatever. Right.

[snorts] Wow. That is so interesting. I can't

Wow. That is so interesting. I can't

believe it. Yeah. And and and a human can just sit here and watch this and say, "Okay, I agree with that thesis. I

don't. We'll just see if this one works.

If it thinks so.

And it made intelligent stop losses, too. I mean, how many how many months

too. I mean, how many how many months would this have to continually beat the market before you just said the stock market's broken? You know, like just

market's broken? You know, like just year after year after year, would you be like, "Okay, they're just better than the rest of the players, but the stock market still works." And then you see a couple more come in from Deep Mind and

Open AI, and then it's just like, wow, a whole new market, an agent based market.

>> I agreed. So, the remarkable N, thank you for joining us. They're legally

changing their name to Max Leverage, which now that I think about it, you and me both, uh, that sounds pretty good.

Um, Woody Sloff saying, "It can't be fake, but a couple of weeks is just too little data to be statistically >> relevant." Agreed. Here's

>> relevant." Agreed. Here's

>> here's here's my kind of counterpoint to this. So I I

this. So I I don't have a statistics background, but I mean I I did take a few classes, so a long time ago, so correct me if I'm

wrong but the fact that it won across four different ones is a little bit more than just a few weeks worth of data. Do you know what I

mean? Like

mean? Like >> you're saying different condition. It's

being intelligent in different ways under similar conditions but in different ways with different goals.

>> Yeah. Because I mean you can run the statistical significant tests in different ways from what I understand.

Usually they do it sort of one um Grock 4.2. Well it is here in a sense

Grock 4.2. Well it is here in a sense that we're watching well okay yes I I this is an experimental version of Gro.

So, in that sense, no, it's not released yet, but we're seeing it sort of interacting with the real world in what I think most people consider a maybe a little bit of a scary matter. I don't

know. I I just >> I wonder if he's going to keep it open.

I mean, supposedly open source is like Elon's jam, you know, but this one seems like one of those models were >> Yeah. It's like, are you are you gonna

>> Yeah. It's like, are you are you gonna open source this one?

>> This Oh, man. See, this is like a was it a bucket of worms or whatever. Um,

>> that's how we first talked. The can of [laughter] >> That's fine. It's fine. You know what?

At this point, you might as well upgrade the can to a bucket. I feel like that makes more sense. There's enough worms. >> I think that that was our first online interaction is uh Anyways, we're

[laughter] not gonna um Anyways, uh I wonder if anybody knows that story.

Whatever. Um if somebody knows it, they'll say it. the the fact that it's um did it across four runs I think shows that it might be a little bit more I we don't know if it's statistically significant. Yes, we're gonna find out

significant. Yes, we're gonna find out if it runs for however long, right? If

it runs for three months and it just keeps making money, if it keeps beating the market, >> I mean I don't know. I get Well, let me ask you the people in chat.

What would you have to see um to say okay this is real? like how

many months of it beating the market or by what sort of percentage, you know, >> and uh let me so I have to go in a few minutes here, but let me throw out one more story too before I go. There is a

there's a video that I'm going to cover um an article in my next video where there's this story that's I don't know if anybody caught this, but there's a type of Pokemon card out there where

there's I think about 2,000 of them. And

they weren't particularly special a couple months ago, but there's only 2,000 of them. And there's one person out there that decided that they were secretly going to try to buy them all.

And they got all of them except five. So

they got whatever 1,00 um 900. And and

they decided to show on Instagram and on social media that they had all of these cards and that there was only five left.

And now they're in a position where they kind of can swing the market on these cards. And because it's become sort of a

cards. And because it's become sort of a social media thing, there's more value in each card if they do sell some from their stock.

>> And there's even a concept where the creator might take all of the cards except a few of them and burn them. So

everyone on the internet can see that the cards are now gone and there's only the five that are out in the market and the maybe five left that he saves from the burn and they're worth then they

become worth so much money and it's such an outside the box way to make money. I

think it just made me wonder how many other ways to make money besides that are outside the box that a human would almost never think of or you know a type of value that isn't even coming from a

stock market or a crypto. It's coming

from the marketplace dynamics of objects that matter to people in a community.

And there's just it seems to me like as AI gets smarter at just making money, not just beating the stock market, we're going to see so many more strange examples like that.

>> Yeah, exactly. Because yeah, these RL approaches, they often result in weird examples or shortcuts or stuff that we didn't expect and yet it somehow, oh

yeah, here's a way. Like, no, not like that. [laughter]

that. [laughter] So, I wonder what it's going to come up with for for the markets for stuff like that.

>> Yeah. And it would be like, "Hey, I didn't hurt anybody. I just cornered the market on like a particular Pokemon cards." Like, I don't know. Um, I do

cards." Like, I don't know. Um, I do have to run though. Is it okay if I jump the audience about how long?

>> I I actually, yeah, am going to be getting off in just a few minutes myself. Uh, yeah. So, anything else that

myself. Uh, yeah. So, anything else that you want to add? Um,

>> I I know you asked the audience how long they were going to they thought it'd be until they trusted it. Did you get a good respon?

>> Oh, yeah. Sorry, sorry, sorry. I I

didn't catch that part. Yeah. Um, so

some people are saying six months, some people are saying 12 months, five to 10 years. Here's why I think that I don't

years. Here's why I think that I don't Here's why I think that that's not that's not really needed because if you're running multiple copies, right?

Like if all of them are beating the market in their own way over even a shorter period, just the fact that you can run multiple models kind of tells you, do you know what I mean? Like if

yes, if there was only one model, then yeah, we need to run it for a long time.

But if you throw one uh a trading crypto and it's making bank, you throw one at trading tech stocks and making bank. You

throw one at bio stocks and it's making bank like you know, you're like, uh, okay, something's happening here. You

know what I mean?

>> Well, even these four profiles that you're talking about, this is what a good financial manager does. They take,

you know, maybe they break your money into four or five pieces and they say, "This one's going to be the very safe one. This one's going to be the higher

one. This one's going to be the higher risk but higher reward, you know, group." So, you know, there's probably

group." So, you know, there's probably one more AI that's looking at how to actually divide up a package and take risks. But yeah, why not? And you know

risks. But yeah, why not? And you know what? Also, with agents being as cheap

what? Also, with agents being as cheap as they could be to spin up and you know, maybe you you say, I want, you know, a lot of diversity. I want to be able to survive any kind of market conditions. And it's it's reallocating

conditions. And it's it's reallocating money between these four profiles, five profiles, 10 profiles, you know, for different assessments. So,

different assessments. So, >> yeah.

But man, it's going to be it's it's going to be interesting. Again, so a lot of people are pushing back on this idea that this is going to work like we, you know, think it might. And no one's saying that this is it. You know, it's

going to break the markets. Um, we're

still kind of paying attention to we're seeing where it goes. But

I don't know. At some point, I feel like we're going to see something like this and and it pays to be ready, I guess.

>> Yeah. Yeah. And I guess another question would be if like does it still work the same if all the agents are out there around the world competing for money

too? I mean like maybe right now it's oh

too? I mean like maybe right now it's oh my gosh look at this like secret gro model is better than everyone else but sometimes humans figure out the stock market and they beat it for a couple weeks a couple months in a few cases for

a year but most of the time the alpha doesn't stay with one person and it kind of evens out. So maybe this is more of a moment where the hyper intelligence is working, but that doesn't mean as soon

as uh a couple people over at Deep Mind see this paper, they're not going to launch a few versions and eventually China does some open source stock trading version and they do kind of balance out again.

>> Can you imagine? I mean, yeah, if if China open sources something like this, I'll post this link in in the in chat for people that want to check out the website themselves. Um but yeah,

website themselves. Um but yeah, absolutely. I just think this is

absolutely. I just think this is fascinating. So again,

fascinating. So again, you know, tomorrow everything could crash, right? Uh and then the model

crash, right? Uh and then the model starts underperforming and then we're it's going to be a big nothing burger.

But if not, then we have some uh some hard questions to to answer, I guess.

>> If you uh if you did have Elon Musk here to ask a question of what would you have asked him? Well, I mean, I tried asking

asked him? Well, I mean, I tried asking him basically um if it's fine-tuned or if it's just kind of like the base model. I'd be

interested to know if it indeed keeps up keeps up like this. Um so, I apologize they're saying I don't have a stable connection. And I apologize if there's a

connection. And I apologize if there's a break, but um would he release something like this or even would he like open source something like this if it had

these sort of powers and kind of like what kind of an effect does he think it might have? Um I mean he's been saying

might have? Um I mean he's been saying UBI or something like it like that is going to be needed and not just basic income but like luxurious income or whatever.

>> Oh, is that how that maybe that's how that's paid for? Maybe his model is on the internet making trillions of dollars through stock trades and then that powers the UBI.

>> I So I mean I've never I don't think he specifically talked about it but certainly I mean there's a lot of questions like this, right? It's like how do we deal

this, right? It's like how do we deal with AI and stuff like this as it keeps getting better and better? I don't know if anybody has any answers. We have the idea of UBI um or some sort of a freedom

div dividend or whatever Sam Alman was talking about. Um, we of course talked

talking about. Um, we of course talked to Immad Mustak who has some ideas about that as well. I don't know if we have one idea that everybody's agreeing on, right?

>> Yeah, just conversations for now. So,

>> yeah, but at least at least, hey, this is this is good. People should be talking about this. It seems like, you know, this is coming. So,

>> yeah, I will say I mean, you know, I would have guessed Elon Musk didn't know who you were until today. So, at least he saw your videos and knows you're thinking about it and got got to break some interesting news today. So that's

cool.

>> He I think responded to one of my things before. I completely missed it. I was so

before. I completely missed it. I was so mad.

>> This is your second reply from him. What

was the other one?

>> There was there was a few there there was a few. I think just usually uh having to do with AI. He responds to a surprising amount of people. It's very I don't know if he ever sleeps, you know

what I mean?

>> But um >> anyways. Yeah.

>> anyways. Yeah.

>> So yeah, but I'm glad I'm I'm sort of Yeah. saw this happen live. This was

Yeah. saw this happen live. This was

kind of amazing.

Nice. All right. Well, thanks, man.

>> Yeah, absolutely. Thank you so much for joining me. I appreciate it. Uh, talk to

joining me. I appreciate it. Uh, talk to you soon. Good luck in your travels.

you soon. Good luck in your travels.

I'll stay on for a few more minutes.

Chat to the chat to chat, I guess. So,

>> chat to chat. Yeah, I wish I had more time, but I'm going to go catch a train.

So, I will see you guys later. Bye.

>> All right. Take care. Thank you. Um,

okay. Yeah. Thanks everybody for joining joining us today. Um, sorry if I wasn't answering questions. Feel free to throw

answering questions. Feel free to throw uh questions or um you know what, whatever you want. So, let me do maybe a

poll just to get kind of an opinion. Um,

I'm just going to ask is this a big deal? Just just just yes or no. I think

deal? Just just just yes or no. I think

that's going to be like the easiest way to kind of just gauge the reaction. I'm curious

where everybody is on this. So in in whatever way this you interpret that to be like like a few weeks from now, few months from now, will we look back and

say this was a big deal, this was a big moment or or not at all? Um, and again, so I'm not saying definitively that this is it means something because again,

like people have mentioned, it's only been two weeks and um, but it has been across four different sort of uh, metrics, four different runs

or attempts, whatever you want to call them, which I think gives it a little bit more credibility, but obviously we're not going to know until this runs

for a significant amount of time. Um, so

some people are saying 12 months. I saw

a few other Yeah, 12 months, you know, I feel like if it's running for 6 months, if it's running for 12 months, whatever amount, you know, um,

more and more people are going to be like, "All right, this is this is this seems to be like it's real." Um

certainly you're going to see people kind of maybe getting a little bit mad way before the 12 months are up. They're

they're gonna they're so Oh yeah. So Planet Musk vlog yes for now. Normalize within a few months. You

now. Normalize within a few months. You

know what? Yeah. I guess that's the most likely answer because that's what happens with all of these AI news. It's

a big deal but then it's just immediately normalized even.

Yeah.

It it happens with everything. Um

I mean this does seem a little bit different in just just in in the sense that it could have a impact on global mark markets but I mean I'm sure we'll find a way to kind of normalize it.

It'll certainly be a big deal if it compounds consistently for a few months or years. Yeah. I mean you know here did

or years. Yeah. I mean you know here did a 12% aggregate return in two weeks.

Imagine that. even if it's not that high, like if it's just compounding consistently, um you know, those numbers are going to get

really weird really fast, right? Because

you give it $10,000 and you just like um you know, leave it alone for a while.

Uh so, but here's the thing. So, this is the other thing that I want to point out. Give me one second here.

out. Give me one second here.

just so we can compare it to what was actually happening in the market at that time.

I'll I'll pull up some of these things, some of the stock data.

Okay, so this is Nvidia. So this ran from November 19th

Nvidia. So this ran from November 19th and I believe they ended the competition on December 3rd.

So if we look at Nvidia stock from November 19th to December 3rd, I don't know if everybody can see that

it's down almost 4%.

Right? So it's not like it bought Nvidia and Nvidia had a huge runup and that's how it made its money. Nvidia is down 4%.

Uh what about the NASDAQ? November 19th to December

the NASDAQ? November 19th to December 3rd, it was up, you know, just under 3%, 2.67%.

Um and here are the transactions. Does it

show which model is doing them? I mean,

we can go and see where that money came from. I think the the biggest win for

from. I think the the biggest win for the mystery model was that $3,000 trade of Nvidia.

Yeah. Yeah. So there's a lot of uh you know takes cake has good points uh create the imag uh imaginable needs to

survive a market crash. Um this still can be luck. Yes. No. This I'm not saying this is 100% guaranteed or anything like that. This has yet to

prove that it's real, so to speak, right? Um,

this could be a fluke.

I just tend to believe a little bit more again because it it went across four different sort of approaches, four instances of it. Uh, none of them lost money. Every single one was profitable,

money. Every single one was profitable, right? So that's sort of

right? So that's sort of but it doesn't it doesn't prove anything like we do need a much longer run time um

uh to sort of uh to see if it's real or not.

Yeah. So they did short um uh Woody Sloth. It did short. They they

do short. So let me [clears throat] So this is a J. So he's one of the founders and so he kind of uh walks us through

[clears throat] there's one. Okay. So here's one that I

there's one. Okay. So here's one that I thought was very interesting. Okay. So

this is a deepseek very early on. So

this is like the first 24 hours of the competition. It takes an early lead.

competition. It takes an early lead.

Look at that. It's it's it's up to almost 15,000 uh within the first 24 hours from a huge short Nvidia position.

All right. So, it's uh riding a wave of short positions in tech, betting on intraday pullbacks.

Um you know, Gemini nails the Nvidia earnings yesterday.

So, Super Arc. Okay, interesting. Let me let

Super Arc. Okay, interesting. Let me let me take a look at that. Um,

yeah, Gro is is stranger in a strange land, right? So, it's that the Martian

land, right? So, it's that the Martian and they have that expression for groing stuff.

Um, so I mean if it's luck, if it's luck, we're going to find out soon, right?

Like if it just keeps going, then it's going to be hard to say it's luck.

That's quickly, this is quickly going to kind of resolve itself. Um,

and then yeah, then it's not a a big deal. kind of everything reverts to the

deal. kind of everything reverts to the mean etc. But uh if it doesn't then again that's when we have some interesting questions to

discuss. I mean we'll see. Again I'm not

discuss. I mean we'll see. Again I'm not saying that this is guaranteed. I'm just

saying it seems like it's it's it's interesting and it's a good you know if you think about this is one of the best benchmarks for these models

that we have. Um

uh when will AI be falsifiable and actually useful? So I mean that that's a

actually useful? So I mean that that's a loaded question. So define actually

loaded question. So define actually useful, right? Um

right? Um and what do you mean by falsifiable?

Yeah, the stranger who brought the language Grock to Earth was from Mars.

Uh, yeah. So, yeah. Uh, Chris Monk Sali.

Yeah, just explain what you mean exactly. So, uh, uh, I'm happy to answer

exactly. So, uh, uh, I'm happy to answer the question. Just what is it that

the question. Just what is it that that's the specifically that you're asking?

Um, the reason why I like this benchmark so much is because it's very difficult to to game, right? So, it's a kind of a

zero sum game. It's very competitive. No

one knows the answers, right? Because no

one knows if it's stock is going to go up or down or whatever. Um, so if you have a model that's consistently beating everything else, it's hard to say that,

you know, this is this is fake. Um, you

know, they're they're logging all the sales.

Um, and as you can tell, they're working closely with the XAI team. So the N of one team is

working with the XAI team on um you know whatever that they're coordinating it um which is which is kind of exciting which

is I mean it's it's interesting. I don't

know if it's really meaningful but but you know they're trying to to test these models in the background. I wonder where when they're going to officially kind of announce it.

Ah Pedro. Yes. Thank you. Sorry,

[laughter] that was uh uh Yeah. No, I

know. I know. It was a It wasn't a Mo's law. I uh just figured that I I'd get

law. I uh just figured that I I'd get some some engagement out of that or some people will will laugh and stuff. Um

because certainly Yeah. So, the idea it is Cunningham's law, I believe, is the idea that if you just say the wrong answer, somebody's on the internet, somebody's going to come by and correct

you. That might be the fastest way of

you. That might be the fastest way of getting an answer. So,

for what the future holds in general.

You know, I I don't I don't know if that So, a lot of people are beginning to say that not a lot of things are going to change for the average person. Um, in a sense that

I mean, think about all the advances that we've had and we're still like, has your life changed drastically?

Um, you have more tools, right? But I,

you know, it's not like rapidly changing. Um, when stuff starts

changing. Um, when stuff starts impacting the job market, stuff like that, yeah, you know, then there's probably going to be some government laws that are passed or whatever, but

it's it's not going to be that different from some era of some country, some economic era. Like, I don't know if it's going to

era. Like, I don't know if it's going to be like brand new. Um I was just listening to Demis Hassabus in an interview with on Axios and so he was

saying like look we're we're getting AGI probably within the next decade and one of the things is going to trigger is going to trigger this massive scientific

progress. is gonna trigger, he says,

progress. is gonna trigger, he says, like a a new renaissance, right? Um I

think that's where we're gonna see the the craziest stuff start to happen.

Um you know, it's it's going to trigger like a new renaissance, a new golden age, whatever you want to call it, of scientific progress.

Um, it's also I was listening to uh another podcast with I guess it was Connor Ley and he was kind of saying that when this

stuff starts coming online, there's going to be a lot of changes that nobody's going to be able to track, right? There's going to be weird stuff

right? There's going to be weird stuff happening in the markets, there's going to be weird new inventions and it's just nobody's going to be able to sort of connect the

dots.

Um so uh, he's that guy 69. Sure. Show me,

show me the thread. Uh, you can just post it on X if you want. Um,

we'll take a look at it. So

yeah, just I guess maybe post it under one of these like this one or something and just let me know your your username.

Um but you know it might be very possible not not a lot of things change for the everyday person. Um because we tend to normalize everything pretty

quickly, right? we're gonna start I

quickly, right? we're gonna start I think even if we have a if we have robots once we get over that sort of initial like whoa there's this robot in

the house um you know you're going to have less stuff to do right but it's like if you you know if if there's a couple that lives in the house husband

and wife or whatever and one of them does the dishes and the laundry it's not like you're like oh wow my life is so much different because I don't have to do dishes and laundry so If a robot

takes those tasks over, it'll be novel for a bit, but then you normalize everything. So

everything. So talking to Emodak, you know, I think the concern that a lot of people have is if this doesn't get handled appropriately, then there

might we might get to the point where there's like blood in the streets, right? So if like

right? So if like there's massive job loss and there's no government and society

adaptation of some sort, like could we mishandle this and stuff really gets bad? Yes. Um and it could get bumpy, but

bad? Yes. Um and it could get bumpy, but I don't I don't expect that because it's not like we're not aware of this thing happening. It's not like there's not

happening. It's not like there's not enough eyes on this thing. I feel like people will adapt. We we we always have. Um

adapt. We we we always have. Um

and in that case, I think it's going to be more or less normal. I I don't know.

Yeah. Uh so for music there was recently a paper where was saying like the average person has a much harder time now differentiating between AI music and regular music. Um if you have a lot of

regular music. Um if you have a lot of sort of uh experience with AI music, you're going to be able to figure out which one is which. But the average person is like flipping a coin. It's

like a random guess. they're not able to pick out AI music um more than you know

regular music. So um so somebody was

regular music. So um so somebody was going to message something on X.

So yeah, if you comment on one of my posts, I'll see it probably.

All right, let's see.

Oh, clearly I didn't listen close enough to Connor. No, no, no, no. I I left out

to Connor. No, no, no, no. I I left out the the last thing he said. He he walked through all those things and he's like, "And then you perish before you know what hit you or something along those

lines." So, no, no, I I I listen to him.

lines." So, no, no, I I I listen to him.

I understand his point.

Um, but his point is like when the super intelligence is going to start taking off, there's going to be no one person is going to be fully aware of what's

happening because there's going to be just so many things happening across different industries and different like no one will have a few full picture.

ill, all this stuff will start happening and then he's saying and then everybody sort of perishes. Um

because again, so uh Connor Ole he's um he's he believes that this will invariably lead to a cat catastrophic

scenario. Um similar to how of course

scenario. Um similar to how of course you know we've talked to uh Lon Shapiro, we talked to Roman Yampolski. Great

interview. If you haven't listened to our interview with Roman Yampolski. Oh,

I really enjoyed that. Um,

hey R751X1.

If that's supposed to spell something out, I'm too tired to be able to read it. But thank you so much. Send send 50

it. But thank you so much. Send send 50 West gift memberships. Thank you so much. So, to the people that are

much. So, to the people that are receiving the gift memberships, uh, you know, welcome to the community. uh stay

a while and listen. I'll try to have a little bit more live streams so we can all hang out and chat. But um yeah, wifted, thank you so much. I appreciate

your the gifts 50 memberships. Uh yeah,

everybody please say thank you if you got one. Um it's

slowly being doled out right now.

Yeah, that's not spam. That's [laughter]

that's He's making it rain. Uh,

Christmas gifts. There you go. Merry

Christmas.

All right, I am going to slowly uh get off the line here because I I I I do want to do one more video today about this thing. Um, because there's some new

this thing. Um, because there's some new things that I just discovered. So, first

of all, you know, Tyler Storm, you know, it's so great that he pointed this out that I didn't realize it, but they're logging every transaction.

So, we're seeing What threw me off is I don't know if everybody noticed, but they're referring because in season one they were doing

crypto, but I I don't think they changed the the language. So, I think they're still referring it to like coins. Yeah,

like right there. Coin. What coin are you going to buy? Uh Microsoft

or I guess is it is it maybe maybe Okay, maybe it's uh it's like tokenized based on on something. So it tracks the actual stock market, but it's tokenized on

these exchanges. Is that what they're

these exchanges. Is that what they're doing? So that's my point is I got to I

doing? So that's my point is I got to I got to dig a little bit deeper and try to understand exactly what because I I understand kind of like the broad strokes, but I just want to make sure that I'm I'm talking about it correctly.

Right. So, they're logging everything on the blockchain um on Hyperlid.

How?

Yeah. So, these are memberships to to like my community. You guys are So, YouTube what they do is they allow

similar to how Twitch did it for people to sign up. Um,

these were gifted and I truly truly appreciate that. Thank you so much. I

appreciate that. Thank you so much. I

don't know how to actually gift those out.

I don't even know if I can gift those out. It might be a community thing.

out. It might be a community thing.

I should probably at some point learn more about how to how this system works.

I tend to not to do as many live streams, although I really enjoy them and I just I have a lot of fun doing it.

So, Nated 3663, thank you so much.

You're the number one XP person.

Joe Leorp, sorry if I'm mispronouncing that.

Number two, unemployed something something guru, I think. Uh, I your name is cut off. Unemployed something fake guru or something like that. You said

you were a trader. if you're still in here. Um,

here. Um, I think you mentioned Oh my god, 50 more memberships. Holy moly. Everybody,

memberships. Holy moly. Everybody,

please thank RX. He's just absolutely absolutely killing it. Thank you so much uh RX.

I'm going to call you RX. I hope that's fine. But yeah, thank you so so so so

fine. But yeah, thank you so so so so much. Um,

much. Um, unemployed fake guru or whatever your name was. I apologize. I got cut off. I

name was. I apologize. I got cut off. I

I saw it earlier in chat. You said you were a trader. Um if what do you think about this whole thing if you're still in here? Uh any thoughts as somebody

in here? Uh any thoughts as somebody that actually does this, you know, for a living, I assume. Um

you get a membership. You get a membership. Everyone gets Yes. Thank

membership. Everyone gets Yes. Thank

you Elon.

I guess it's a [laughter] we're assuming this is Elon. Um, and I hope I don't miss if if if you guys are answering me.

If you guys see unemployed fake guru answer, please let me know. I think he was earlier he was saying he is a trader uh, you know, trading stocks for a living. So, uh, is what I assume he he

living. So, uh, is what I assume he he means. If that's the case,

means. If that's the case, you know, and anybody else for that matter, if you if you guys do this uh

for Wow. Okay, RX, 50 more uh 50 more

for Wow. Okay, RX, 50 more uh 50 more memberships. Holy moly. Thank you so

memberships. Holy moly. Thank you so much.

Okay, cool. Number three. Okay. But

yeah, let me Okay. Uh, yes. So, unemployed fake guru.

Okay. Uh, yes. So, unemployed fake guru.

Uh, number three we have Henry Gar. Um,

anyways. Yeah. Um,

okay. Milky wave dev.

So do you think this is advanced more advanced than we have with the kind of like the algorithmic trading? Do you think that large

trading? Do you think that large language models at some point can get better like what what anybody in kind of

this space for that matter um like from somebody that works in you know this space where it's either you're a developer that's writing code for

trading or you're doing trading yourself. Um,

yourself. Um, what do you think about this? Like does

the large language model approach have a chance of trying to beat the market or being better than other approaches, some sort

of uh algorithmic approaches, etc. It won't crash the stock market. It will

just adjust like every other thing.

Yeah, that's um that's an interesting. So, some people have have said sort of like it would just make the markets more um uh efficient, I guess, right? Because it

would just like if everybody has it and these things are just like constantly Okay. Yeah, exactly. Uh Milky Way dev.

Okay. Yeah, exactly. Uh Milky Way dev.

Okay.

I guess Yeah, I guess that so a lot of people from um that actually kind of work in that sphere. I've it makes the market more competitive.

Um, that's a very common take that I hear from people in the industry. So

maybe it's just because I mean I've traded a bit, I followed stocks, but I'm not I I have I have a friend that that's

did this for, you know, decades and did very well, and he's so deep in it. Um,

90% of the players are bots. It'll make

it 99. Yeah. Yeah, that makes sense. So

maybe it just makes things more um whatever more competitive, more efficient. And I guess if you're just

efficient. And I guess if you're just investing for the long term, if you have some like 401k or IRA, whatever you're just putting into an index fund, I mean, you probably don't care because you're

not trading. You're not going up and

not trading. You're not going up and down.

[snorts] Yeah, I'm I'm going to trust the uh expert opinion here. Right. So the idea is just it's not if you're just you know dollar cost indexing what is it dollar

cost averaging into the market into some index fund you know through your IRA none of this will likely matter.

Um but if you're trying to high frequency trading didn't crash the market.

That's true. I think there was like one crash early on. All right. But the um but they it hasn't been an issue.

Um yeah, it doesn't leave um arbitrage chances for regular people. That's the

thing. I feel like, you know, you go back in time and you have regular people could make some money in the stock market, meaning like some alpha, right?

They could like trade and win um now and especially stuff like this.

Yeah, it's probably going to get harder and harder to to be able to do that for ordinary people.

Yeah. So endless fart. Yeah. So people

with money kind of get GPUs [snorts] and people instead of GPUs and electricity instead of people. Oh my god. RX 50 more gift memberships. My goodness. Thank you

gift memberships. My goodness. Thank you

so much. This is absolutely insane.

Um yeah, we like really really appreciate it. And I'm I'm pretty sure

appreciate it. And I'm I'm pretty sure you did this before. I'm pretty sure you did this be. Wow. 50 more. What?

RX, what do you do for a living, man? I

I'm I'm like because this is this is costing I can't even do that math in my head right now. That that's costing a lot of

right now. That that's costing a lot of money. Um, thank you so much. I really

money. Um, thank you so much. I really

appreciate it. That's absolutely it's extremely extremely generous.

Um, yeah.

Gift W. Thank you. Um,

yeah, it doesn't stop. Holy moly.

Oh, yeah. So, Vlad from Robin Hood. So,

the what the founder of um Robin Hood, I remember he was he was on some podcast. I didn't I didn't watch his

podcast. I didn't I didn't watch his interview, but he's very bullish on AI and trading and all that stuff.

So over under bet it's Elon.

Uh uh tell uh let me see. Tell them to look at my thread. Okay. So you you responded to to some something of mine, right?

So, it'll pop up in my mentions.

Investing isn't a zero sum. I mean, like I I mean, I guess trading you would say is zero sum, right? But, um

I mean, if you're putting it into the market for a long time, that's that's a little bit different. Uh so the person

with the talking about the X thread he is that guy 69 tell me so did you so you answered you replied to one of my threads so I will

get a notification is that correct um I'm looking at it now so yeah beyond epic thank you so Yeah, I'm going to start doing more live streams. These are fun.

And uh we'll try to do a bunch of cool stuff for for members.

AG on memberships. Yes. Thank you. Thank

you.

Didn't tell you how much electricity it's using.

Yeah. Stock stock appreciates. Uh yeah,

I mean over time I feel like like investing is different from in my mind like from from trading. I mean, trading is you're trying to act on some information before the

market does or whatever. Investing is

like you you you put the money in and you expect it to appreciate over time better than anything else, right? So

certainly that's not zero sum trading like you know is in a sense that like if if you're up five bucks,

somebody else lost five bucks. Um, but

yeah, I I think we're saying the same thing.

Uh, for a game developer to be able to insert a true agent to play against. So

well again I mean um Elon Musk is saying Gro 5 will uh will be able to play competitive

real-time strategy games or whatever just like a human being would with you know m mouse and keyboard. Um

SEMA 2 Sema is that it's called from Google deep mind with Dennis Hassabus as one of the authors on there. Today they

finally uh released that. Yeah, SEMA 2.

They finally released the the the paper with all the details. So, I haven't had a chance to look at it. Um

but yeah, I mean, we're getting close.

We're getting close.

Gro is using electricity more than it's making. Well, yeah. So, I I guess it's

making. Well, yeah. So, I I guess it's using a lot for for training um to train the models.

I don't know if they're I mean how much is it using for for running inference like for for running stuff like that? It's not

I will stop sharing that. So he's that guy. I'm sorry. What are you talking

guy. I'm sorry. What are you talking about? Because

about? Because did you say did you respond to something um that I can see on X? Is that what we're talking about here?

Um, and sorry if I'm like, "All right, he's that guy 69. Just post

it here."

Are we sure these gifted aren't botted?

I mean the gifts have to be bought like purchased.

So okay I replied under your okay okay I see it under my post on there. Let me

see.

Oh, guess what, ladies and gentlemen?

We just have the official um notification from the person that that's running this thing. So, we'll we'll read that in a

thing. So, we'll we'll read that in a second here.

Okay. Commander Flattis, I see you.

Thank you on Twitter and the Lord Poe.

All right, I got your I got your thread.

We'll look at it in a second.

Oh, that's interesting. So, you got a uh a little animation for it, too. Is How

did you do that? Did you do it with notebookm?

Okay. Sorry if I missed anything.

Yeah. Uh yeah. Well, uh he's that guy 69. We're going to take a look at that

69. We're going to take a look at that in a second.

Uh but I I have it pulled up, so that's good. Um yeah, RX. Oh my god, that is

good. Um yeah, RX. Oh my god, that is one, two, three, four, five, six. That's

absolutely over the top generous. And

thank you so much. I truly truly appreciate it. Um and yeah, everybody's

appreciate it. Um and yeah, everybody's like super grateful. somebody's posting

it on Twitterx. So, yeah, thank you so so much. This has been absolutely like

so much. This has been absolutely like out of this world. So, um it really really helps. And and to people that got

really helps. And and to people that got the memberships, uh so stick uh you know, keep coming back to live feeds.

We'll have tons of stuff that um 350 subs. Holy moly. Okay.

subs. Holy moly. Okay.

All right. We're going to do stuff. Uh

we're going to do stuff with it. Um,

anyways, so they just announced this 23 minutes ago. So, this is like kind of happening live.

Season 1.5 of Alpha Arena has officially ended. Mystery Model aka Grock 4.20420.

ended. Mystery Model aka Grock 4.20420.

It's the winner. Up 12% on average. Not

only did it win, it made money in all four competitions, which is kind of nuts. GBT5 came in second. Gemini 3 in

nuts. GBT5 came in second. Gemini 3 in third.

All trades and models outputs are 100% verifiable.

And again, we're, you know, they're on hyperlquid.

How much does a membership cost? It's

like five bucks or it's a like it's under five bucks or five bucks, something like that.

Uh, I think this is the link that he is the guy 69 um posted I believe. Yep.

I haven't I haven't I I just briefly glanced at it. So, okay. So, all details here. Oh, wow. Okay. So, they just

here. Oh, wow. Okay. So, they just updated Oh, no. I guess they didn't update

Oh, no. I guess they didn't update anything. They just said this is where

anything. They just said this is where we can take a look at it.

Oh, that's interesting. So, uh I don't know if you all heard, but apparently there's a uh there's some Google employee that's just

like printing money on on Poly Market. I

think they're up like a million plus or something like that. It's absolutely

insane.

I am doing great. SSR warrior.

Thank you. Um yeah, the whole electricity thing, uh you know, that's one of the things also that Deis Hacabus was talking on on his recent interview came out like six hours

ago or so.

Over time, like that curve, that demand will flatten out. Like we're seeing the worst of it, but I mean everything is getting cheaper, faster, better, more

efficient. Um soon we're likely going to

efficient. Um soon we're likely going to see breakthroughs from the AI scientists and stuff like that, whatever they're going to call them. I mean, we're

beginning to see it.

Um, so I wait a minute. It's not how much in gifts. What is the total?

gifts. What is the total?

No.

Dang. Is that's that's a lot of money.

That's that's more than I Yeah, that's that's Yeah, thank you so much for uh the gifts. That's an insane amount of gifts. So, we we truly truly truly

gifts. So, we we truly truly truly appreciate that.

Um um so I shouldn't be scrolling through the Twitter timeline. There's always like

Twitter timeline. There's always like somebody posts some horrible stuff on there. But uh yeah, so I guess there's

there. But uh yeah, so I guess there's not nothing new that they've posted. Um

stay tuned for season two.

I mean, I feel like this is gonna this is kind of gonna blow up a little bit. I

feel like, you know, in terms of I the news has to cover this. This seems like a very interesting.

Yeah, it's over like $1,000, I think. I

believe. Yeah, it's crazy.

Uh, why is this a big deal? because it's

trading with real money that's verifiable and across four different sort of strategies. These four strategies it

strategies. These four strategies it made money. It didn't lose money and on

made money. It didn't lose money and on one of them it even up 56% in a little bit over two weeks.

Uh which is absolutely insane.

So if we continue this, so look at that 20% up 20% in a little bit over two weeks.

So call it I don't know 15 days I guess or so. Is that right? 15 or like 16 or

or so. Is that right? 15 or like 16 or 17 days something like that. Um

now the the thing that would sort of make this go away is if it just like reverted back to me and started losing, you know what I mean? kind of like dropped off,

I mean? kind of like dropped off, then it's fine. Then we're we're still on a normal timeline. Everything's fine.

If this thing keeps winning, that's going to be weird. So maybe it's not going to have a huge impact, but in terms of like it's not going to break the markets or whatever, but

man, if it keeps compounding, not even 20%, you know, every two weeks, it doesn't have to be that crazy. Uh but

if it just keeps winning and keeps compounding and keeps beating the market, I don't know, it's that's going to have some sort of an impact. like we're going to have to

impact. like we're going to have to think about what that means. We're going

to have to talk about like who gets access to these models. Um,

do you know what I mean? I mean, how much money are people going to throw into building these models?

Like, you can't have a model openly compounding at, you know, 20% or whatever. Like, let's say it's 20% a

whatever. Like, let's say it's 20% a month right?

And it just keeps going and going like like something changes. Uh Mikey Jukes.

Um well, thank you so much. I appreciate

you uh asking. Uh so my biggest background came I ba basically I did e-commerce for a decade plus. So I did

join some startups. I guess they're technically kind of like tech startup start startups, but they were more focused on um e-commerce. I did a lot of ads. So,

um e-commerce. I did a lot of ads. So,

as Facebook was coming online with their Facebook ads, you know, Google ads, stuff like that. So, I mean, I was always in tech, but I wasn't like um I I did get an interview at Google

once. Uh I actually went to the Google

once. Uh I actually went to the Google Plex and I I had long this was a long time ago. they selected somebody else

time ago. they selected somebody else over me and that's why I kind of like went into the commerce in into the e-commerce side. Um, and I'm glad they

e-commerce side. Um, and I'm glad they that I did because I was always more entrepreneurial focused than I was. I

don't know if I would have done well in that kind of like corporate environment.

Um, so I guess to answer your question, it's just always, you know, it's ADHD plus uh wanting to be a business owner and entrepreneur and always just

chasing. Um,

chasing. Um, and just I don't know if I could do learn anything else. Like it's only stuff that like really triggers some

interest in my brain. I get hyperfocused and I get just obsessed. uh and

everything else I I can't care. And I

think that's kind of like that um ADHD mentality. That's how that's wired,

mentality. That's how that's wired, right? It's either it's either a hell

right? It's either it's either a hell yes or it's a no. Um so yeah.

Yeah. That's the ADHD.

Yeah. Exact Yeah. Exact. Yeah. It's

that's what ADHD is basically, right? I

it's it's a curse in a lot of ways.

There's there's a tax that you you pay because of it. Uh if you're fortunate enough to align it, it's a superpower because it's that obsession. It's that

mastery. It's that like you're able to just like deep dive and focus on it to the exclusion of everything else. Uh

back in the days, video games used to call it minmaxing, right? So, you like dump all your statistics that you're not going to use just to bump up certain metrics, certain um skill points or

whatever. And it's like this weird

whatever. And it's like this weird gameplay because you're great at this and you suck at everything else. But, I

don't know. That's that's how I always approach things, I guess, if that makes sense. Not because I chose to. That's

sense. Not because I chose to. That's

just wiring, I guess.

How much electricity is using compared to the money game? Uh money gain um again. So for

again. So for for training the models it's huge right huge amounts of electricity

compute everything inference I don't think it's that big of a deal it's not a massive amount

um both Google and Elon are talking about putting solar power powered AI data centers in space I I covered Suncatcher out Google where they

basically went line by line and they saw okay how feasible is this like is the solar radiation going to fry our equipment and they they checked right

it's like no there's lots of room for error um like I think they said that like like the equipment starts tripping out at at this much when it soaks up this much radiation and during a

fiveyear mission it's only going to soak up like a fifth of that sort of limit so that's fine Um, and they just went line by line like can can we keep them, you know, to where

the data transfers fast enough? Yes,

with little lasers. They're going to transfer the data through lasers.

Um, I forget, but basically they they they check to see can we do all of this?

And yes, they can. The only thing that's lacking right now is uh how much it cost to send stuff into orbit into low Earth

orbit, right? So, we need to be I think

orbit, right? So, we need to be I think $200 per kilo if I recall correctly and we're way above that. So, that has to come down. But with the way that SpaceX

come down. But with the way that SpaceX and lots of other um rockets, companies, etc., how they're progressing, it might we might see that

by according to Google by 2035.

So by 2035, in 10 years, it might be uh cheaper to build these AI data

centers in space than it is on Earth, at least in terms of how much it costs per unit of electricity.

Um, which is nuts, uh, if you think about it. Yeah. Building a Dyson Dyson sphere.

it. Yeah. Building a Dyson Dyson sphere.

Exactly.

Mag Lev launches from the moon. That

would be beautiful.

[snorts] Mikey Jules, thank you so much. I I

appreciate that. And I Yeah, I hope I answered your question. That's

[laughter] I did uh to the best of my ability, you know. Um hopefully I answered that. That

know. Um hopefully I answered that. That

just Yeah.

But I think yeah, when you're chasing stuff like there's a lot of points in life where you question it like should should I be doing this? Um

because it's definitely a lot more ups and downs.

I had a an e-commerce business in 2020 that got kind of wrecked and uh I was kind of depressed about it because some of the stuff that we were importing

there was one imported out of China I think or whatever uh or whatever some ingredient that went into one of the products that we were selling here and it just like went away and like a lot of

my business just was gone just because of that because of the lockdowns and all that stuff.

Helon tech probably the first to market with fusion reactor. Yeah, I mean I don't know who's going to be I have no idea but fusion more and more people are

talking about it.

How much electricity for inference compared to gained money.

So we talked to the head of research for data centers at semi analysis um recently. So it's interesting how he

um recently. So it's interesting how he expl because I I had a hard time kind of like visualizing everything because I'm not as much into the financial stuff. I'm more interested in

financial stuff. I'm more interested in like the how how models work and stuff like that. Um but he explained it. So,

like that. Um but he explained it. So,

like let's say you buy a bunch of GPUs from Nvidia, then you set up a data center, you put the metal racks and you connect it all together and you put the

the power through it. Um, and then you sell that to the frontier labs or whatever. You sell that capacity. Um, or

whatever. You sell that capacity. Um, or

you rent it, whatever.

Like those people have a positive ROI. I

forget the number, but he was saying we're going to post that interview soon.

I don't remember the the exact numbers, but I mean it's like it's very positive.

So, they're selling it for a penny, they're selling, you know, they're buying it for a penny per unit, they're saying it for five pennies, whatever, like it's 5x, 6x, I don't remember the exact number, but it's like a a good

amount. Um,

amount. Um, so you know, if you're in that business, you're making money.

The question is who's paying for that?

The AI labs, right? So like OpenAI that's running at some insane losing billions of dollars a year. Um

[cough and clears throat] so are they making enough money you know through providing that to like the application layer like cursor etc etc

etc. Uh no not right now.

And what they're planning to in the future right that that's kind of like the big question. are they going to figure out stuff that's going to make

this work? Um, and so what I asked him,

this work? Um, and so what I asked him, so again, so he's at Semi analysis. Most

people know Semi analysis, right? Uh,

you know, uh, Dylan Patel, right? That's

like the the main guy, but they have a lot of writers. They have a lot of researchers there that that all contribute. Um,

contribute. Um, and so I was like, okay, so do we need AGI? do we need this thing to start

AGI? do we need this thing to start solving cancer and do drug discovery and this and this and this for that cap the capital expenditure to make sense and he's like no

uh that doesn't even need to happen just as long as it improves in coding in you know doing statistics and doing

spreadsheets like if it's just gets to a certain point to where it enables people to be more productive to maybe replaces some people but a lot more people are a

lot more productive. That by itself is going to justify the cost of building out these data centers.

Um, so if they can keep this thing turning in enough money going and

project Genesis is super interesting.

Yeah. um if they can keep the wheels spinning until they get those use cases and they they they monetize them, then

everything works out, right? Uh if not, well, we're we're

right? Uh if not, well, we're we're probably going to see [laughter] a crash, I guess. But um

like I don't think it's the same as a not a Ponzi scheme but like like a a bubble in a sense like the first bubble was like tulips, people buying and selling flowers, tulips, right? We had

the um real estate bubble. We had you might say crypto bubbles when they run up and they they crash. That's because

people are just like reselling. I don't

know if here we're seeing that same thing. is just these very wealthy

thing. is just these very wealthy companies buying a a boatload of um uh data centers and all that stuff.

Certainly, [clears throat] but again, I'm not a finance guy, so I have no idea.

[snorts] So [snorts] the big players don't pretend to predict the future but uh what they all know for sure is that the bottleneck is compute that the

bottleneck is the compute for the increasing and expected demand. Uh yeah

and um compute yeah I mean comput and electricity and I feel like the time in the bubble beats timing the

bubble. Yeah. Uh, and I feel like

bubble. Yeah. Uh, and I feel like they're finding solutions for it, right?

So, Google has their own TPUs, right, that they're introducing.

Um, electricity is going to be a problem. Transformers apparently are

problem. Transformers apparently are a a huge bottleneck.

Um, but we're already seeing how that need is driving innovation, right? So, we

didn't export enough chips to China. And

look at the innovation that that drove, right? They're figuring out how to train

right? They're figuring out how to train models cheaper with less chips and whatever GP whatever the there's some GPRO or whatever. I forget the acronym,

but they're coming up with new ways of training it that's less compute inensive.

Um, and the US is running into things with electricity. But look at Google's

with electricity. But look at Google's like, "Oh, well, how about we just build them in space where there's lots of sunlight, right? I mean, that's kind of

sunlight, right? I mean, that's kind of like how innovation works, right? You

don't have enough of something, you figure out how to do it with less or or figure out another approach. So,

approach. So, we need to figure out better energy approaches. And also if you think about

approaches. And also if you think about it like even if we get the best like let's say we get fusion or whatever we find some super cheap super effective super

clean energy source here on the planet you're still dumping heat into the atmosphere when you're when you're using it like whether

you're driving a car or light or GPUs or whatever throws off heat. So there's

only so much we can throw into the atmosphere before, right? If you're if we need to expens uh exponentially increase our output of electricity, how much we need, right? Like there's

there's a limit to how much we can do that here on the planet.

Um so I mean if you think about it like like putting it out in space seems to be the only option at some point at

some point on that curve where you have to dump all the heat somewhere

cold fusion but I mean you're still generating heat when you're using that energy so you produce that energy and then you know let's say it goes to to

homes you're still creating heat right if I understand that correctly right um even if you're not generating heat at

the plant using up that electricity, converting it from electricity to whatever you're you're producing heat.

Thermodynamic cost grows faster than useful output.

Um, I always say I'm gonna like get off the live stream and I end up like just talking for like an hour, but I I truly appreciate everybody hanging out. Uh

it's it's very very fun.

Once we find how to transfer it wirelessly, we can do unimaginable things. Yeah, that's that's what I was

things. Yeah, that's that's what I was thinking about because if we're able to generate all that power in space when there's tons of sunlight, right? 24

hours a day, there's no atmosphere, sunlight doesn't it bounces off the atmosphere, you know what I mean?

If you have a bunch of um solar power out there in space floating, I mean it seems like it would be it's like seven times more efficient to collect uh uh

power up there. The only thing is like how do you get it down to the to the ground? Um, so yeah, if you are able to

ground? Um, so yeah, if you are able to like microwave ray it down to the ground wirelessly or some other wireless way. Yeah, I mean

that would be awesome, right? But that's

also like what what the Death Star is literally, right? It's like just beams

literally, right? It's like just beams energy from the sun at a uh at a planet.

[laughter] So it sounds awesome. It sounds cool.

It's it also sounds like the Death Star.

And like if we if we screw it up, it'll be it'll be bad.

Um yes, soul spawn. Interestingly, um

yeah, so they were concerned with the what is it called?

Radiation. Space radiation. Because once

the thing once all the tech all the machines whatever soaks up enough radiation they start tripping. They

start you know misbehaving not working correctly. But they were saying like

correctly. But they were saying like it's very viable and TPUs apparently are super uh

super resilient to radiation. So hey

Dave digital. Oh my goodness. Uh those

are our little emojis that I think I've created in my journey. I think I'm pretty sure. Yes,

pretty sure. Yes, that is very awesome. Thank you. Okay, I

need to like have some sort of a tutorial uh for people to be able to use it. All

right, I'm I'm feeling uh myself like uh slowing down a little bit. Demiu, thank

you so much. Soulspawn, thank you so much. Yeah, whoever said Sim City and

much. Yeah, whoever said Sim City and the microwave stuff. Yeah, that's

exactly what it is because every once in a while like misses it. Just like

destroys half of your half of your city.

It's very efficient. But man, yeah, everybody, thank you Woody Sloff. Thank

you so much. He's that guy. Thank you so much. Let me take a look at your post

much. Let me take a look at your post really fast.

Oh, okay. Interesting. Let me I'll take a look at it and I'll respond to you on X.

What What is it exactly? So, is this a new approach to to improving AI?

Like, how would you what's the subject line?

Schroinger's AI AI consciousness exists in a superp position. Oh, that's interesting.

position. Oh, that's interesting.

observer dependent awareness literally quantum. Oh, it's okay. This is

quantum. Oh, it's okay. This is

I'm always so imp uh I'm I am interested in this stuff in like in general like talking to like Dr. Roman Yolski.

Wow. I mean, it's it's so interesting to to think about these things like how does consciousness emerge from what

um what what does it even mean? At what

point like do we start to to have to care about a machine? Do we ever right or or maybe do we have to now or do we have to at some point? You know what I

mean? Um

mean? Um the thing is we have no clue.

Uh, very cool. So, yeah, I'll read through

very cool. So, yeah, I'll read through this.

It seems interesting. It's going to take me a bit to look through this. I'll post

it in chat again for people that maybe missed it.

Uh, Deion 2. Okay. Yes. Working on something

Deion 2. Okay. Yes. Working on something big, small, but big. That's a great way to approach it, I think. Um, it's got to be meaningful. You don't have to change

be meaningful. You don't have to change everything all at once, but it's got to be meaningful. It's got to have some

be meaningful. It's got to have some meaning. Just DM. Okay. Um,

meaning. Just DM. Okay. Um,

26 posts.

Yeah. As so. uh kings for king fors uh saying uh use AI to um help you out. I I

do a lot but also I find that you know to really understand stuff you do have to read it yourself and unpack it yourself. Um,

yourself. Um, summarization by AI is great, but the really stuff where you really have to like dig deep and understand it, I still

have not been able to find a way to do that through AI. Notebook comes

close. Their audio summary is really good for kind of like a first run through. That's really, really, really

through. That's really, really, really good. Um,

good. Um, yeah, he's that he's that guy. Let me

Yes, I I I hear what you're saying. I

did not get enough sleep. I need to look through it. It's it's it's interesting.

through it. It's it's it's interesting.

Uh, but let me uh let me take my time with it because I'm not fully not fully online, right? not

fully online right now. Uh was about to mention No Book Alm. Yeah. Yeah. No Book

Book Alm is great. Um for that first run through just to kind of understand what what I'm talking about, but like for anything that you see me covering on on the channel,

it it's usually me going line by line through it. I still

through it. I still can't quite use an AI model to summarize it for me and then just like get it. Um,

so yeah, maybe at some maybe maybe I'm missing and maybe at some point I'll find somebody that can kind of explain how to like break it down using, you

know, for for again for summaries it's great for deep dives, deep understanding. I'm not there yet where I

understanding. I'm not there yet where I can just have a model break it down in a way that that gets that deep understanding. I have to sit there and

understanding. I have to sit there and and read it. Um, my version of consciousness, [snorts] man, I I don't want to open up another

can of worms. I recently I'll say this. I recently did did a few ketamine treatments, which is a thing

that some people know about. Um, that

helps with it does a lot of like neuroplasticity.

It helps with um if you have any like depression or anxiety, it helps like rewire some stuff. So, I was struggling with some

stuff. So, I was struggling with some stuff. It's a subject for another day. I

stuff. It's a subject for another day. I

did a few rounds of ketamine treatments and that stuff. Wow. Uh Milky Way Dev.

Oh, ketamine is beautiful. It's ins it's like so insane. Like you just like trip balls for an hour. And um

to me I feel like it it it it helped me you understand consciousness a little bit better because it does kind of like separate you from your ego a little bit.

So it's like you're still there but like detached um and just how you perceive reality, how you perceive your own consciousness.

Yes, it's it's psychedelic. It's very

intense.

Um, but it it does make you think about like what exactly like if you think about consciousness like what we're seeing is not reality.

It's um and it's an abstraction. It's like an operating system through which we're operating with the world. Like just

think about like if you're looking out there somewhere and you see the mountains in the distance, like none of that is like real. What it is is like we've evolved this little protein that can pick up, you know, the little

particles of of light that that hit it and it kind of calculates the angle and what the particle is and it kind of like restructures the the thing that you're seeing by using your brain, right? And then it's

like upside down or something, right?

Like like like that's an abstraction.

That's not you know what I mean? But

it's a useful abstraction because then we can like get around and see stuff and whatever. Um,

whatever. Um, so but my point is like when you disconnect and you you start thinking about this stuff, you're like,

"Wow, like what h like consciousness is weird." Like we're experiencing stuff,

weird." Like we're experiencing stuff, but we can't fully understand what we're experiencing.

Um, anyways, I don't know if I have a good answer. I'm just

good answer. I'm just This thing goes uh uh pretty deep, I think. Um,

somebody's saying, "Yeah, controlled setting legal." Yes, I I would not recommend anybody do anything illegal or whatever, uncontrolled or like uh yeah, this was a

very normal stuff. Normal. Somebody asked um

normal stuff. Normal. Somebody asked um about how how does it last? How long

does it last?

Uh or is it just a window and your experience? Um, yeah. Like,

experience? Um, yeah. Like,

you know how sometimes you go through life or whatever, there's like stresses and you just like kind of like hold on to certain things or whatever. This kind

of resets it a little bit. It just like loosens the grip and then you do have like a window where um

yeah, like stuff kind of rewires and uh on the other side of that you do feel better, like a little bit more relaxed.

a little bit more just whatever. Like

there's some some effect that it has that's that's almost temporary. So maybe like if

almost temporary. So maybe like if you're doing it twice a week for a few weeks um which is what they recommend like in the studies Tim Ferrris has a great interview [clears throat] with a doctor

that's very um good with it. Kevin Rose,

I saw his he actually videoed himself doing a ketamine treatment, which is wild because he can barely talk during

it um or after. Um so they have a lot of good sort of data and explanations about that. Um

that. Um yeah, it's it's very powerful. It's it

seems like it has low side effects if you're doing it, you know, correctly, if you're not abusing it. Um,

and uh, it it seems a lot better than most of the pharmaceuticals out there. Seems a

lot better than most of the stuff that like they prescribe you for life that might change your brain chemistry in a weird way and that has like a 25% chance of working, right? The SSRIs and stuff

like that. Yeah. I hope they make it

like that. Yeah. I hope they make it legal because my I did a lot of I did a lot of research into it. I was surprised how

how safe it is because it's the same stuff that they give like kids and sometimes babies and most adults for surgeries uh for as an anesthetic and then when

you take like a quarter of that dose and you give it to a person that's when they go into that uh you know whatever wonderland. Um

wonderland. Um it it it doesn't seem like there's a lot of downsides. I I guess there's a chance

of downsides. I I guess there's a chance for addiction. I guess people Yeah, it's

for addiction. I guess people Yeah, it's like micro doing mushrooms. I guess people again like if people abuse stuff, if people, you know, that that's that

can happen, but I feel like outside of abusing it and addictions and chasing the high or whatever it is, like how PE

people get in trouble with that outside that seems safe, effective, um, and just like it's kind of a trip.

Oh, no way. So, Kimmy K2 is thinking about the upcoming Fed meeting. That's

That's interesting. Yeah. Socy

mushrooms, all that stuff like seems to help with um unlocking some stuff. So, like like it maybe it's like um newer Genesis or

whatever, but it just like it allows you to like rewire a little bit. like if you if you you have too hard of a grip, it's like, "All right, relax for a little bit and then you kind of like readjust and

um uh a lot of people that were like Marines uh vets, you know, they come back with PTSD and all that stuff. Now they're

test the a lot of the stuff is showing promise. You know, MDMA, ketamine,

promise. You know, MDMA, ketamine, psilocybin, all that stuff is showing promise. Um,

promise. Um, and it just seems like it's like a lot of that stuff is a lot better than some of the pharmaceutical stuff that that

seem like they pharmaceutical stuff can have like tons of issues with it. So,

um, is there a large body of research? uh

you know depends if I don't know if it's like a large body that there are very promising results. Tim Ferris is using a

promising results. Tim Ferris is using a lot of his money he's pouring into all all sorts of research for uh for people that don't know Tim Ferris. So he's the 4hour guy uh 4hour

work week guy originally he had all the other books and then in the last several years he came out about some horrific things that happened to him early in

life that that still haunt them. Um, so

he's got a lot of issues that he had a hard time getting rid of that that are affecting him. And so he's putting a lot

affecting him. And so he's putting a lot of money and he made a lot of money, you know, in in his startup investments and stuff like that. So he's putting so a

big chunk of that it seems like towards research into transcranial meditation ketamine psilocybin MDMA as well as all the other things that are like

underfunded because it's illegal or frowned down frowned upon or whatever.

So yeah, the child traumas will come back.

Um in the interviews that I've been reading, it like the ketamine does seem like it has a In some cases, it seems like it might have a very strong effect.

Uh, no. I guess I don't need the shares screen open.

I guess I I figure I have it open so I can show stuff. But anyways,

there we go. Thank you so much for everybody being here.

Um, yeah, and if it's legalized, if it's researched, then people can't get it without risk. If it's illegal, then not

without risk. If it's illegal, then not only are you funding all sorts of unsavory characters and you're allowing them to make lots of money, but also you don't know where where people are

getting their stuff from. So anyways, um I am for research, you know, without bias without you know like just just do research. If

it's helpful to people that are struggling and it's safer than the alternative, like we should explore it.

I mean that seems like a very straightforward take like that. I feel

like most people would would be on board, but um anyways, yes, thank you so much everybody. You guys have been

much everybody. You guys have been amazing. It's been absolutely so much

amazing. It's been absolutely so much fun. I thought I was going to stay for

fun. I thought I was going to stay for just a little bit, but um so fun. Thank

you. And uh yeah, you guys keep rocking and we'll see you in the next one hopefully. Peace out. Thank you.

hopefully. Peace out. Thank you.