Dario Defends Himself, Datacenter Panic, AI Doomer Trap, Senate Toss-Up

Key Takeaway

When evaluating a technology or policy, separate the headline fear from the incentives underneath it. Today, pick one change affecting your work—AI, automation, a new tool, or a company policy—and write down: who benefits, who bears the cost, what evidence supports the claim, and what would make the transition feel fair to affected people. This simple exercise improves judgment and helps you communicate change in a way that builds trust rather than resistance.

Episode Overview

The panel debates Dario Amodei and Anthropic’s approach to AI risk, arguing over whether AI oversight should be government-led, industry self-regulated, or kept transparent and open. They also discuss backlash to data centers and automation, the economic insecurity fueling populist politics, and the role of affordability, housing, wages, and trust in the upcoming U.S. elections.

Full Transcript

Transcript of Dario Defends Himself, Datacenter Panic, AI Doomer Trap, Senate Toss-Up from All-In Podcast. Auto-generated from episode audio; may contain minor errors.

Okay, everyone, please. please return to the number one podcast in world. Your favorite podcast, podcast allimp. We had it again. huge, huge list meetings with me. Key questions for saxophones Chimat Friberg and yours Jake House's boyfriend. Topic one. Topic first. Let's go right away. to the point. None No more empty talk, guys. . Dario spoke up. Obviously he is listening. the All-In podcast and was forced to answer for a little discussion Gavin Baker last week, that many people inside, or some people supposedly, perhaps, inside Anthropic, they think that they will become one private company in the world at a certain time moment.

Some people from Anthropic School Douglas answered absolutely untrue. Whoever said this to Gavin, he lies. And Dario decided that he would come out from cover, and he entered the chat with epic two-parter essay on regulatory pots. He said, quote: "I know abbreviation Silicon Valley, where regulation is equal to regulatory the excitement that equal to concentration of power, but I always thought this is too simplistic a picture of exchange messages." He said, quote: "I don't I agree that my messages were disproportionately negative. They were approximately equally balanced between risks and advantages.

It required a certain responsibility. Quote: "On today the most accurate criticism companies that engaged in artificial intellect, including anthropic, is that we have not yet fulfilled their big promises to benefit to the world. This is completely our fault. Most often the phrase is quoted: "That, what will work is this actual treatment cancer, not shiny marketing issues ". Sax, uh, you Are you talking about anthropic? a year and a half already. He He didn't answer. Gavin comes out and he suddenly, uh, very talkative and answers. What are you?

you understand from that fact that he is not answered you, and then answered Gavin, and then we Let's get to the details. He doesn't want to be with you. quarrel? Justice for the sake of J.K., he answered me. I have I mean, I don't want to. to get stuck, but The reality is that he did not answer that what Gavin said. He I could easily say, what Gavin said , he said, it's not true , as Shalto did. He did not speak out. on this occasion.

Instead, he was really waiting for question regulatory delight. So, on In my opinion, that's it. that's what attracted him here : he is a little irritated by that we constantly we accuse him of regulatory enthusiasm. And this is it, What am I talking about? last year, that they participated in complex scheme regulatory delight, based on intimidation. And these are the ones moments that he actually reacts. And I I think he doesn't. I like that. he feels that it contradicts him reasons that he dishonest and adheres to all these rules with the right reasons.

AND my answer to this like this: look, there is no doubt that Anthropic was extremely aggressive in the pursuit implement the desired them regulatory frames as on level state, and on at the federal level. This is regulatory delight. Now, Look, I think what Daario could be sincere in their beliefs. I am not I say he does it. for financial reasons. Nevertheless, he seeks to capture state apparatus for your own sake political order daytime. Just here nothing to say argue. I don't think so. that there are a few more things, which can be said from about this.

So let me first go to point number two. He says he doesn't carry responsibility for negativity around AI, that he was balanced in in his image of AI. Oh, really? You you expect us to there will be amnesia about the fact that about one and a half a year ago? You are very specifically said that within one to five years 50% of employees entry level knowledge will be lost work. And it wasn't. just some kind of speculation about idols in the podcast. This was what they designed and strengthened to full-fledged media campaigns.

You can you remember that President Obama even retweeted this, and many others officials Democratic Party did the same. Moreover, this scheme repeats with everything research anthropic blackmail. Remember, he has team in the company. They are supposedly studying consistency. They create it very artificial research, where they are over 200 times offer this model, until finally will get what they want the result that will attract the attention of newspapers. And again, my the idea is that that it's not just someone, who praises podcast about risk AI leveling.

It they are actually create research, develop it, and then they promote him to create they need headlines. And Daario, his sister, who is company president , and other leaders 60 minutes, having waited breath, untwisted this thing. And from the moment publication of this a lot of research people were reviewed him and said: " Look, this thing "very artificial." Including with the British security institute AI, in essence, said, what was this thing created in conditions high pressure, and they designed result. So these the guys created campaigns to force people to be afraid of AI.

Why the public So afraid of AI? They they think that he will take their workers places. They think, that he became Terminator. This is it. those fears that Anthropic instilled in the blood of the media, and no company, and you must say, therefore no founder- CEO did not do more for promotion of these fears, promotion of these fears and intensification these fears than Daario. And the fact that he doesn't see it, makes it, I would said, insincere or someone who has delusions. I am leaning towards the last one.

So, this is point number one. And I really I think that almost everyone who is read and followed carefully after reading, felt the same. Daria, you have really a blind spot regarding one's own behavior. By the way, can we to talk about it a minute or not? Yes, let's do it. it. Just so develop what you You say the problem is what? we have now, actually quite problematic. If I had to take three different action vectors and now combine them, in you would be very good scattered, disorganized and, frankly speaking, false narrative about artificial death intelligence, and he actually established , and now it is growing, and in as a result of that is happening?

I was really shocked, when I saw how governor of texas Abbott issued this decree regarding processing centers Texas data and limited their growth. Josh Shapiro in Pennsylvania just did the same. Axios reported that Republican the party sent out memo to the leaders of various companies that engaged in artificial intelligence, in which essentially saying: "Hey , stop harassing people, because we are about to we will lose these decisive race Republican party to the Senate in Ohio, largely because of this issue." So you have Dumerism , and now you have political pressure from both sides, which, in my opinion, opinion, extremely dangerously.

And then the third leg of the chair– this is growth profitability. By the way, Ministry finance just now doubled the ransom bonds. If put it all together , as simple as random observer, I would said we did a few really stupid casual mistakes in the United States. And the code of everything this is that, what if you combine everything it's together, growth profitability limits money supply from a point view of how people want to invest in risky assets, because it is risk-free the bet continues grow.

It leads to that people are less likely invest in everything data centers. The processing centers themselves data are closed. Where does that leave companies operating on the front line? They are at the greatest risk. Why? Because they are not have investment class. They are in the worst situation from the point of view of their balance sheets. They are not have the opportunity access to computational resources needed for success. So even to the extent you are growing actually make it worse my problem is that meaning that you more and more is needed more capacity to support this growth that you don't you can finance, because all the negativity set people up against you.

So everyone roads lead to big sign questions and moments, when I scratch my head, when I think: how are we found themselves in this situation? But was it Is it a coincidence or is it was management risks with good intentions, whatever it may be was, we are now in a very dangerous situation which we could honestly avoid. I completely agree. Fry, your thoughts on this movement against processing centers data or regarding How did Daario leave? directly and wrote double message? As Do you think he is sincere?

is he delusional or something like this? Why don't we steal it? Manet? Yes. And if we let's put ourselves on manager's position advanced laboratories, and this the leader sees model capabilities, which we may not we see, because we see every day we work in corporate environment, creating things, doing things benign, in a harmless way . And they sit there, testing models under pressure and understand what is ways to do jailbreak models, to make them do what they don't must do. AND then you can use them for things like meadow development weapons, creation viruses, creation infectious diseases that does not exist today.

You can you use them for development cyber weapons to make them do things on the internet like- here is the hacking bank accounts and stealing money people, etc. And you you can encourage them to do things that can widely to influence society, creating fake media, fake videos , fake images , spreading them through social networks, cracking algorithms. I I mean that algorithm X is public. So theoretically agent can read it, copy, understand and find way to manipulate the general public. AND now you realize, that during creation these tools, where there are many benign positive, world's leading progressive programs , you see all the bad ones things that can to happen to them.

And you you can continue try create systems and continue try create ways, to prevent this. You can create own code that checks each incoming request, which checks each API call. And yet somehow the network agents may be used for performing certain actions, and they can start to do what we don't we want them to did. So let's say you Sit there and eat. the head of the company, which creates this wonderful technology. You completely believe in her opportunities and benefits for society, its benefits for economy, but you don't can you understand how keep it under control and how stop social engineering, bioweapons and cyber weapons.

What are you Do you do it differently? Good. Good. So, you Do you remember what this is? an exceptional way that is exceptional by. I like it. There is one thing you must do if believe that this is true, so that even more improve this thing. You can't confuse tokens thinking model. Then What do I mean? If only compare models with open source with models with closed border, we can't understand what it is tokens of thinking. That is, you give him hint, you get something in response, but everything that is in the middle, confused.

Claude It's confusing. Good, they do the same open AI models. And I would say that if I I was really afraid, first of all What I would do is I would try to find way to debunk these tokens of thinking, so that a third party could actually to legitimize what you are talking. Because What is interesting to me is that I haven't seen the same one. Dumerism from none companies that develops models with open source, including Nvidia. And the thing is because with models with open source you actually you can to see how they think.

So you you can see theirs discrepancy in real-time time. And so instead of this we must essentially agree with theirs interpretation tokens, which we do not we can see which can only see they, and what they can to interpret as, as they want. And not condemning, Is that correct? wrong, I would said give them a chance to make the right one thing. Do this damn thing accessible to others people. Let someone another who also has to a certain extent prudence, will sit in room and play with you.

So let me to express this. So, Saks, Demis put forward an argument for benefit of regulatory FINRA-style systems, where industry participants, who understand technology, fast developing and cooperate with commercialization, actively viewing systems and models through each other industry consortium who hires and appoints scientists for work on these things. And they just do this instead of so that it belongs and managed government and created a certain scheme regulatory delight. But this You were pretty for a week. against the FINRA system in in their tweets. It seems you supported her the first time when Demis her first time published.

Me it seems that Demis's proposal, to which was subscribed Daario, Sam, Elon, You are against it now. But this could be a solution, if there is real here concern, considering Chimat's remarks about what the industry can to control alone one. Let me to clarify. I believe that creating a new one regulatory agencies for AI, developed by FINRA model, which is quite sleepy regulatory agency that writes rules for stock exchanges brokers in enough mature industry, which not dynamic at all, is a terrible idea. And to things, FINRA widely perceived as protective only existing ones players.

And talk. with any startup founder who had to to deal with them. They believe that FINRA exists to protect large banks from failures. It's just not very good model for adjustment. Maybe NFA, National futures association, which is also one of these self-regulating organizations on financial markets, there are the best alternative because it is enough dynamic market. Good. So, what you you say, very interesting , namely you mentioned the term self-regulatory organization or SRO, and there are many different models, and one of them it can work, and that was my opinion.

I I think that the most appropriate for these circumstances will be something on like the MPAA, which does not report at all before the government and was developed the film industry in response to the code " "chem." And the government said: " Get together, guys, or we will have to to do it for you." AND MPA then advanced standards, which then led to more obsessive, harsh government actions. So this, by the way essentially, a model that, as I said I could to support. But is this and the model that supports Dario?

No. Dario and I think that Demis wants FINRA, which, according to is essentially a regulatory agency with development of rules, which not only publishes standards, but also conducts preliminary model testing. Good. And then it will report to by the government, and then will officially be enshrined in law. AND the reason why everyone they say: "Elon supports", consists in because I talked to Elon about this. Elon is not supports FINRA for AI. He could support MPA for AI. And this is mine. opinion: Demis wrote something very eloquent in this podcasts, Sash, with I say from the beginning about self-regulation MPA.

I understand. But Listen, here's mine. opinion. Demis wrote something very high equal. So, when you you say something like : "Well, everyone supports "the idea of ​​Demis." There is many versions of the idea Demis. Some of them can be normal, and some bad. And I have to mean that if this FINRA model, it is not will work. Long having talked to Dennis about this, I I think that the general structure—this is an SRO in correct structure . However, FINRA often used as example SRO simply because it is the most famous SRO.

But you You are right. I think that there is a mechanism by which by which this the thing must to function, which differs from any other SRO. But the idea is because the industry unites, and, on According to Chima, they considering models each other. They control this. They receive independence, but not controlled by the government . And well, that's what I I think it is important. AND to make it clear, this idea of ​​FINRA for AI, Dario demonstrates that it not an SRO because it is not self-regulating.

It regulatory organization. AND part S—this is full fig leaf, designed to to hide the fact that what's new regulatory reporting agency before the government and is engaged in previous issue, testing and approval models. I call it DMV for AI because I think that all these models line up in waiting in line conducting tests, and then release, and this terrible for us slow down. You ever- Have you ever been to... This is a terrible idea, and it just repackaging what Dario always wanted, namely FDA for AI.

Now he says it's the FAA for AI. Do you know how much? time taken by the FAA and FDA approval of new products? Yes. Yes. China. Advantage China. Yes. That's right. China will win if we will do it, and this will become clean regulatory enthusiasm. You should take a look. on a Chinese missile, which they planted yesterday. You know, she all-steel, and was held good analysis of that how quickly they were able do it from the moment the last explosion. Then they completely made steel, and Elon is still waiting for FAA approval for your steel Starship.

You could watch behind this in China, where their regulatory the authorities allow them move a lot faster, and in as a result they are starting to recruit real speed in space race. AND then, separately, many biotechnology moved behind border with the USA from the same the reason you indicate, namely the cycle FDA approval is as follows long, and the stages are so slow people down. AND China completely reviewed my rules regarding clinical tests and the third phase, and then, having completed this, everyone biotechnological investments in the world now moving to China.

The fact that Daario could want and called , particularly in their recent blogs, to creation regulatory FAA-style aircraft, and what is it, in my opinion, the right model for AI, fully discredits us, and yes, China us deceive, because they are not subordinate everyone their companies with this restriction. Now, this can return to your steel master for a second. What I think we need there should be no DMV for AI, doesn't mean I against AI security. So, let's let's talk about some things that can be make.

So, according to first, I think that good thing these Frontier Labs are taken seriously to safety. Sam just published a post, It seems like yesterday, in to whom he said that actually they are are going slow down the release some of their models Frontier, because believe that their security package is not corresponds opportunities. Good, This is responsible. To things, do you know, why, in my opinion, Are they doing this? Ago what are they carrying huge and sufficient responsibility for usage responsible measures, because if they If they don't do this, then will be subjected all kinds of lawsuits of responsibility for product quality.

Look what's up now happens with Meta. On they are being taken to court, because supposedly theirs the product led to depending on social networks, and now on them are suing over a trillion dollars. And this happens immediately after they lost the court jurors in Los Angeles for 6 million dollars, where the plaintiff filed for them to court for creating problems with perception of her body. So we see massive lawsuits about responsibility for product quality against the big ones technological companies such as Meta and Google, and if OpenAI and Anthropic are smart, and I I think they are smart, they are not want to be next.

So these companies have enough reasons to protect against responsibility, and not just fast bring out your fashion product. Good, security is part of what they should get it now. As the industry should coordinate their actions with each other? There are proven and reliable ways do it. Companies in the industry publish articles. They publish technical articles. They publish blogs. It good ways exchange of advanced experience. They are visiting symposiums and conferences and perform on them. They ask comments within open standardizations. There are ways to by which the industry is very productively cooperates on issues security.

But which one common denominator everything I just did said? They all open. They are not secret. And it does. these judgments contested to to the extent that Jamat I talked about that before. what we need see the logic that is behind this. And I against these companies met secretly, whether within the framework of something like FINRA or something another to install standards that are not transparent, which are not are subject to appeal. And you know, What will come of this? Regulatory capture and anti-competitive behavior. So there is ways to do it, which will protect the public, which are open and transparent.

This is what what should we do strive. You won't see that this will happen. And I will say Why do you? Again , technical details have meaning. If suddenly competitor A, B and C will be able to see every route and gap, which generated by your model as an artifact transition from incoming tokens by the end of the week tokens, you are essentially reveal a significant part of your recipe. And I just don't I think that if Anthropic does not want to, even if I am an API client, share with me tokens of thinking from my company, they definitely don't will share them with OpenAI and anyone else in room.

Good. Good. Let's just Let's do one more step. So, let's assume that Dario is successful, and JL, I'll put you question because I took upon himself Moderator duties of this segment. Let's. Yes. No, I have meaning that I...No, Sax has...this...this... Sax log cabin...Ya wanted to hear, so let's just... No, honestly, it is Saxon Willis...I I wanted to hear yours. full position. Let's just say that Dario is successful. There is certain regulatory scheme, which, in fact, as Sachs notes, slows down the industry, slows down development models, and Chinese Frontier Labs achieves significant progress, start jumping forward, and they get this RSI, it's recursive self-improvement, and their models become the best in the world, and they only become better and better and better.

Jacob, let's start with you. Good. And then we will go around topic. Good. The best in world moderator. What are the consequences of this? will have for American worker, for American business and for American consumer in a world where China has the best artificial intelligence and all processing centers data? What takes place in our world? Which consequences? Because everyone constantly hear: "Oh, China will win. China will win." Whose is this? Are you worried? What is it does it really mean? I I don't think China will win.

I think that it will be, you know, not there will be one person, who will receive the entire prize . I think it will be gradual improve. So I I don't think it has to. value. I, you know, I thought about why. actually this is happening negative reaction to data centers. And it was a little shocking to see how Abbott and Shapiro, because these two people were the largest supporters processing centers data, right? AND if you twice click on this, they just want make sure that you do not connect to networks.

And, you know, water problem, about which we talked about here, not too serious, but I think she is was very revealing. And I think that you and I talked about it on rear channel: or they both do this is because their voters, people who are for they vote, really care about these processing centers data, and they want stop them. So then you will have to to work in in the opposite direction. Why do they want it so much? stop centers data processing? Why Americans are so against it artificial intelligence now?

And perception AI, of course, already existed in advance formed by science fantasy. So we it was necessary to work to overcome vision Terminator with Skynet. But Dario told everyone: "We "We'll lose our jobs." Elon said, "You don't will have to work ", it was more positive point of view into the world, but he said that Sam Wallman conducted research regarding the universal basic income, and he told everyone that you won't have to work. There is a UBI that says such things to people who must to work to to earn a living, and who do not have savings, and which are watching, you know how administration for the administration does not can tame inflation, regardless from whether they Democrats, or Republicans, and then they observe, as a minimum salary it just remains as it is, and barely fluctuates until the stock market is beating records, and then all we, all of us, leaders technological companies, we become extremely rich .

And you wonder why people are now accepting socialism, and why in you are Abbott and Shapiro, who say: "You know, that I", and "and Roano, worst of all, they say: "I I have to be anti-capitalist, I I have to be...I have to "slow down." "It because, you know, they should have felt something relief, but this didn't happen. So what? people who don't feel that they have some kind of power? They are looking for strength, and if to look at what happened, they exacerbate guillotines, and, you know, this metaphor has wake up people.

Or that's murder general director of United Healthcare, whether it is people who they throw incendiary bombs to Sam's house Almana, we need wake up and to see why people So upset? Ago that they cannot survive. They have problems with accessibility, with salaries, and what they can they do? They can vote for socialists, and they can actually stop centers data processing. And exactly This is what it's about. Nobody, none ordinary the American doesn't care AI safety. This is not the one. the discussion they should be discussed.

Discussion for them is why I not the rich, but everyone else -Yes? I would simplify it a bit. your argument. I I really think that you came across something important. But the point is because, I don't think, they are thinking about..." That's exactly right. I I actually think that they are thinking about it just like about that, who won for the last decade, this generation technological oligarchs who won, enough unpleasant. They are not ambitious, and they are not those people on whom can be compared. They not Steve Jobs, even in the worst sense.

AND now they are watching on this next generation and say: " Wait, we let's move on from billionaire, ten-billionaire, seven billion to present trillionaire." "And this those guys. These guys even worse than last time. Exactly This is what happens. So I don't think that's it. as specific as you you say, Jason. I I think it's vibrations. And the vibrations are terrible. AND they look around around; they are as follows: " If we just we can do it slow down, maybe another one will appear a group of people, and they will really have intention to do the right thing, and that's will make me feel better, and I I won't be so reluctant.

to hate the fact that What are you about to earn? "trillion dollars." « That's exactly the feeling. AND people standing on front-end show from this reason, not deserve trust. This is what it is. . Sax, let me. to give you mine question. And me too I just need it. need to hear someone's answer to this is a question. Good. I will answer. your question. I I would like to... Yes. So mine question again yes: if will take place deceleration regulation with the development of AI in the USA, and China will overtake races that the consequences this will have for American business, for American consumers?

How is this? will actually affect them, will they win? What does this mean? What is it means? Help people to understand. So, for example, if I I'm sitting at home, okay, let me to show you counterfactual assertion. Imagine, that China won Internet. Imagine that all major internet companies were Chinese. I am not I can...I need to so that you...I need you so that you can imagine for me. Tell me, what...give me the scenario of what it is means. So imagine, that instead...I I know that people don't like big ones technological companies, but imagine, what if Google, Microsoft, Meta and all these companies would actually be...this would be BU, Alibaba, Tencent and everyone else.

I have on I mean, it's obvious that..." it would be much worse for the United States. Our the economy would not be so dynamic. She wouldn't be like that. advanced, and all this the wealth would go to China. There are also military, or, you know, double, application of this technology, especially with artificial intellect. Algorithms power all kinds of weapons, drones etc., so we will behind them. I have on mind, see: America will not be number one country in world, economically and militarily, if we we will lose AI.

That's right. simply. Do you agree? For example, what is your opinion about the world, American business , consumer, individual person, in a world where China has, quote, "SHI?" I don't know. What do you have? What do you mean by "one"? I I don't know what that means. . Good. So, I I'm trying to understand. What am I striving for? If you could have guessed it , that would be easy- consequences deceleration regulation. So, if we let's slow down AI, which the consequences this will have for people?

How people understand that this means? Yes. Unless we actually ban , and I don't even... you know how to impose a ban . So, I think this is just a useless exercise. Forbid open code, I think we have everything It will be fine. I think that the consumer benefits. I I think they will win. American companies, and I think what will win American economy. And the reason, why does it win, is that we we find out, I will show you are really an example interesting insight.

AND this is about closed vs. open. Good. Three years ago there were only models. Yes. And you were looking at some kind of benchmark, and there was a Freeberg ball, and you said: "Wow, that one, in Who has the higher score? "better". And then for the last few years it appeared that called " equipment", right? Yes? And equipment just wraps model, right? As CodeEx, or Claude Code, or Hermes, or OpenClaw, is equipment. You know, What's interesting about models? open source and closed code? When you take the model from closed code, it shows very well benchmark yourself.

When you wrap it with your own equipment , its capabilities are decreasing. When you take the model from open source and If you use any what other model from open weekend code, its capabilities are improving. What is it trying to you say? I think that we find out, and these the data will become a real tsunami during the next of the year that exist all kinds of ways using technologies from open weekend code that is significantly more productive and significantly cheaper than alternative from closed on weekends code.

If America won't stop us, us, consumers, business owners, from use the cheapest, the fastest and the best. If they will not stop us in this, we will win. At the moment we let's discuss which private American Will the company win? loses. I'm coming back. to the previous one argument. I don't think so. that Americans are so very worried about whether win or lose any of these companies with closed source code. And I I think you see it in politics. Therefore, Friberg, my clear the answer to you: until we will not do something draconian, like attempt to stop open weekend code, American the consumer wins.

IN I am bad for you. news, Chimot. Coming soon prohibition of open source code. They are not going to call it that. You know what they are will they say? They will say ...that we just must apply those the standards themselves open models that and to the closed ones. Good, this is how they do it are taking a step forward step by step. Allow explain to me how actually works regulatory delight. Therefore, first of all, to you need to get this regulatory apparatus. Okay, Daria. wants FDA for AI, but in he is not there sufficient political support for this.

So instead they use this FINRA Trojan Horse for AI. They they call it self-regulation. Actually, that's not true. but in any case, it helps them to move from the dead points. Now they created an organization , which establishes standards. Now in them is previous model testing. Then the pressure increases, to codify this into law. So this is What happens next? AND then they say: " Look, all these standards should apply equally to everyone models". "But here in Why is there a problem? Open and closed models, as you know, technologically are different.

Return to Daario's testimony in Senate, it seems, in 2023 year, where he said, that open models, if they are enough developed, there are dangerous. And why? Are they dangerous? Ago that they cannot centralized control, to track, cancel and remove, as closed service, such as the one that he places. Good. I I understand. Please, let me just finish, but then you can answer. Yes, you you can answer. Daario basically says they are open models are dangerous because of these characteristics that they have. And these characteristics unchangeable.

Technologically they are not can be changed. After how you you release it open model in the world, you are not you can roll it back and you can't exactly to control how people it is used, because they launch it on own equipment. Daario says exactly what this makes open models are dangerous. So they are going to make the body on standardization will say, "Well, we have to set standards "AI security". "By the way, Daario and, uh, OpenAI, they will finance all of this. They will provide all calculation.

They will stand behind this. They will be the ones who will coordinate actions with government officials, because, honestly saying, people in government have no idea how monitor, to control and establish AI safety standards. Technologically it is them far from being able to. So they will turn to these companies and will say: "Tell us "how to do it." "AND therefore, standards will be established, and then it will be very a simple matter justice to say that standards must to be used as to open ones, as well as to closed models.

Open models are not can answer requirements are the same, and gradually they will be disconnected from this. I understand. but American overnight economy suffers paralysis. Why ? So what you're saying is regulators never did nothing ok, let's go we will answer with our own with your mouth. So let's let's take it again example Coca-Cola. Coca-Cola— this is extraordinary profitable, good managed business that generates money. He already exists, I don't know, over 100 years on now. He there will be many more hundreds if left his fate.

It global business. Right? Let's take Pfizer. Same situation. Well managed business, global business. And now you eat them you say that if you don't tell them, hey, where would you didn't work, you must follow American rules . Okay, that's one thing. Or inside America you will be told that you must to adhere to our rules, and other governments will have others regulations. Now what? such... What to do with Coca-Cola? Introduced slower, more expensive, less effective, less valuable model in United States. Of course they are. will do, and their business in the USA is paralyzes.

They will have the best prosperous business abroad. They will invest overseas. And you you will see how straight foreign investment to the United States will grow sharply. And you you will see how straight investments in others countries of the world will grow. By the way, you saw this in China, when China did everything these steps to paralyze Chinese technological companies. I think that this is the best analogy. There was a moment when China found itself on crossroads where large oligopolists, large technological the oligarchs tried seize power, and the tip of this spear there was a moment when Jack Ma said that very famous speech on the eve of Ant's IPO Financial.

What happened? Xi Jinping inflicted impact. They took full control. They said: "Now we We work together. Everyone who is worth something, will have a representative the Chinese government in to your council directors. We have golden voice. We we will work together". When you look at that happened to the direct foreign investments, and that What happened to theirs? economy, it fell from the abyss. It will happen here too if it will happen. So they can try, and everyone They will say: "Okay." AND CEO Koch will be redirected all your capital for the borders of America.

The same Fizer will do, me too. AND anyone. What about you then? Are you going to do it? Pick up ours passports so that you don't could travel? So I think that's it. will be very noticeable the way is simple destroy the United States, and you can do it do it if you want, but you will see how it is almost instantly plays a role in stock markets and in capital allocation companies. I agree with you. regarding all negative consequences if we let's ban open models, but this is not means that this is not It can happen.

Look at all negative consequences ban centers data processing. Centers data processing will amount to 3% of GDP. This essentially creates all the increase in stock market. But they don't vote. against this. They vote against, so that two people, whom they don't like it very much, became trillionaires. This is what they are against. vote. This is a big part problems. Jake, what Do you think? Well, I just wanted to add to this. We we already know that will happen to AGI Freeberg, because your initial the question was similar in an attempt to explain to the Americans who will happen when China will win.

We already we know what will happen, when they win. Products will become cheaper, and, uh, you get to them more access. If they will solve the cancer that they will do with it solution? They just will save it for himself? No. They will sell it Americans, likewise how they sell phones are cheaper, yes just like them sell cars cheaper, and we we will try them block. And then, that Americans will do, we will lose work through them. AND so I think it will be way to explain it to them .

But if we want to change the situation, it's actually simple convince Americans, that they will have a job, and that it will benefit them . The American people afraid of losing livelihood. This is the beginning and the end. AND you know, my experience in investing in zipline, Uber and a bunch companies with autonomous driving. And here, let's say, poverty. Money is not a problem here. . But, in my opinion, this will affect autonomy and jobs. Zipline just concluded partnership with Uber. I I invest in both of these companies, and they are going to make a million deliveries per day.

Then you see huge progress Why. They double number cars on roads approximately every 6 months. Tesla is about to switch to cyber taxi. And people in to your community feel that they have there will be no money, which will remain in their community, because these drivers, you know, even though called often, this money at least remained in the community. Oh, they were making money. not enough. Oh, they earn only 18.6 dollars 21 dollars on an hour, whatever it is there was no. But now we We take it away from them.

AND Where are these going then? money? Where are they going? money for a robo-taxi, a trip to Vimo, zipline, where is it all is going? They rising into the cloud, and then they return down in Kremniyev valley, polarizing even more wealth. And that's where you are. You'll see, I'll predict. this is now. They Vimo will be banned in New -York, Boston, County Colombia, and they will limit them licenses. We must find out how to explain to the Americans that we are going raise the minimum salary.

Something that when I say that we should increase minimum wage fee and to focus on this, people say: " This is madness." But when we talk about experimental tax cuts, people say, "Oh yeah, we we spend experimental tax cuts. We we will change the cost capital gains. Perfectly. Let's let's try and "Let's see what happens." But nobody wants to try to raise minimum wage fee. Nobody wants explain to people where will their work be in future. And exactly here, you know, whoever expressed this opinion, I think it was something like this: who Are there any spokespersons here?

To us need spokespeople who will claim that this technology will give you can get great job. How not paradoxically, the who explained it best this, or these two, were, obviously jensen who said that artificial intelligence is not will replace you. This will be someone who works with artificial intelligence, and then Zuckerberg, who said, "I will give you a bunch of free ones tools". IN in his manifesto he directly said: "I will give free for you tools. I will give you opportunities. You get agents. Your life will become better." Daario and Sam didn't do it.

AND actually did a truly wonderful thing. Tsuk Tsuk did a truly wonderful thing, and just created this new one program, or almost training school frames, similar to professionally technical school, where people can to learn how get various work from laying fiber optics, performing new construction work with blue collar, which we have a huge deficit, and therefore in fact difficult to increase salary or something like that . This is a great program. Yes, yes, I mean, we need more, and how Zuckerberg suddenly became the best our spokesperson industries, as if This is a very low bar.

. I mean, this is ... Look, I have mean that all this it boils down to that the fact that Daario, not exclusively, but again yes, he brought the most specific figures and that's why I did it the most attractive and the most vigilant for the media. 50% of employees intellectual elementary school affairs level will lose work. Not far away in the future, but already in a year. By the way, we already crossed this annual mark. He updated the information? He said that I Am I wrong?

Not quite. Sam, by the way, said, and must give it to him due. Need to pay tribute Dario, and he should start talking about work. No, but Sam really expressed himself and said: "I was too negative. I was wrong." Dario doubled his opinion. As far as I know, he never gave up their statements. He idol. So, but I have to I mean all your thoughts, yes, I agree with you, that hostility and anger towards the centers data processing stem from fears about work, as well as other fears about AI, but it job loss is main anxiety.

AND there is none of this evidence. We have been for years we are living in an AI boom, and there is none of this evidence. We see only creation clean workplaces. That's right. Everything was vice versa. I mean, Listen, I had... I was huge. fears about this reason. AND I will also say something else. Business opportunities so wonderful. I have seven vacancies. So I I constantly say: Wait. Wait. This is important. Wait . Jake has something important here. advertisement. No, I Listen, I found out that if you can to convince your command to accept these tools, so many appear business opportunities, what are you after all hire people for working with people in your organization, who accept these tools.

So we created so quickly syndicate, created our programs that I I thought: you know, we need to add several people. So I thought that we we are going automate this, and we add. I'll just say more. one thing in common remark. Silicon Valley completely lost that what made her great. This was a place for weirdos, for people on periphery, for people, which really were idealists. And now it became something far from it. It equivalent to what There used to be Harvard. This is a place of accreditation.

This is the place you go come to to say: "Oh, I'm here" , you know, to Silicon Valley. I worked at a startup. You know, I got into YC". You know, when we arrived, we were looking forward to meet Bill Campbell of the world, Mike Morris, John Doors. These There are no more people. We wanted to work on Steve Jobs. Its there is none here. And in this break everything we did this focused on making money all possible ways, with all my might. And I think that in this there is nothing bad, except of what has now come the moment when we this needs to be better to recognize and try to do a little better.

And every time, when we don't we do, they give us a chance to become better, and we will not carry hatred. Hostility, which, as you said, David is building up, and that's it what we are seeing now. Yes. This is not no doubt. But Can I protect? Dario? Let me just to say something in forgive him, My earphone is out of order. in order. You said that want to protect Dario. Actually, I I think he is very idealistic. I am not I think the money is his.

main motivation. Just what is it? ideals? He wants create these new ones superintelligent machines love and grace, which will distribute all resources society, while people are spinning with your fingers. This is not the one. an idea that one believes in the average person or which attracts her. In fact, this dystopia. And exactly so many of them dystopian messages . And it is because of this idea, whatever not idealistic were his motives, that's why he doesn't inspires confidence. I have I mean, I really I don't think he's right for us.

lies. And what he actually is trying to do, it's to capture everything wealth for yourself. I actually think that he is trying to create this future, but this is not the future, which, in my opinion, wants American public. And to things, he must protect it. I mean, if you think that create singularity, and, let's see To be honest, he probably leads in this, in creating recursive self-improvement, he must protect this is the future. I see, like all these people they sing on EXO, something like: "Oh my God , Dario, thank you for You came, you know, thank you, who came down from the mountain Olympus and slept together with us on X and finally got involved in social networks", and it's like: no, he has it make.

You know, if are you going to take take the initiative regarding the creation of this crazy of the future, you must tell us, Why are you doing this? Well, he has more. speak. I think that This is a key point. AND returning to your opinion about what for him it's not about money. Seven co-founders this, as far as I am concerned I understand. It was documented, hmm, I think in a pile of press- releases, but they never about it they say. They donated 80% of their shares.

So, for example, 10% anthropogenic assets in some trust, which with two or three trillion, you know, 10% of them will be the largest, well, this will be the biggest democratic group. This will be the biggest, no matter how they used, but why don't they talk about the greatest trillion dollars for What's up with EA? What can go? wrong? Well, I mean, let's hope they direct him to something that really has a positive effect on humanity. But they, they, and only they, they only allow about 1.8% each.

This is a term of art. Yes. Well, then I think that a direct transition is that you are just me I think we can. move on after this, this is critical importantly. What do you need? to learn? You have AI, to speak. No, no, no. You have arrived. to plays, she is very simple. You have reached number one, you must write everything independently, that he clearly did, and you you know that Jensen did. Are you sure about this? I I don't know. Me it seemed like it was he.

It was very similar to me. on Claude. You think you know. Claude's voice? I have on I mean, he's the father. Cloyd. This is true. Actually Claude may sound like him, not the other way around. It could be the other way around. with causality consequential relationship. But he used exactly the same scheme action, which consists in because it is necessary exhaust the audience. It must be said a thousand words like this Zuck did, and then get ahead of problems. This is where leadership comes in. is collapsing.

Need to talk about that there will be more, and then never apologize and not to yield either one step at a time. Simply to attack someone who is attacking you. Fake news. I I just don't think so right now. that he can win dispute, being transparent because it sells fear to motivate the government to give him control and create this dystopian the future he is in believes. At least, that's mine. the essence " create a program for a generation tools". To come up with such ideas.

Do more like this. Friberg, the last, the last, the last, the last. I I think it's worth it. to think about whether Do you believe in recursive self-improvement. AND recursive self-improvement— this is a theory according to which at a certain moment AI model can to launch a whole bunch agents for execution of a large the amount of work, and they can to work together to create a new model AI, and this new model AI will be better than previous one. And in a certain the moment it all becomes virtually completely automated and self-evolutionary.

AI now constantly is improving. And this an idea of ​​what you have singularity, as some people call this, or you have ASI, artificial superintelligence, which arises from this abilities. But this ability is recursive RSI self-improvement. Anthropic stated that they not sure if this is will happen, but this perhaps. And they expect people from industries I work with talked, they think this is real, and such arise philosophical debates, because companies have a different view on the way forward, where everything really should be built by man, designed by man , and there is no such thing: " Let go of all this, let people separate from AI and let AI to develop", and this really like in the movie* She*, where the thing is simple takes off and is developing.

But if this happens, and if RSI is real construction, then you almost fighting. It's like A fool's errand. And this fool's errand in the sense that this whole the idea of ​​intervention human in stages model development. You are scaring again. people. Well, no, I don't actually, This is my opinion. If RSI— this is a thing, and by the way, RSI not necessarily leads to some dystopian result. It's simple. means that the models everything is improving faster and faster without people, without people necessarily designed, designed and designed the following evolution.

And then you have not necessarily a place for, oh, us just spent one big one model launch every 6 months. We we are going to contact state regulator or to go somewhere and get approval. There is just this constant police service and monitoring, which has to take place at a certain time moment. And the truth is in because everything you it is necessary that this to accomplish, that's what you don't need governments in the US to say " yes," or governments in China, to say "yes" or governments everywhere to to say: " I know that you are you going to say." You need satellites in space.

Do you need some... So you need a certain sovereignty, a certain space where you you can, and I don't know if it's in space. I am not I don't know if this is related. with biology and agreement which he had just concluded in Kazakhstan, and you there is something unique geography. This may be in Iceland. This could be a place where in the ocean people are are swimming. It could be in the ocean. But the point is that if RSI is a thing, you have there is no human process when you intervene to to control this thing.

So, you don't must have a cohort of scientists from advanced laboratories that do this manage. Theoretically, smart enough person or group smart enough people can to promote the RSI factory anywhere in the world or anywhere in space. AND your attempts ban centers data processing or attempts to ban or create a regulatory authority. That's why I I call it useless. matter, because for open internet and computing technicians needed only chips, power and connection. And if in she has chips, power and communication connection, it can work. And I don't I draw a dystopian a picture.

I just I'm talking practically. pragmatically, that all laboratories speak about RSI. So if RSI is where it will lead, then laboratories can go and create centers data processing in Iceland, in China, where whatever, whoever gave them freedom is wonderful question. Unless Wouldn't we be better off? Wouldn't it be better? to have systems where we have are these really there laboratories where we have there are these processing centers data located in USA, where we have over jurisdiction where they we have the opportunity monitor, to control and control them, not just let them to appear in some a random country or on some random territory?

So, I I think, my opinion is that we should involve and to keep the advanced laboratories, not try, yeah, you know, drive away their. And then here is the famous the image of waiting, but why human progress is being made not bad, and then recursive self-improvement, and it just comes out out of control. AND I will tell you why I I consider RSI legitimate, how close are we before this? I think this is what people would like know. As far as people around whom you know, honestly saying, until now, there is a huge problem with slowing down time in development of AI models.

And, and I mean that we usually we are considering something that requires linear time, as a certain number of calculations, to achieve something. Happening unpredictable changes that actually it is unknown which one will be the next stage, which the next step of change in terms of time. Time models are strongly change from point development perspective artificial models intelligence. And that's exactly it. so crazy that so hard to predict, therefore no one really can predict, when will the IAAF take place, when will it happen the next breakthrough, and then suddenly everyone They say, "Oh God, we're in singularity.

They just solved eight unproven mathematical theorems per week per with the help of some quite advanced models. Now it will be second front line model". So this is moves very much fast, but a lot models are a mixture expert concepts where you have such smaller ones models that calling each other to do something, and then you summarize it and you call it big model. And these smaller ones models over time look much more similar to those observed in biology, where there are genes, a gene creates a protein, and proteins interact very complex in ways that we don't we can completely to develop.

Networks proteins perform certain tasks. And I I think it's the same. evolution is likely will take place with models where they are broken down into smaller models that makes them much more efficient and more productive. And we we see a lot of this in because of how they develop many models. We we call it a model, but actually under it works a lot models that work together. So if models really develop like this way, then RSI is perfect system for definition the right path, that is, how many exist submodels?

As functions look like each submodel? As they call one one? This evolution, probably, will take place without human architect. She will be developed by using the system testing and deployment, and she will be completely automated. AND that is why this approach, which actually solves this problem, eventually becomes future development model development, so I think RSI, probably is the future . And when will this happen? again, it will be where there is a data center and power sources. I remember a few years ago everyone said : "Oh, general artificial intelligence inevitable.

General artificial intelligence inevitable". And where it? I mean, maybe we already It seems we are immersed in it. Okay, fine. But is he turned out to be Terminator, or wrong? No, that would be superintelligence. I have meaning that general artificial intelligence is the same smart as any man. Superintelligence is that, What are you doing? approaching, I understand. Free to talk about how far away it is not so. Or I think RSI can happen? Yes. I think what can happen many blacks swans. Very difficult to estimate the probability these things.

What do I know? exactly, that's what it is, as you say, I don't I think it's something new. state regulatory the agency will stop this, if this is so, therefore that it won't happen quite technically difficult. All that is will actually do it will concentrate power around this thing. It will limit the possibilities. You know when it's time reaches old age, let me to talk about aging. We just want to have this technology in this country, here's my point vision. And I just want to go back to that, what was I asking about earlier.

Why don't we we want China won? Because we we want the best laboratories. We we want the first breakthrough system was here. Well, Dario made this argument. Daario actually made this argument in in the context of attempts prohibit export chips, which is one of his hobbies, because Anthropic really stood behind a significant part political activities, aimed at sales restrictions Nvidia internationally equal. And the argument was that we we want to have a center data processing, full geniuses in the USA. And what after all, you know, for what will happen AI used, yes this is so that you could create all these agents who are scientists of the level laureates Nobel Prize.

And if we have thousands or millions such, and in China only dozens or hundreds, then we win everything. And that was the whole point. argument. But now we are actually we prohibit centers data processing. We have there will be no centers data processing, complete geniuses. And he did. everything possible to restrict sales chips with our to the allies. So, I think, our allies too will not have centers data processing. So we fully shoot in your own way, and, probably, it would be good if Daario truly believed to the idea of ​​centers data processing, complete geniuses, which he defended the past year.

He must to come forward and to say what we do big mistake, over-regulating data centers. Here's how we do it. I didn't hear him. performed in such a way keys. And this causes you the question: whether Are you fighting only for the losing side argument, is there positive version the sides you are on ready to fight politically? Now, Freeber, regarding your the last one remarks on RSI, I I think it's true. important, again, black swan, very It's hard to say whether this is will happen. I, you know, I really think that we we are heading to OSI or we already have OSHI or something like this, but most important thing about OHI what we have choice.

I would much rather It would be better to have several OSHs, than just one, right? Yes? For example, Elon says that the way support AI consistency or OSHI—is to make it maximum aimed at searching truth. I like it. like it or not Yes? This is the way prevent it awakening. That which Anthropic says this is the way support consistency is to have this constitution, full of liberal values, and you teach her on it, and they have some kind of ethic with EA or someone else who writes these things.

So, there is different approaches. I want, so that they all create complex AI to to control and control these things. What I don't want is this just Skynet or, you know, awakened Skynet is worst possible outcome. AND you know, if you're already on to go deep for a second in science fiction, then, for example, was series about Terminator. Yes, he was. decent. Its stopped. He was not bad. Its canceled. It was many years ago. But in any case, in the series they understood that the plot was that for victories over They are Skynet created a good artificial intelligence (AI) ) to combat Skynet.

They realized that people they just can't deal with it independently. So they created a choice. Yes. They taught it should be more friendly, you know, friendly to people. So, in any in any case, if already to come true science fiction, I think the choice is yours. to be our distinctive trait. Competition should be ours a distinctive feature. We we want to have several winners. We are not we want a small number of elites stood on this path . And this is what you you will get if will move in FAA FINRA direction.

Break Friedberg. A new one just arrived nickname club. Here he is. This is how you sounded. Friberg. Dock Brown. Marty. Marty. Your future is not predetermined. You you can create your own own future. You can you use RSI. Doc Brown. I am not trying to be...I'm not I try to be a supporter of RSI. I I try to be pragmatic. In the future you you will return. If RSI is the way it is will go, you definitely won't want to try regulate AI to he stopped because someone else wins, and then you literally hand them the keys to RSI, because they suddenly will be able to achieve this is faster.

And you don't Do you want to ban? data centers, because then they will be outside the office. Good. This week It was a strange story. AND Horowitz supposedly is under investigation. All this according to the report Bloomberg that their investigates Ministry of Justice for over a year now. All this according to Bloomberg, according to with something called chapter 8 Antimonopoly Clayton's law. It law of 1914. To this for over a hundred years, prohibited interconnected management, that, I I think it looks like old-fashioned English name directors, for example, the council directors.

AND conflict seems to be arose because Ben Horowitz is included in datab board of directors bricks, while another partner Andre Horowitz, Martin Casado, sits on the council directors of five trains. Not I know, have you heard about this story. It was one of the main news this week. Uh, It's obvious that Dre and, uh- uh, Ben Horowitz collaborated with, uh, Trump and the big donors. Uh, add it. this to this mixture, like you want. Why is this? is happening now? I can tell you. Andre's enemy and Horowitz, jealous their profits, says: What are we talking about?

They don't understand how venture capital works capital? Like, it's like invest in one company, but not to another Is it possible? I don't understand. What did Bloomberg say? They said that they are under investigation through this conflict, as described, between these two companies, which means that the enemy reported them Ministry of Justice , and then told Bloomberg: "They are under investigation." Well, that's why I I'm bringing up this topic, this seems like this strange, and then to me I think I will do it. assumption that CEO one of these companies angry that other partners in another board of directors reported this Ministry of Justice and Bloomberg at the same time.

Interesting. Well, okay. Well That's it, people. Let's get down to business. I think we are fast. coped with this. Let me just to say that you saw the distribution, for example, the cursor, and then SpaceX and Horowitz in lights. They are devastating. . So, in there were two of them the last few weeks. I mean, God. My, I congratulate them. How many billions they earned on this? Well, the cursor, and what? they did not earn a lot of money? Offer open router only for seven dollars.

Open the router was more like a push for the emu. was something like 6 or 700 millions, but the cursor was decent. It it was like a snack , but SpaceX was the main dish, which cost about 157 billions. Yes. Now you just talk between to others, even if do you mean that venture capital firms do not should invest at competitors. The reality is that startups exist in very dynamic environment. They often change their business models, and two companies that start at completely different areas, can eventually compete one with another.

So, this is various members board, and they, probably can to set, you know, firewall between each other. Yes. Yes. Sorry. This is not companies that do not submit documents. What confidential information has any of these private enterprises? And, by the way, Andrea Horowitz could also just assign there independent director and on this to finish. So, if it was some kind of very important thing, I don't know why the Ministry justice by this is engaged. Freebrook is right. No, because when you file a claim for Ministry of Justice , they are technically, if open the file, you are technically are under investigation.

And then you can you tell about this to someone to the clown reporter, who knows nothing about Shiola, and they just repeat it. This doesn't mean anything. Good. And then the Ministry Justice doesn't tell you, that you are under investigation. And then they don't say either to you when it closed. Just like it's simple one-sided PR a move for the enemy to create some chaos. Ah, he's destroying it. God bless him. May God help them. I wonder what it is anyway such a rule. Hmm, but we are a little had fun with this.

So, er, let's move further. I want you to, Friberg, joined to me and Sats Bestie Chimat at the All-In Summit Summit, which will take place September 13, 14 and 15 in Los -Angeles, that's it. announced. And you know, who will be the main one hero of the party? Those the same guys who Madison Square is selling out Garden. I know that in Friberg is personal relationship. Let's, tell everyone who will be Chain Smokers, this will be amazing. Therefore, they will be with us on closing party.

It It will be great. And then we will also arrange more one " Wow. Complete delight Universal Studios. So last year we had there was a whole party in Universal, which was stunning. Uh, where are Sax and I? were on stage with Diplo. Jay, you were there too. It was wonderful. Yes, I danced on stage. Who is Diplo? DJiplo. Come to that one. concert. We have others special unannounced guests, which we are not going to declare. Therefore, That's all. We finished doing advertisement. Good. We know that Democrats are favorites on victory in the Chamber representatives on by-elections, which are just around the corner.

Polymark has them at 88% favorites, but Poly Market now has Democrats by 52%. This is actually draw. Uh, current balance of power, Republicans 53 seats, Democrats 47, and this means that, I think, Democrats will have to turn over four of them without losing none. In 2026 35 places will be selected. THERE ARE -e, in Florida and Ohio are happening special elections. 23 of these 35 places occupy Republicans. Important races, for what are they worth watch. Alaska, Maine, North Carolina and Ohio. 25 February, the day before the beginning of the war with Iran, Republicans had a 60 to 41 advantage.

She reached 70%. Here is yours. polymarket. Sax, what are you? What do you make of this? Why Republicans are going to experience defeats at intermediate elections? Well, first of all, survey, as is known , are mistaken in the summer before the elections. And there is some data from this reason. So, if you are you going to spend field surveys, let's just let's talk about whether are political survey in general accurate, because it not a survey. It polymarket. Just so it was clear, survey somewhat are different.

Yes. People. Yes. But the survey form a public opinion now. Therefore, political scientist Patrick Graffini analyzed over 3000 surveys during the last four elections cycles every two years , starting in 2018. And he saw huge partiality. The advantage of surveys in all four cycles was in favor Democrats, and the average polling error in all surveys was D plus 3.7 regarding the result elections. And I think that this problem polls last sometimes it got even worse . And you see it there, where, it seems, observed oversampling young progressives in these surveys, because it's not the case only in that survey provided preference for Democrats over the Republicans.

They also provide Democrat preference from DSA over more moderate or centrist Democrats. And you seen this on several elections where a candidate from DSA should have won with by a large margin, and then he actually lost or actually lost, as in Wisconsin. So, in anyway, I I just think that the survey is very unreliable at this time of year year, and many surveys actually are created as internal data campaigns to motivate your donors. So, essentially, people upset, and then throw people out positions. It's simple. trend, and that's what happened again and again.

Yes. No, I just think that survey of some kind in an excessive manner overestimate Democrats and progressives and have partiality. This is what what the data shows. So I wouldn't now gave this too much attention. I think the Democrats, I think they have the right. to say that they are favorites on victory in the Chamber representatives. I I think the Senate is now uncertain, and I would said that Republicans will eventually save him, if I had to now provide. And I I just think that when you will return from the Day work, right then will really begin campaign before by-elections.

And I think that Republicans need to focus on the right messages in this campaign. And I I think there are many. good ones. They just should remind to him, who are all good things. On which, in your opinion, the three main ones must concentrate ? Well, look, everyone. excessively focused on war, and it was in the center of attention of the news, and, I hope the kinetic phase of this ended and not will start again. AND assuming that that's right, I think so Republicans can remind the electorate about everything that happened the last few years.

I think there is a good one. a story that can be to tell. So, by- First, remind me. to the people that the border was closed during the last one and a half year. We are not just slowed down Biden's invasion. We completely stopped. Contacts on border fell sharply to the lowest level in 50 years. Illegals not even anymore are trying to cross the border. It was so successful. On- second, closing border and deportation of hundreds thousands of the worst criminals in reality made the country much safer .

Number of murders decreased by approximately by 20% compared to last year, and the murder rate is the lowest in 70 years. So, that's number two. On- third, mortality from overdose drugs again decreased by 20% thanks to the reinforcement border control. Fourthly , the Americans saw 17% growth tax reimbursements thanks to because the president is not introduced a tax on tips, tax on overtime work and tax on social software. Therefore, this is real increase in net salary in as a result of the decrease taxes. Also tax regulations in a big beautiful bill stimulate purchase new equipment.

We talked about accelerated depreciation. AND, maybe that's why optimism of the small business has just reached of the highest level almost a year later, and that's it more small businesses say, who are planning to hire employees in the coming months. So , I think we have really strong economy. Cost prescription drugs and car insurance fell last year after a significant growth during the times administration Biden. I think that cost of prescriptions underwent medication the fastest annual decline since 1963 year. So there is something to it.

speak. All provided a great the value of the price jump on eggs at the beginning administration, but few people worried to note that in July observed 39% decrease compared to the past year. So, look, I think that Republicans have many good things, about which you can speak. They just need to remember everything that happened the last two years. Okay, Fber, you have it. thoughts on what will happen on these by-elections, and what are people on are based on their decision. Things are expensive. People don't can allow for themselves housing, security health or education and are leaning towards this DSA concepts.

Was frightening statistics regarding young people conservatives who support socialism. Seriously? Wait, conservatives, such as MAGA people, you say, support socialist trends? How are they then? conservative? Well, I I think they are stubborn. change or somehow change in a way calm about issues, that worry them. 53% of those who identifies himself as conservatives, up to 40 years support grocery stores, working under government leadership, and not democrats, conservatives. Take, for example, the first line of this articles in the Wall Street Journal. Fox News poll in July 2026 showed that 61% Republicans have a positive attitude to capitalism, compared to 72% in 2019 year, while the share those who think so strongly approvingly, decreased from 54 to 41%.

The proportion of those who say that system rigged for rich, increased from 30% in in 2018 up to 42% among Republicans. So I I don't think that's it. a fundamental limit. I I think that's the problem. inaccessibility, about which I constantly I claim, and I I affirm this with on the first day of the show, there is function of state expenses. The more the government spends on education, security health and housing, so these are getting more expensive things, not the other way around. They must have possibility to provide access to these things.

But when the government intervenes in these markets, these things become more expensive. And no one wants to admit this in neither party. And I enough time already visited Washington to know that no one doesn't really recognize and doesn't care about this the moment that everyone think that theirs the work consists of to get more money from the government for their voters. Each a person in each parties. Very little principled conservatives. Many of them are conservatives for position, party affiliation, affiliation, but they are not conservatives for principle in traditional understanding, small government , free markets, etc.

AND I think that this is motivating factor. I would say that if to look at macrocycle, macrocycle is that we printed like this we have a lot of money so much debt, 30- annual percentage rate this week's rate is 5.3%, it continues every day install new 20- annual records. I I think we we will continue to encounter problem inaccessibility, and will eventually happen some big movement towards more socialist the policy that we We'll see by 2028. That's why I continue. to believe that someone like AOC will become president.

I think what will happen this year wave, and then in 2028 will be bigger this year push because I don't I see that arithmetic or fiscal The situation has changed. so that in fact change the landscape accessibility. I just I don't know what to do. to do from the point of view of politicians, and I haven't heard, so that one of the parties talked about real solution for solving this accessibility issues until 2028. And then another big problem, which I was talking about, this assets. People middle class, people who worked in private companies, people who are not dependent from social provision or state pension funds, all of them had 401k retirement accounts and owned housing.

So their fortunes grew together with rising inflation assets. So, 50% the richest Americans have 178 trillions net assets, and the richest 50% Americans have 6 trillions net assets. This is not billionaires. This is the whole middle class. Middle class owns everything wealth in United States. AND it's because they owned assets. They owned houses and stocks. The cost of houses and shares rose sharply. So if you don't have house and shares, you were left behind. AND so I don't see how from the state point of view we are not politicians let's end up on this very unpleasant, winding, dark the path of socialism, because alternative policies that can be would be to introduce to solve these no problem now are nominated by none party.

Therefore, I think, what will be big socialist support. I already talked about it, obviously, during many years, and I I think we will see, how will this manifest itself in these intermediate elections, and this will create momentum for on the way to the 28th district. And listen, I have mind you, this data surveys very alarming. This is not only Democrats. This is not just a party system. This is a horseshoe. That's 63% of Americans who live off their salary to the salary, but not can afford this to allow It Americans who will vote for this system.

And it doesn't matter, What party is this? They just vote for her. Approximately 20 cents from dollars go to our interest from federal side. THERE ARE- uh, inflation, we thought that we will achieve this. For the whole year it was 3 or 4%. Um, it didn't work. reduce the indicator to two, and the concept of GDP, what are we talking about here they said, again begins to grow, and if we can do it achieve, we will be able to to get out of this way growth, this also did not happen.

It is obvious that war became big disappointment for many people in both parties that were promised that there will be no war, uh, and here we are 6 months later in a row. What, in your opinion? opinion, it will happen during the next, e -e, midterm elections, and then in 2028 and for the release of Birch- Point? Is there any a way out this? This is debt. spiral, is that it, with what can we do to fix? On yours opinion, I think we can to correct.

I just I think it is necessary. will. And I think that we it is necessary that corporate America took up the matter and fulfilled her part of the work. And to us governments need to mostly not intervened and did not have no fancy schemes regarding the new inventive legislation, which, in their opinion, maybe solve the problem. I I don't trust anyone. survey. Simply To be clear, I I don't believe this data. surveys. And I think, that we have reached a very dangerous point, to which we have already reached with Media where the media lies.

We all know now, that the media lies, and the media cannot be trusted. I I would like to now to take a risk. However, I don't I think this is something very dangerous. Questionnaires are lying, and I don't think so that they can be trusted. And I think the media act in coordination with sociologists, to find data that will allow them write a title, which, I hope, will allow them get clicks. So I I would just start with this. Questionnaires the media is scolding. So what?

actually takes place on places? I don't know. But I don't trust that. show this data. What I know what helps and harms the current to the president? I think what if we, as in corporate America, we will get possibility to present better the face of capitalism, we will allow the system in her current to continue work. And I think, that all these things, if we will not wake up, they tell us that corporate America no longer deserves on trust. Corporate America does not desire anything. Corporate America doesn't do what causes a certain trust and respect.

AND particularly in Silicon Valley valley we are completely lost the script because reached the very end. And everyone in this involved in one's own right to the extent that it or a pearl, from one hand, or arm gets stuck in a jar with cookies, on the other. You know, I posted story with a jar of cookies that Anthropic made in the field health care. A jar of cookies. Why are we in this? I have I mean, this is so easy to avoid. It headwind for the current president, even if he doesn't has none before relationship.

What is a fair wind? I I think that the way out is war in Iran may to be huge a fair wind. I I think that structural solution energy problems in the interests American citizens may be a fair wind. I I think that ensuring that so that everyone long-term and medium-term GDP investment could actually bring fruits, maybe a fair wind. IN all these things still have many levers that need to apply until November. And therefore they can create enough significant incidentals winds, but headwinds The winds, unfortunately, are quite strong.

And the White House need a real one coordination, so that to collect a bunch of these general directors corporations, collect a bunch of generals directors Silicon Valley and tell them: "Guys, enough of this nonsense. Play polite in the sandbox, because you are about to lose standard operating procedure which allowed you exist. Yes. And they need to be clear to convey this message. Yes. And I can say you the exact reason, why democratic socialism will win, will continue win. Um, and these far-left candidates from the Democratic parties will dominate.

And this, you know, wine Republicans. They very tight controlled all three branches of government. And if you look at what participants react podcasts, uh, and professional stack preferential podcasts, and I listen You know, not everyone has them. episode, but I, I, I, I I'm diving into Tucker, which is currently creating own political party, Joe Rogan, Tim Dylan, you know, all these people who everyone supported Trump. They quote investigation Epstein, what's good, what whatever. I don't think so. what is special significantly, but for them That's right.

War in Iran, obviously, is the main because they consider it a problem, because they were promised that This will not happen. AND then it's up to you thoughts, Friberg. This problem with inflation not resolved. And she just continues . And we will sit here, I think, because 6 months or 12 months, discussing all possible investigation Trump administration . And we need to begin to consider that , what, what will happen, to your thoughts, Shamat, and I I thought you were summarized some broken parts capitalism, I think, okay, but we need details consider this, and I last week only raised the issue extreme behavior in capitalism, where we over-optimize .

I felt that service providers delivery to Sax, and I had quite lively discussion on this about the past weeks were just too smart, squeezing too much a lot of the latter miles literally from the side Amazon. The way we we treat people, which are located at the bottom of the list employment, there is the reason for this, and we we can't decide question of value housing. and not to do it This is the accent. Impossibility to resolve the issue cost of care children. Impossibility raise the minimum salary.

I I think we should consider all these things and say that or Trump should consider all these things, and he must build a second one half of yours the term is around these things. More no cryptocurrencies, no wars, no attacks on our allies. All these performative things and just mistakes should disappear, and we must concentrate on what he is on convinced people. People feel that they convinced that this will be populist the presidency which will take care of working men and women in this country, but everything was the opposite.

Sax, I'm sure that Do you have any thoughts? We will give you final word on inflation because you talked about it, and Friberg spoke about it. I mean that you say that the cost life will lead everyone to socialism. I I agree that there is some risk of this, but let's Let's break this down a bit. more details. If to see where the largest took place inflation, then it always in these sectors ...Economy in which government the most involved. So health care and colleges, as well as housing where the government has huge subsidies and regulation, it seems only makes the problem worse.

It is quite logical that if the government is for something pays, then suppliers have huge incentive to collect as much as possible more. And in fact many of them too are engaged in fraud. So we we know that these areas full of abuse. But these are the areas that are the most inefficient, and where prices are skyrocketing are growing. I have on attention, hospital services, it's simple exponentially. AND then, if look at the spheres, where we don't just have there was inflation, and prices actually decreased.

This applies to everyone. technological areas. This is mainly televisions or computer software software, mobile phones, similar things, prices decreased significantly. To things, this diagram shows the entire period from 2000 to 2026. So this is 26- annual period. And for this time general inflation rose by 96%. So the price level is almost doubled. for this 26- annual period. On happiness, on average, on average this is right. But, as I I already said some things, such as security health, steel almost 300% more expensive , while computer software software decreased by 67% real dollars.

Tim sometimes average hourly rate Labor increased by 135%. So at least average hourly wages have increased approximately 40% more than the inflation rate, but obviously still this is felt inflation. I think mine the idea is that How to solve it? We we know that prices are decreasing due to active competition, and not through reinforcement government intervention in these markets. And, unfortunately, it happens that socialism causes problem. And then it happens that groups like DSA, claim that the way problem solving— it's more socialism. So this is a disease that claims to be medicine, and then people they accept it, and they everyone accepts it more, and illness is getting worse and getting worse.

And this is it the spiral in which I I think we are taking a risk. to get in. We are in a spiral now. socialist debt. All these expenses now. We are not in...we are not in... death spiral. Not worry. Ahem, we are not in a spiral death. It's great that you They said this because it seems what do you really have hot idea. We haven't yet. there. America feels good. We are promoting . The economy is great. THERE ARE things you need to fix. I would say, start with corporate America.

Let's let's get together and let's start doing what is honorable, interesting and ambitious. Yes, I think you are. completely not understood the essence. I I think it's real. The point here is that if he hadn't started war, we wouldn't have the same problems with inflation. We don't have would be the same problems with expenses, and that would be completely another picture, and this was his job, when he promised American of the people who will focus on accessibility. And he controlled both House of Representatives and the position of president.

And he is ready. to do any decree regarding cryptocurrencies, wars, anything. But he did not publish any decree regarding accessibility. Maybe you don't noticed the moment when Chamas said that this there will be a fair wind, to get out of this war. I think that To things, kinetic part already ended. So, let's hope. I I mean, this it just shouldn't have been start. Um, me too I think but in any case, look, past president year adopted its economic program regarding the big beautiful the bill, and I I think it works.

We see a huge investment boom. We we see how trust in small business grows to the highest level for a long time. They say that will create more jobs during the next year. I think that economy feels good good. I mean, that we have these problems . I'm not saying we have them. no. And the expenses are an old problem. Not I can't with you. agree. However, I I don't think so. Democrats have solution to the problem. AND democrat socialists definitely not problem solving, because they want higher costs.

Death consists in that we are now we stimulate young people people in Republican parties to accept socialism. Moderate accept socialism. They are all looking at this is what they say: "Well, this "sounds better." If Trump will not populist, perhaps these people are populists . "And that's exactly what I have on my mind." under the pile deaths, because they simply will continue spend money. They are going to increase speed of spending. Maybe you have That's right, that's scary. By the way, I I don't think we are now.

in this. We take risks. to stop only on one, only on the last point. They made an assessment the cost of political platforms, proposed DSA. Remember, we already talked about it earlier. And, according to estimates, this will provide new federal spending from 71 to 212 trillion over, by the way, general number prisoners. None all American populations 182 trillions. 182 trillion. Yes. Good. But I just I say that's exactly how much the program will cost TSA if you guys Do you think the root our problems consists of overspending and, by in essence, in excess governance.

Unfortunately, that's not what DSA will do... No, but here's what will happen. And I just I will allow myself to walk along sequence of events. Sequence one, they They are introducing politics, starting with 2028, under leadership comrade AOC, and government office, uh Roano, Minister housing and communal services farms and urban development, and the rest of the Pilot Bureau are organized for launching these new programs. These new programs are launched, and expenses rapidly are growing. Market bonds sold out for 30 years. And now 30-year yield bonds increases by 7%, 8%, maybe more.

Things are getting better. more inaccessible. So then the question is in that, well, we don't we can overcome this with fiscal point vision. We can't just force the Fed start buying bonds. There not enough purchasing ability to to lower rates in order to reduce inflation. Therefore we need more income. Ago now we need to introduce national tax for confiscation assets. And there is no enough money in billionaire class or class of 50 millionaires. Total net the capital of all people, whose fortunes exceed 50 millions of dollars, is only 23 trillions.

You can take 10% of them net capital, and this will only cover 3 fiscal months expenses, today's fiscal expenses. So in the end, the One way to achieve this —start conducting mass tax arrests. And here are the arrests, arrests. So, this is just appeared this week , KO by KO from the KO Institute. AND again, this shows that the lower DSA spending limit is 71 trillion. Upper limit—211 trillions. Now, if you compare this to such things as corporate profits or, uh, private wealth sector, corporate profits are only 35 trillion.

Forbes 400, all the wealth The Forbes 400 is only 6.6 trillion. Therefore, this is less than 10% of lower limit of costs DSA. I keep saying you that this is not about mathematical realism . It's about anger towards the class. people who become more and more disgusting and distrustful. But people can't either to afford buy a house To people We have been neighbors for 35 years. room. And people go to grocery store and shocked the fact that they have to spend a whole salary or a third of the salary, so that buy products.

It's not that people don't want to buy... yacht. The thing is that yacht class just... No, I mean, well, No. I mean, look, in each socialist, hmm, revolution, each revolution that we seen historically around this, always there is an elite split society, and they usually are classified according to different. This is wealth. , these are Jews, this is some kind of race, these are immigrants, there are certain organizational purpose, some appointment identity, which says that it is reason unavailability. Ultimately, dysfunction relations between the government and all the others organizations that make it up society.

These organizations can to be states, it can be private companies, it can to be NOS. Exactly relationship dysfunction government with others organizations fundamentally leads to that things are becoming more expensive and dysfunctional, and the actual market for gaining access to things for a separate a person experiences failures. And in this the moment you have to identify, and you do not identify government, because the government is the totality of all people. So, and remember that government is system, means of action. So that people organized, they should organize ...the government needs to go and then eat will have to have alternative a system that can to destroy.

You must specify finger, and you must to point a finger at someone. And if you are you going to to accuse someone of In 2026, the best class of people That's what billionaires are for. And, honestly, even better than them. But in the end, all of this technology, right? Data center- This is a temple where people pray. billionaires and trillionaires. This place where they go to to make their sacrifices. This is their gathering place. So I think the center data processing is something like, you know, temple on the mountain.

And therefore I think the center data processing connected with all this. There is no rational discourse about centers data processing. It type: no, no, look, if look at this through the prism of Rene Gerard, everything is displayed. It like billionaires accuse, and trillionaires, I have mean, even worse, and because it will allow catharsis, which will allow society reboot and realize: wait a minute, oh Oh my God Yes, there is a whole generation, although there is one more generation ahead, and only then this a long process, this long.

Here's an idea: next president or trump in to his second half just will publish it, will publish it construction decree a million homes, two million houses, if possible will have decrees. If can start wars without approval Congress, That's not how it works, is it? wrong? What is it? You know that this is not That's how it works, right? There are local... He, of course, abolished all laws, that prohibit... But he did everything it is possible that the centers data processing got green light. We know this.

And he worked with by state governments. I I'm talking about how he said: you know what, we made a great work for centers data processing. We made a great work for crypto industry. AND now we are going to make two wonderful things. Wait. Let me to finish. We we are going to do two wonderful things. We we are going to spend exhibitions. We we are going create vocational schools and build affordable dwelling. I just I'm going to...Do it. within the next 24 months. We will find way to mobilize private sector, mobilize public sector.

We let's build more houses and we will provide you excellent education, which turns into work. This would be great for Trump. Federal government controls the largest amount of land in United States. So he could do something to do on federal states without the need for a pile other state and local nonsense . So, there is May Freddy M. Now she can be is not located in places that are actually make sense for commerce and for human habitation, or who already live there, because I don't think so it actually borders with most cities.

So, unfortunately, it's not his control, even if he just wanted to say: let's Let's do it. I think what if the president can spend a couple hundreds of billions on this war, he can spend 100s billions on affordable housing. AND, maybe he can't place them on Manhattan, but he can place them in other places. And he must fight for that the rules...yes just as he fought for regulation processing centers data and energy. He expressed his opinion in this very podcasts at the summit artificial intelligence. Hey, we're going.

clear all paths to nuclear energy, to get rid of gas. We we strive for beautiful clean state federalism. I I give him credit. for this. Now do it. it. Now do this. for houses. Let's I will see how you are. do this for houses. Trump, use that one the energy itself. They just passed draft law on dwelling. They just approved draft law on dwelling. When did we last hear? Trump? Unfortunately, his written by Elizabeth together with Warren, so I don't know how much he good, but Yes.

In any case , I'm just suggesting him a decision to save him inheritance. Okay, everyone. As always, you Take care of everyone. Take care of audience. Noyce, Thank you. Noyce too cares about you. Okay, rap. I need to go. Come on. Good, all. This was another one. amazing episode. I love you guys. It's nice to be in podcasts. Good thing that The quartet is back. Come on. See you later. We...Let yours The winners are passing by. Rain and I opened source code for fans, and they just went crazy from this.

I love you. Queen your best My friend left. This is my dog ​​drawing attention to your driveway. Oh, God, Myasher will meet me, oh We should all just to take a room and to arrange one a huge orgy, because they are all just not on that are not capable. It's like sexual tension, which they just need somehow release Moisten your phlegm. To us need to buy merch. I'm going for everything. I am going to

Go deeper

• All All-In Podcast episode guides