Nvidia's Historic Quarter, SaaS Comeback, Bessent vs Druck, America's Debt Crisis, Cancer Vaccine

Key Takeaway

Treat AI as a leverage tool, not a substitute for judgment. Today, choose one recurring task—writing, research, CRM updates, or analysis—and use an AI assistant to produce a first draft, check facts, or organize context. Then personally review and own the final decision. The episode’s broader lesson is that durable advantages come from combining reliable systems of record with adaptable AI agents, while resisting simplistic narratives that predict either total replacement or effortless transformation.

Episode Overview

The panel examines Nvidia’s record earnings, Salesforce’s rebound, and why AI agents may strengthen established systems of record rather than destroy all SaaS businesses. They also debate America’s debt and refinancing risks, the appropriate use of AI in authored writing, parental controls for social media, and personalized neoantigen cancer immunotherapy.

Full Transcript

Transcript of Nvidia's Historic Quarter, SaaS Comeback, Bessent vs Druck, America's Debt Crisis, Cancer Vaccine from All-In Podcast. Auto-generated from episode audio; may contain minor errors.

So, hello everyone. Welcome to the podcast number one in the world. Your favorite podcast. Favorite your podcast favorites podcasters. Yes, it is. All-In podcast. "Main" four", " Fantastic Four », original The quartet is here. David Friberg, just after interview about theory "deep state" conspiracy states" with Eric Weinstein. Surprisingly. You made a small trip to Washington. Excellent episode, Friberg. It was fun. Eric— a unique person. I I also saw Katsios. Two sides of the science medal in America—it was the basis of those two interview. They were interesting.

Who is Eric? Weinstein? This is a scientist, who works with Peter Thilem, graduate Harvard. Did something truly capital. Yes. And he had podcast during short time. He- heterodox thinker in science. What is its purpose? heterodox view of science? THERE ARE many things that everyone perceived as fundamental foundations of modern physics, for example, string theory, and he thought so nonsense and openly about it declared. So he became a kind of exile in scientific communities. This is nonsense, truth? He justified his position and proposed alternative theories who create their own kind of "theory of everything" in physics.

And after this scientific his community rejected, people they call him a charlatan and a fool. He talked about UFOs, about Jeffrey Epstein, and people say, "Well, this guy— marginal", or use others terms to convict its. Over time, you know, the essence is similar to that Katsios said: many What science imitates this herd mentality where everyone has to line up, agree with the same general theory, otherwise you will be kicked out. You don't get you don't get grants you don't get a position work, you don't invite you to read lectures.

You have to go to keep up with the mainstream in science, otherwise you exile. And before it was all about I know. But very rationally speaking, that while string theory it is not proven, she remains unproven. So if you want to use adjective « nonsense, this normally. This is probably will offend some people, but in essence he more right than is mistaken. She unproven, right? It's like Jay Bhattacharya, who says, "Hey, you know, COVID could to originate from a source in laboratories. And when people said that, they replied that they create theories conspiracies.

You are spreading disinformation, although no misinformation no. If you don't are you talking about absolute truth, then this is a theory. All started with theories about where COVID came from. You are talking about level of corruption in very narrow field science, namely about financing apparatus and approval grants. And getting jobs teachers and professorial titles. So if you don't follow the rules, methods, theories, if you are not in the mainstream, You are being kicked out. To By the way, I agree with this. I , I don't know, Eric.

This is what led to stagnation in science in America, because if you don't part of the mainstream, You are being expelled. AND because now everyone have to think anyway, you don't have heterodox thinking. You don't have thinking outside the box, and this means that we we are not expanding anymore horizons and not we open a new one. AND this is quite spicy question now. Here why did I leave and talked to Kratos, said otherwise. Probably in 2026 a lot more gradualness than It was in 1926.

Absolutely. That's right. Remember the post general theory relativity. It looked absolutely amazingly. But people are all they studied it anyway, evaluated. They are not discredited her with with the words: "Oh, you idiot." And if you think that this general theory relativity is real , you idiot. That is happening today in science: if you say things that too far from standard approach , you are being punished. You they are expelled. You exclude. Have you seen Olympics? Saw Robot Olympiad, Saxon? Look at this. I saw. Yes. It pretty crazy.

IN This Olympics was theirs. in China, and... Oh my god, they copied how Friberg runs. Look at this. Yes, this is definitely a sprint Friberg. Friberg Sprint. Sprinting Friberg. Look on this. He retreats, and then boom, explodes. This is my sprint. This is like me skiing right now. Yes. Look. But here Jason Kuhn sets records in 42 years old. Here is Jason Kuhn is right here. This is Jason Kuhn. Yes. With a bare torso under poker game time. Very good. Put on a shirt. back. Kun, you know, I, I I was thinking about it, Sax.

Do you know what I realized? I think you are 100% right. regarding PR. And I I realized that the Chinese, PDA, simply brilliant in PR. Look, What did they do? They gathered a bunch Chinese people fill stadium and cheering for AI and robotics. They are rooting for it. They support it. And then Dario says that all work is over. We will be the last. company in the world, and This is the end of time. I I cover my rear. We are not talking about Dario this week.

Your grand theory is that in a country with a billion population not there were 4,000 applicants see how it works running in circles. You Is there really only one theory? It is true that China AI-minded lot more optimistic than we. Optimism about AI in China exceeds 80%. It means that they answer to question: whether Do you think that AI will bring more benefit than harm? Over 80% of Chinese they say "yes". In the USA this indicator is about 30%. So I've been saying for a long time, that the US is ahead China in everything except one—optimism.

AND this is the biggest risk for us in this AI race against China—to do what will harm us ourselves, because we are too pessimistically are set on of what it can do AI. I think that's all. setting. They are forcing you feel. Do you see this? Show me how the robot explodes on the 16th Just a second, Nick. That's all. directed. It setting. It setting. They make you feel sorry robots. You...They are forcing you to cheer for them. It will become something notable. No, it's a red robot.

The funniest thing was, when the red robot lost his legs, it was a few tweets that I I find it very funny. One guy wrote: "Oh my god, he tore it off." your HDMI, your processor". And then...and the second one, which is not funny because I know Christiana McCaffrey, He is a great guy. He's like Christian. McCaffrey on third weeks when you see how he just has The leg comes off. Very, very cruel. Congratulations to the CCP. By the way, Optimus, you guys, accidentally missed latest videos from Optimus from Elon?

He showed someone you? Didn't see. He doesn't publish them. for the public. I saw the last Optimus. He would have already won. half of these competitions. Did you see this? He makes crazy progress and does not show it's public, but I saw. Without sparks, that fly out of it housing. No sparks, but I saw the video, about which I told him: "Is this CGI or is it real?" AND he said, "No, it's real. I need it. just sent this morning". And this sounds like: "Wow, Optimus will become the most product sold ", I think." Yes.

You know, the most difficult thing is how I I understand, with these by works—this is what you can you teach them to take a specific action, like running in that race or jump, or whatever. But the hardest thing is when they are faced with a condition, which have not been to programmed, and especially when they has to to interact with something in the physical world— Here's where it gets tricky, Yes? Can they to fold laundry or to lift something, not having broken and not dropping it, or what?

Anything? And especially something they weren't taught to do, you know, This is where it gets tricky. here. Yes. Precision Optimus and his ability to do new things. You seem to be hit the spot. I am not I want nothing more. say. You were playing with Grockbot, right, Sax? Yes. Oh, yes. You're in trouble with Grockbot. . You squeeze. maximum with Grockbot. I just reached my limit. using Grockbot and I need to update. So 200 bucks. I was...Yes, it was. actually very interesting. Yes. What? He just does what you tell him you say, right?

Like, he just does work. Yes. What's great about Grockbot, so that's what he is always on. He is in the cloud. You you set up yours agents, and they can continue to work, even when your computer is turned off or it fell asleep. Others, you know, agent frameworks that tried to do it, like OpenClaw or Hermes, they were just on yours desktop. So if the computer fell asleep, they stopped. You had to use a bunch Mac Mini, and they are also were very difficult in used, very energy-intensive, super complicated.

This is certainly the simplest option . This still requires a lot of effort to configure everything right, or whatever yet, but it's fun, it's the easiest thing ever still, and you really You see what this is all about. . Yes, that's enough. surprisingly. I said him, what is it worth make multiplayer regime. So that it could be possible to accommodate two agents in one room, but then to be able to to accommodate two people from your team in one room, and you know, having people in the loop, this will be...

such a great idea. It will make this thing to go crazy viral. Yes. if you can just add, because then, you know what What's cool about agents is what you really are You have a swarm of agents. IN there are actually a lot of you various agents that perform various tasks, and they are getting better at various things. You know, some people ask, why not just have one agent? It turns out that if you several agents, then they acquire more context and experience, therefore they become more specialized.

But yes, you could then use each of these agents as a channel, invite people there, and they would take in this participation, and then added more one person. Yes, this is... Yes, then it can to become very, very viral. Yes, that's great. product. OK, Let's start working. IN We have a lot of things to do. Nvidia just published record report on profits. I have on mind you, they just blew up the market, and we are also a little let's talk about Salesforce, which returns, and about the end SaaS apocalypse; Nvidia shares rose 9% against the backdrop of huge indicators and more bigger predictions— $26.2 billion income per quarter.

Chamate, this is growth. by 106% compared to last year, more than twice. Waiting for Wall- Street was 92 billions. They gave 70% forecast growth of the next year. This is much higher. what I was expecting Wall Street—45%. So realize this. Figure, from which all going crazy: 60 billions of net profit. This is the most profitable quarter core business any public companies in history. IN Google and Amazon were higher one-time quarters, Of course. Here is yours. schedule. Clean Nvidia profit and revenue quarterly—they just tearing apart templates.

Promotions are growing. They are again achieve their historical maximum, already close. Before the report Nvidia shares only rose by 12% since the beginning of the year, on par with the S&P 500, but they were lagging behind general index chips. I think, obstacle, Chamate, is that people believe in approximation competition. We are about Let's talk about it. Nvidia— the most expensive company in the world. 5.5 trillion. Poly Market gives 79% probabilities. Here is yours. Poly Market, our wonderful partner, Poly Market. 79% probability that Nvidia will finish the year on first place, 6.3 million volume bidding.

In other news, Salesforce grew more than than 20% on Thursday after also reported about wonderful neighborhood. Revenue increased to 11.3 billion, which is 11% more compared to last year. It corresponded expectations, but they exceeded the forecast adjusted EPS by 80% : $5.90 vs. $3.27 dollar. They did big jump, because they are big investors in Anthropic. They raised their annual forecast up to 46 billions. You were right, come on Let's have a party. his circle of honor. Here he, circle of honor. Circle Chamata May 15. What is your opinion about this?

You were quite critical regarding the SaaS sector in previous episodes. I think it's tall. market segment where Mark works where big ones work monoliths, entirely safe. People are beginning to understand: Hey, wait. a minute. This is much more difficult than us thought. It is not that " "boop-boop-boop," he entered request—and “beep-bop”, everything works. This is not true. works. I think we a little oversold. Now I think that this consolidation and revaluation can to take place in opposite direction. So, what is opposite agreement? The opposite agreement is who has constructive clean income withholding, who has a negative the outflow that was really predictable, and this another way of saying, Who has the best?

relationship. I think that these guys are ready for success. Okay, that's it. Chamat, what do you think? about Nvidia and Salesforce, which surpassed everyone expectations in their quarterly reports about profits? There is a lot here. to discuss. So, maybe we'll start with Salesforce and then me I'll switch to Nvidia, that's why what I think Nvidia is good transition to conversations about national state finances. Ahem. I am not I know. For me in May was quite it is obvious that exactly Salesforce was significantly oversold. I think what I tried to say then, and, maybe now I can to formulate more precisely , —this is what we approaching end of the second phase of AI.

So, if the first the AI ​​phase was models, and you can think about the model as about brain, then the second phase AI is tools, right? and agents, which is similar equivalent to providing brain, a pair of eyes, hands, notebook for memory and keyboards for printing. So, you take this primitive object that was really good at issues and answers, and do its more useful, almost as autonomous creature. But we will soon let's start this the next phase. At least that's what I see in enterprises, with which I work with, namely the fact that this is still not enough.

And now you need to do this autonomous agent much more knowledgeable, give it context and allow to do the work much smarter. This means that you take the brain, give it his arms, legs, limbs, eyes, etc. But then you have to teach him to be lawyer, service representative support, trade agent or someone anything else. And so that do this one next step, you it takes a lot contextual information. Therefore, people who today control large accounting systems, about what I said before and on Why am I emphasizing this now?

occupy an incredible important place in ecosystem, if everything are doing it right. AND I think you see that Mark finds way effectively work with models, support them, but at the same time build own and start move to the next phases. And I think that's it. therefore its index clean keeping income is so high. That is why his income high. That's why he raises forecasts. And that's why The stocks have soared. It seems they increased by 43% since the moment, as I said, that they have reached rock bottom.

So it took... Yes, today they grew by 20% only on this. Friberg, you were on CNBC or just on quarterly report? I was in the office. Benioff on preparing for the report about yesterday's earnings. Dario was there. Dario and Mark spoke on CNBC. They arranged studio for a report on profits at 60 floor, where we are everyone had dinner. And then I participated in report call from Mark, because, you know, about a year ago, I think I said to you, Chamat, we somehow created over the weekend CRM tool within the company, using cloud code and Cursor, all This was quickly put together and launched, and started use.

AND then it turned out: " Need to add this function, change that one", and also accounting, security, access, and where is the storage data, which guarantees protection? There were many disputes regarding all functions that we finally had to to develop in order to make a system really useful and scalable. Tim sometimes we have companies were conflicting interests about which software we provide we can develop— what is unique for us, for example, software for plant breeding, which is our main competitive advantage. So, how us to create a better technological platform for selection?

How do we create software that will provide tools to the people in the laboratory, greenhouses and breeders? So we everyone started building these internal tools and later understood that our time, effort and energy better to spend on development of working processes that are unique and create value just for ours organizations, instead of in order to recycle software provision that others have already done so. Where your best ROI, best return on investment time spent and best return on investment invested dollars? AND it's not about perestroika communication a tool like Slack, or a messenger like Gmail, or CRM systems, such as Salesforce.

Then we talked to Mark on the phone, and he says: "You know, try Salesforce" . I will help you with everything. to adjust. Come on, Do it. And I'm like: no , no, no. And he: yes, yes , yes. He is like this persistent. You too You know Mark. He incredibly persistent. He added his own best seller in correspondence with me, and He wrote to me every day. Are we working already? We already Are we working? Until I said to the seller: Okay, go to mine.

office. He is going to my place. office. He is everything configures, and it's done. And then we realized: oh oh my god, this is real much better way for us. Mark asked me tell this story on the report about income and share looking at these changes. I believe, and I have already said about this earlier, that horizontal platforms—these tools like CRM that work for many industries, whether Gmail and Slack —they are omnipresent. Excel no one will be remodel, the real value is in software provision, unique for your niche.

Right here AI plays a role in creation individual work processes and programs. This is how I see it. this is real "SaaS- Apocalypse". It rather vertical SaaS apocalypse, that is software provision for one specific one niches. And right there, in my opinion opinion, everything is collapsing. I completely agree with you. A platform like Salesforce. I completely agree with you. I completely agree. with you. SaaS. Let's talk about Marc Benioff. He unbelievable the seller who experienced "SaaS- Apocalypse", found a way out situations when it the company already written off from accounts, and on a large scale integrated AI into its platform.

You know its a long time ago. Perhaps, you will tell about him as about leader, and what we we can learn from his example, when company in the eyes of the market almost died, but then it became again SaaS leader. Well, this narrative is about " The "SaaS apocalypse" was clearly exaggerated. This conversation is about what SaaS is dead, today simply shattered to pieces, because Salesforce shares increased by more than 20%. You you can view my past performances, I I already said that. many months. I am not I believe that the main accounting systems, such as CRM, will be thrown out and replaced by something, created "on emotions." I'm talking about It's been a long time.

Corporations need confidence. They have compliance requirements . They need professionally managed software provision that worked and was being debugged for years. I mean, just think about all bug reports , which were filed regarding Salesforce for 20 years, right? And this the most difficult— do everything right. When I I work with all of these. agents and AI, you are you dealing with probabilistic things , where there is a multitude of edge cases. When you work with main system accounting, you just want to be sure, that she works.

And her the cost is quite low so that you don't wanted to expose danger one of their main systems, removing it and replacing it with something, created by hand "on the knee". So I I think that all this story about The SaaS apocalypse was exaggerated. Meanwhile, Jason, regarding your argument, I think that Benioff does everything. right. Can to say that this is an approach "if you can't to win— join", because he entered into this agreement with Anthropic . What is this? agreement? First, he is going to offer models Anthropic inside Salesforce .

More importantly, I I think it's true. key point. He integrates Salesforce into Claude. He is ready. let Claude be interface to have peculiar primary relations with user. And now Claude will be able to get access to all data in Salesforce, all workers processes and all actions, which can be performed user. Others in the words of Benioff willing to take a risk mediation, so that to ride this AI wave, because he knows—that's it what he ultimately wants client. You know, you don't Do you want to go to Salesforce or any SaaS- application to write their agents.

You want to create agents in that AI- program, which use, and do you want these agents could work on all your SaaS- platforms. So he ready again to give a significant part of the relationship with users to remain relevant. Why is this? ultimately good for his business, not badly? Because, as I said, I still I think that you systems needed accounting. You need, so that AI agents addressed these systems and received data from canonical sources of truth. You you just don't want to around this was any probability variability, and you you want AI agents could record data back into these systems and workers processes.

So, here does not happen complete replacement, and superstructure of new levels, right? IN you have a base level data, then programs and work processes, then agents, and top—interface user with artificial intelligence. AND Benioff says AI now it is the interface user. He actually admits that is betting on it. But here's the thing. I talked to him this morning, and the thought that he expressed and which, in my opinion, opinion, underestimated, is that in Salesforce hidden many values. That is, we know that this is incredibly wide and a rich platform with many functions.

Regular user will never find them everyone, but the agent— will find. So think about it. about all cases when agent now will approach user and will say: "Hey, I noticed that Salesforce has actions that coincide with because you ask me make. Do you want me to? did you do that? AND user will answer: "Yes, you know, Salesforce changes as accounting system". Yes, Sir. Yes. That's right. For now accept agent as advanced Salesforce user, who is perfectly able to to work with him. You don't must hire dear team consultants to understand Salesforce.

The agent knows how it is make. So, he can reveal the full the value of Salesforce and all these functions. AND when will you start to trust one's own agent, you just press the button « always allow", and now the agent will be able to fully use Salesforce capabilities. So , actually, right there everything is moving. But now arises question: how SaaS- applications fit into this wave of AI and agents? Needed great API and great CLI, interface command line, through which agents will interact with SaaS tools.

I think earlier majority developers SaaS products spent a lot interface time user, and now they really deserve it to think about agent interface , and not just about interface user. You must create best agent interface. And, I think , you must to abandon the idea, that you are completely will control relations with customers. You must consider agents, created by AI- companies like expansion of your platforms, and not to feel threatened by you are losing relations with customers. If you you can do everything this will become a source truth and system accounting, then you have big ones will appear opportunities.

I must to say, he is one of most underrated managers from product and sales in our industry because I talked to him in Davos, Sachs, about Slack. And he asked: " Tell me what you are. "You are busy." And I replied: "Listen, You have all my details. I I can't get it. access to your data . My team wants leave Slack. They want use this open-source product itself so that I can get all this information". He, again, how are you said, Friberg, connected me with sales team, and now I have a full administrative access to our Slack.

Slack built a product called Slack Today, which they gradually are being implemented, and he keeps promises OpenClaw. And what he did Sam, having redeemed founder of OpenClaw— immediately after he bought the founder Peter, and he began work at OpenAI—OpenClaw just went out, no more innovations. I I'm not trying to say, that Sam did something bad or what OpenClaw terrible, they just lost their advantage. And now Slack Today version is about seven out of 10, and I I notice that she is becoming 10% better each a few weeks.

Therefore, he will do parallel connection, Sachs. He will have own agents, and he will allow you come with your own. His agents can a little behind, but you too you will receive them free inside Salesforce. It like two paths to victory, you understand. Listen, I think SaaS- products can to offer their own agents, own AI, co -pilots and all that. But the reality is, that most actions with agents will take place outside these products, they just want interact with them, receive data, use them as an accounting system, source of truth, and then record actions back when it appropriate.

And I just I think that the founders should support this trend, rather than fighting it, because if you will fight for control and say that only our agents can interact with system, then people they will just leave you . So you're playing the long game. with vertical SaaS, or you play on downgrade, as we Friberg? I think it depends. from a specific company, its features, system reliability accounting and other factors. No, I don't think so. vertical SaaS has accounting system. I am not I think that's what they are engaged.

I I think that their workflows are the key point where such horizontal monolithic companies, like Salesforce, unique; Workday, Oracle with their ledger, this is very difficult, SAP, IFS, It's hard to see the way, how can i play these businesses so that they became a repository and single source the truth, as you said. But vertical SaaS— these are work processes and processes, and I just curious what's up will survive this. Well, listen, I think, that's a good question, and it all depends on situation, but just look a few months ago everyone said that horizontal SaaS is not will survive.

And one of main positions Leopold in that fund consisted in playing on downgrading all these SaaS -products, including Salesforce. And what we we see now, turned out to be a mistake or not at all irrelevant. Honestly, I don't I wonder if he understood. the difference between vertical and horizontal SaaS, because that was a problem financial short sale that is, he needs was to choose big stocks to short sufficient volume for getting what you want shoulder. So I think that he could just to adjust everything to your own logic, to make it sound right, but I I suspect that it was trap through leverage, where he thought: me must have a massive position on growth behind others assets.

And his prime broker said: well, you need to find shorts, liquid shorts , and he replied: okay, Adobe and Salesforce. Well, but listen, justice for the sake of, I think that this was dominant a few more statements months ago during "SaaS apocalypse", and Benioff shares fell heavily on market. I mean a 50% drop, and now they won 20% back. Therefore, I think, that many people just indiscriminately sold out all SaaS. And I think if I had single out one conclusion from today's news, so this is what simple extrapolations of the future, based on these trends, do not work; idle time extrapolation here was: AI agents can write code, so that's all software software will depreciate.

It was oversimplified extrapolation. Now we see that actually exist certain accounting systems that extremely compatible and complement AI- agents, and if the founders are right they are positioned, they can use this trend. By the way, I I think I narrated. about the loss of workers places—this is another one simplified extrapolation. Wherein consists of a narrative about job loss: now AI can perform intellectual work, so everyone intellectual workers will lose your work. It completely ignores that the fact that there are ways for employees intellectual areas to take advantage of these tools and adapt to them.

So just like Benioff uses trend for agents, intellectual workers use it and will be able to adapt to these changes. If you are implementing technology, you survive. If not— You will lose. I have on mean that everything is enough simply. I also wanted would like to point out incredible Nvidia results. Two things that happened very quiet. Jensen now clearly does betting on open-source. He bought Hugging Face for 12 billions. And he made one of those deals from Poolside for 6 billion, about 20 in total billions on acquisition that is much less than 1% of the market Nvidia capitalization.

AND thank God that in this the country can again to carry out mergers and absorption. But now he has a level distributions for open source. If you don't you know Poolside, they created a model for Laguna coding. So now Nvidia has a place, where developers and hackers go to find out which one open Sorry. Sorry. I am not I think that was all. That's right. As far as I am concerned I understand, because Jason Warner this morning tweeted about the release new model. I was confused by this, so what could they do buy two different ones things.

100 engineers, which are moving to Nvidia, can work on what is at Poolside, and this, essentially a version of Cursor and Cloud Code, which runs on local servers, which works in open source, open scales. No, they also have model. I mean that they have a model, and they also have this agency a platform that can manage encoding. So, I didn't have clear from advertisement. I am not read this carefully. Yes. What exactly did they buy? I think, by the way, we we hide the main thing.

It seems they are, after all messages, conclude these agreements with Poolside and Hugging Face, and this large acquisitions in in the open-source field, if message true. But look, the main one news today, how to me, this is what Nvidia showed stunning quarter, making over 100% per annum calculations. A billion dollars income every day now. Yes. Almost 100 billions of revenue just for a quarter. $60 billions of profit. 75 %gross margin. I have on mind, if you exclude one-off articles, it was the biggest profit, ever received by the company in history.

Jay, let me to emphasize what you said in the introduction. Yes. What really, I think, is the reason why stocks today increased by 8%, this is not just a wonderful neighborhood, but also that they predict 70% growth of the next year, while the market I only expected about 45%. So this is quite amazing. AND this figure of 70% could to be higher, but they said they were limited in supplies. Yes. So, you know what The main conclusion? I I think for a wider market this means that The AI ​​boom continues.

Another narrative that is collapsing today, is that these capital expenditure in AI it's a bubble. That she is supposedly soon it will crack. And what Nvidia actually says with their numbers and predictions, this is what these capital AI costs will last a long time in the future. They have real support. AND It's amazing that after these numbers, even though 8% growth, Nvidia only trades with by a factor of 12 to profit. Yes. And the reason is, what do people think about competition, obviously Cerebras, and the moment that I was going to voice earlier.

I think that Jensen does bet on open source, judging by these two acquisitions: Hugging Face and Poolside. And I think, Sam Altman got irritated tiger. If to look at what he did: he concluded that huge deal on one hundred billion dollars from Nvidia as early as 2025 year, and then announced an agreement with AMD, and Jensen was clearly dissatisfied with this, and he expressed himself immediately after Sam made this statement, what is going on create your own chips for output data jalapeno Jalapeño, and what is it?

was going to conclude agreement with AMD. Then Jensen came out and said: "Well, by the way, we really don't be sure to do these investment in OpenAI", and I I think he is accelerates open source, and next year you you can buy from Nvidia is just all its own stack. He's going for it. He is going seize open source. And this really interesting a moment in time for us, to look at him, because we were sitting here 6 months ago, 12 months ago. Oh, open source never will catch up.

He is working on a project unmanned control with open source, and now also owns Hugging Face, leading a platform where everyone trying out new ones models. I think actually it is the most important story . I believe that the general the conclusion should be like this: we'll see how all these businesses merge and all of them will compete between yourself in all possible ways directions. So, the idea about what exists clear division into client and supplier, in my opinion opinion disappears. Nvidia still significant depends on hyperscalers in in terms of their income.

It seems it was 24 billions of dollars for quarter, but they now created Neocloud business, which is equal to this amount, and It's incredible, isn't it? Yes? Yes. And I think you'll see as more and more people will try with compete with them. So now your customers, hyperscalers, say: "We we are going develop your own chips". Quite logically, that NVIDIA says, "Well, we we are going develop your own models". They are do. And then you you will see how they rising up on the stack and hosted model.

Then you you will see how they provide APIs and services conclusion, and they will become cloudy provider. So, Jason, to yours. remarks, I think, the right way thinking is like this: in business 1.0 you earlier was a business capital expenditure ( capex), right? And people asked: "Okay, What is your payback? investment?" Or you was software business and thought: " Okay, what are yours? operating expenses ( opex), right? And what your payback investment?" That's all merges together. Every company will do everything. Look at these large companies in five years.

IN all of them will have their own cloud. All of them will have their own models. They will all have own chips. Everyone they will have their own data centers. From A to Z. They will cover everything from above down. And then the question will only be in because it's better. And I I think that this actually better for broader ecosystem. It better for all of us. Certainly. OK, let's move on agenda. Nova Besson's strategy regarding bonds criticized him Mentor Stan Druckenmiller. Last week 30- annual treasury bonds reached 19- annual maximum in 5.3%.

You talked a lot. about this, Friberg, because I am interested to hear. your opinion. IN Besson's answer doubled the amount of redemption long-term bonds Ministry finance from 2 to 4 billion dollars 19 August. CNBC reported, what Bessent is considering possibility increase in purchases bonds, and Fox Business stated that he wanted to strike fear into traders who play downgrade long-term bonds. So, here a lot is happening interventionism. Stan Druckenmiller, former mentor Bessenta, published an article in The Wall Street Journal, where opposed him plan. His main the argument was because Bessent is mistaken, trying manipulate prices, instead of eliminate the root cause profitability growth .

America has a problem with expenses. We talked about it again and here again. 40 trillions of dollars. No more Doge. No more cost reduction. AND, Of course, we have a war, which is quite expensive expensive every 5 months. Friberg, seems to be the past episode you said that we add a trillion. So we add 2.5 trillion per year. We let's add 10 trillion during this second term Trump. What are yours? thoughts, Friberge, on of what is here takes place on the bond market, and disputes between Bessent and Druckenmiller.

Chamat, we are with you. we can also to talk about using AI for writing articles, writing yours articles. Yes. I'll start with these relationships, as you noted. Bessent and Druckenmiller closely worked together, they close and respectful each other, as well as Kevin Warsh on similar position. He worked with Druckenmiller the last few years in his office every day. They are extremely close. It triumvirate, and I don't I think that in history there were three people who so close to such positions. One manages the market activity with Drak, another leads Treasury, and The third is the Federal Reserve.

It enough to have an unprecedented such closeness to such positions. AND, You know, I think... What is it? Yes. Right. Caesar. THERE ARE... You named it. triumvirate. Triumvirate. Who is Caesar? Who is Pompey? And who is Crassus? I'll leave it to them. discussion. But if open the curve COVID-era profitability, I think it's worth it. look at the curve profitability treasury bonds. So, this is shows depending from the term: this one-month or 30- annual treasury bond? Therefore, throughout this time, which yield? You know, how much is federal the government has to pay percent, so that to force people lend them money for this term?

And this approximately the era Covid, somewhere in 2020. You can see, that because of low interest rates and the fact that inflationary consequences stimulation that arose during Covid, not yet influenced the market, we observed indicators below half percent, below quarter of a percent up to 3 and 5 year olds bonds, and then slight growth, when was it possible get 30 years old treasury bonds at 1.7%. Again, looking back into the past, always easy rewrite history . To the Federal Government it would be worth it to translate all your 30-year debt term at this stage .

But the debt balanced by this whole curve treasury obligations. So let's see, where is the curve located profitability treasury bonds today. So, that's how much the market takes from federal government for borrowing money for paying his bills. And at the bottom end Treasury curve you see a little less 4%. So, in order to the federal government could to borrow money for one month, he has to pay creditors 3.8% for this money. And if consider a 30-year-old deadline, then now he is 5.2%. Therefore, now for federal government The USA is very expensive to borrow money AND the reason for this— stable inflation.

I would said that it was because excessive government social spending programs and more. AND big problem now is that, that the federal government spends so much, what if he started aggressively reduce costs, then the argument and the concerns are in that it will hit by unemployment rate and will cause a recession, because the federal government is too important part of the economy. This is the main one argument. But this causes inflation and scarce financing. Therefore, this year deficit will be approximately, let's say, two trillion dollars.

And in market result says: "We are concerned fiscal solvency USA in the long term perspective", or " There is a higher risk, so we are going to charge you a higher price percentage: 5.2% for 30- annual papers". Now what does this mean? for the federal government? Today average cost federal debt government is 3.4%. This is what we pay as a percentage in average over 40 trillions of dollars debt that has federal government. On every shift interest rate by 1% the US government should pay 1.25% annually GDP additional percent.

1.25% of GDP in as a percentage every year for this change interest rate by 1%. And now we we see 5.2% for 30-year-olds bonds and short-term rate of over 4%. Therefore, from the federal government there is a problem because within the next 12 months they need refinance 10 trillions of dollars debt. Term repayment of this debt is approaching. The validity of these bonds are expiring. They must return the main amount to the owners bonds, and they will have to again to enter the market treasury obligations to sell new valuables papers and borrow funds for refinancing.

So cost of borrowing will now grow. AND when this value costs are increasing federal government are increasing, and budget deficit is growing. Therefore, mine theory and my argument consist in the fact that Bessant cannot use no measures that would really significant influenced long-term part of the curve profitability. Now in we have a fundamental problem with fiscal expenses federal government. AND I think the note, which was written by Duck— regardless of whether Did AI write it or he himself has none value. The essence of this The notes are correct.

AND she is called to provide cover for Bessant. He says that it's not your fault Bessant and not him responsibility- solve the problem yield curve. It responsibility Congress. It responsibility the president. It the duty of those who responsible for the budget and bills, get up and to say: "We we are going to cut expenses, because if we we won't do this, we A spiral awaits, with which we will collide within the next 12 months when will have to refinance 10 trillions of dollars debt. So, Bessant does his job as Minister of Finance.

He says, "I'm going to market". "I'm going to buy treasury bonds." "I I will try to lower their yield". But he has only a maximum, maximum, maximum trillion dollars, which can be used for debt redemption. Him will have to anyway sell treasury bonds for 10 trillions during the next 12 months. 10 trillion. He must sell them. So even if he maximum will use his/her authority for purchases in the nearest in the long run, it's just trillion, and then he will have to sell 10. So the market is the market, and Dak's note is cover for Bessanta, she returns responsibility Congress and to the president, saying: " Guys, we have to solve this problem with a deficit." That's it.

of a kind histrionics between the two of them, and they as if supporting each other? I don't think what are they allowed to do communicate. I am not I think they are communicate. I am not I think that Daku allowed to speak with Kevin, or I don't I think that anyone with they are allowed to talk to someone one. So I don't think so. that this is happening. But I think Duck sends to the market signal, saying: " Guys, this is not responsibility Besson. Got it.

The burden is not on him. shoulders. He can't to be Atlanta. He is not can lift this earth . He can't decide. this problem. You must act. I think this is protection for Bessa. Got it. Good. That's what I'm for, actually, led. Chamat, What is your opinion about what will happen next? We have already passed the 5% mark. Or Will the system break? How will this affect interest rate reduction rate and how much everything can go wrong if we can't take it under control ? Tactical response on your previous questions about why is he buying bonds on the market: This is called curve suppression profitability.

That is, when you buy bonds, you raise prices, and predicted profitability is falling, which allows Scott be financed from lower costs for loan, because if the yield is 5.1%, not 5.3%, then issuing bonds , you are focusing on these 5.1% versus 5.3%. Therefore, this is suppression profitability tries to help with that, about what Freebrook said, and right in this tsunami refinancing, which is approaching. Good, but listen, let's take a step back. I believe that main problem is that Congress eventually responsible for costs. Congress completely unable to do something, to take under control these budget deficits.

It is happening constantly. This is not Democrats' problem or Republicans. It Congress problem, that is, both sides constantly spend much more than It would be worth it. So, in our debt is growing by 7%, and GDP—between 2 and 4%. It a recipe for disaster. That what Duck said in that note, —need reduce social benefits. He is probably right. The question is when? And who will force United States to resist this problem? Who will force US Congress to face this problem? And he says, that the bond market will start doing it, so that this single price is real a reflection of good of will and trust in United States.

AND because this profitability increases, simple rule says: trust falls. Yield increases, trust decreases . Another thing I would like to say: this takes place in extremely awkward moment, when we should to be in in the middle of a huge financial development to support AI. AND if you go back to a hundred years ago, you know, industrial revolution, great agreement, all this is government The US was financial guarantor, right? They were the ones who who could intervene. AND, unfortunately, because of your financial condition of the government The US can no longer to do this.

Here why, thank God, in We have such companies, like Nvidia, Google, Microsoft, Meta and Amazon, who take this burden on oneself. I wanted to relate it to another question, because in the internet now so much noise about Nvidia balance and all other nonsense. AND I think people are completely don't notice that these The boys are pulling on the entire economy to themselves USA. A hundred years ago this the US government would do. But the US government is not in able to do this, because Congress cannot to organize your costs.

President actually did something incredible on final stage these trade agreements. If you remember, he managed find 2 trillion dollars to pour them in into the US economy for support this development. But practical reality such that these two trillions coming in from other countries, and their processes are long and confused. So we have a situation where all these money must be invested. This has to be guaranteed by the US government. But he doesn't in that financial condition to do so. The president resorted to some magic to make it easier in short-term perspective, but this only temporary plaster.

Real solution, as it says Friberg, consists of because Congress has unite and understand that problem is necessary fix it now. Because if you see profitability if you see three-year-olds bonds at 6%. This is the beginning. a deadly spiral. This will not happen. instantly, understand? So don't panic, but this is the beginning extremely painful process. AND this pain will last for years. The only thing that can fix this now— it is control over budget, and this Congress's task. It It has to happen. Is this year, or next, it is simply necessary make.

Sachs, President quite a few talked about cost reduction beginning of his deadline, but we have little what did you hear about this from commander-in-chief lately. Or the problem is that no politician dares to touch on the issue cost reduction? It too toxic, because Trump is on to his last deadline. He " lame duck" for by definition, or not Yes? His second and, hopefully the last one deadline, if it is not decides to go to third. Uh, but he's not spoke out about this. And is this too much " "radioactive" topic for anyone in politics?

And to you had to spend It takes a little time. You tried their hand at it in Washington. It too toxic for anyone to to take on cutting expenses? He sent there Ilona. It was, you know. , quite sharp. AND this, of course, now does not happen. We we do not reduce expenses are massive. In fact, debt is growing faster than a year ago. So what are you? Thinking about it, Sax? Maybe you can to go beyond discussion only Trump. Yes, look, This is partly a tragedy.

shared resource, so you have 435 members of the Chamber representatives, 100 senators, and each of They have a voice, and there are still president, and none of they are not controlled expenses. Actually right to dispose of finances take beginning in the Chamber representatives, and then goes on. Ago very difficult control costs when so many people express their opinion. If it were only a few people who gathered in the room and solved it, they, probably could do it make. But when so many people have incentive to protect their programs, it becomes very difficult.

That's what consists of part tragedies of the common resource. And reality such that the president does not has veto rights on individual articles expenses. It would be great if I had one. He does not have constitutional full powers to tame it. AND despite what they say Liberals from the opposition, he is not a king. That is, he doesn't have any powers. This, according to essentially, the root of the problem . And you're right, he supported DOGE until end. He is the only one president who once supported something like that.

And what happened? Elon has arrived until the moment reduction of one department or one subdivision. Hysteria was so big, that's it accused of killing millions children. It was complete. nonsense. But Tesla dealerships started throwing incendiary mixtures. And this is another one part of how to carry out reforms? As reform spending? How to tame them, when the media is not on on your side and not wish to highlight Is this true? It the worst part. I have meaning that the media does not have a political incentive for mutual concessions, but they absolutely nothing they do so to cover the story true.

Simply think about how the media reacts to everything it. So it really is a difficult problem, when you start with that no one does not control it fully. You know, this truly a tragedy common, and the media, by in fact, do not perform your work with coverage of this properly. So I, You know, I don't know how. us to achieve reforms in on this issue, and I will only say that if Republicans will lose control over Congress, I don't I think this is something. will improve because you you see now, that political energy in the Democratic parties in full aimed at these huge new spending programs that wants DSA.

So, you know, if you vote for "changes", choosing Democrats, this is also not will help because they they want again only more and more more expenses. So I I don't know what this is. decision. Faure, give us time frame. Obviously, Republicans like to discount taxes. And the Democrats like to enlarge costs. And here we are. It didn't get any better than this. three or four presidents. The last one who is this did, was, seems, Clinton, who balanced the budget, but the situation is becoming getting worse.

Certain To some extent we all agree. Sooner or later it will end, and we will have to implement measures austerity. When this happens, on your opinion? Is there critical point regarding debt ratio to GDP? Maybe people on the other hand just will say: "Hey, I want 12%". You have a bad one. debt. I want 12% per annum behind him. And we just Let's say we can't. pay 12%, and we will have to go to full hard savings. Is there a point? break? If so, then When?

Well, the question is so what is this for will lead? We, as democracy, we have recursive solution: when people feel pain, they react by voting. It takes us back to the question is, will we slide? we are for socialism, because this is exactly what is happening after intermediate elections. So I think, somewhere between the age of 26 and election cycle of the 28th the problem of inflation, and remember that inflation problem fundamentally rooted in government spending. So, the scarce funding has very clear correlation with the cost of housing, health care, education and everything another.

So what? the government spends more on different things, so they are more expensive become. And therefore inflation is stable, because what we thought temporary extraordinary measures stimulation during COVID, very fast normalized and became permanent cost cycles. You know, if we returned in 2019, we should actually have budget surplus in the country right now. Our economy has grown so much since 2019 year, what if we had same level expenses, as in 2019, our federal government would receive profit, repaying debt. Reality in because we transformed these extraordinary expenses on a permanent basis.

By the way, we did the same after the world war financial crisis 2008 year. They were crazy. incentives, and we continued these costs. So the question is, Jake, is because what happens next? If inflation will remain high, but purchasing people's ability will fall, they will vote for the one who proposes decision. And I don't see political message , which would win, saying that federal government now we will spend Less. We are going to reduce the pile everything. Therefore I I'm worried that the only The solution here would be, which is very harmful individual freedoms, many rights and many things thanks to which the free market works.

and these are some of these more socialist principles and socialist candidates. Assets. I think this is kind of critical point. I think that this rather so-so spectrum, and the reaction under this transitional time phase, from the 26th to the 28th year, can be negative. The next big the event will happen approximately 2030–2032 years. This is the moment when in the system social software will run out of money for payments. And I think that in this era many states will become bankrupt or not will be able to perform your obligations.

AND when states cannot perform obligation around 2030–2032, will happen as salvation and restructuring social provision, and salvation and restructuring states. And maybe this There will be a time when all this will break. Oho. So you're saying, what will we see municipal defaults ? Uh, yes. And people in states will announce default on their own debts. This would be catastrophically— to save something like California. Well, because obligation for unfunded pensions in the states so significant. And this is not just to say about debts. Listen, California spent tens of billions dollars for this stupid the railway that leads to nowhere, even where there are no tracks.

And they are discussing political will to at least stop it. It it would be easiest reduction in the world. Just think about it. this idiocy. Nobody says we shouldn't to do this. Why no one says that we shouldn't do this do? That's what I'm saying about the media. A trillion dollars. Everyone who spoke behind the railway to nowhere , it would be worth it to be ashamed until it will not disappear. Trillion dollars for train from San Francisco to Fresno, which one is more expensive for a plane ticket, — This is just idiocy.

It's like in the movie "No" "Look up." Type , look up. It just absolutely pointless. Turn off this thing. Why no one doesn't even say it simply turn it off? It tells you everything that need to know. Private company built Brightline in Florida, and they, seems to have understood, how to do it for share of the cost. Then What are we even talking about here? are we talking? If it were was a good idea, private industry it would implemented. Yes, Friedberg. Private business would see opportunity and would build it.

Here there is none opportunities. In general, these government programs cause more harm than benefits. When the government interferes with guarantee student loans and gives credit to everyone, administrative expenses increased by 6 times, and the fee for tuition soared by 8% every year for 30 years, because the government said, which will guarantee any loan. That precisely with housing, and now no young person cannot allow own a house. The same with healthcare, where they said: we we will pay people for that they stayed at home and didn't worked, taking care of others, and we don't we will check whether really they are do.

And the cost medical services flew to the sky. If you take any of these government programs, each of them has the opposite effect— inflates costs and inflation. Root inflation in this country . The main reason this—state costs. And the reason, why are we finally we will become socialist country, consists of because we are not we restrain state We don't even look at the expenses. in the mirror, saying: " Guys, this is idiocy. What are we doing? And now markets, supporting Chamat's opinion, they say us that this is idiocy.

We will no longer be to finance this nonsense. And to us need, to your thoughts on media, Saka, Nick, Shirley, "60 Minutes" "or anyone to they simply said: hey, these expenses, this fraud- are unacceptable, and we need to send people to return this money. It's like we we burn the money we have we are not there, — Fraud is the first level. There are about 18 here. levels. First level —this is fraud. Yes. Of course. you know, and in view of the tragedy of the common, structural nature this, where there is none no one who would controlled it, need a huge public pressure to it is to master it, and its not at all.

And there is no no effort on your part mainstream media, to create such public pressure. So, how about this? to get out at all? AND I don't think so. is happening right now Yes. This should be a critical moment. Yes. You say that it has a crisis will occur, and then people react. But, returning to in Friberg's words, we we can get socialism, because people may not be connected real reasons crises. And Friberg, this is mine question for you. No, socialism was much more likely 100 years ago, when everything was less interconnected.

America cannot just solve it make. Do you understand Why? Because all our debt related to internal and international players. This is not just a solution. None such a button, which you can press. AND the reason is that you would go bankrupt for one night. Therefore I I believe that reality like this: we will hit the fiscal wall. Something will break badly, and, like during COVID, or under world time financial crisis, or September 11th—these are the only moments when you can change expectations structurally. Structural changes that most influenced for our lives, took place in these sharp moments: 11 September, financial crisis and COVID.

Yes. And each of them turned into "never again" moment "waste the crisis" when it all came down to arguments on benefit of expansion government power. What was your question? to me? Okay, let's take it, for example, housing availability, Yes? It depends on mortgage rates. Now I'm looking at average rate for 30- an annual mortgage—this 6.73%. A little less than 7%. It approximately 150 basis points over 30 years old treasury bonds, right? So how Do you think it could be? average person connect the facts that higher costs mean higher rates for treasury bonds, which leads to higher mortgage rates pond, and in As a result, you are no longer you can allow for myself the same housing, because monthly payment becomes much higher.

That is, do people see this logical connection in that is happening? No. No. But people want someone to blame Every time, when a crisis occurs— in this group, even when are we going to podcast—happens something bad. We always put question: who is to blame? This is the first thing it does man. Yes, exactly. And this is literally... human. No, I mean that this is the human brain. Here's how it works. There is some kind of problem. Who we are shall we blame? This is the first What are you thinking about?

You don't think. objectively about how solve the problem. You You think: who do I need? to blame? This is the first what you want to do. And then the question is, when the rates for mortgages are growing, Who are you blaming? The current president, my congressman blah blah blah, reselect someone. That's all. respond. AND a person appears, which says: "I will give you free housing. I I will give you a free one. bus. I will give you free mortgage. I will take the land from landlord and give "It's yours." You are voting for this person.

So I I'm worried that mechanistic reaction to what you describe, Sax, and this is very problematic situation for people— absence housing affordability— is to to give away the keys to kingdoms to him who will throw wood on the fire . And fire is government spending, and fuel is new government spending to try to "solve" » the problem that caused the fire itself. And I think that's exactly it. me the most worries. So, I think I know. how can we blame Dario. Okay, let's get back to this.

What did he do? That's the thing. Our only hope is AI, right? Because only AI can create exponential growth, necessary, Yes. for the economy to grow enough for us could grow out of this problems. Yes. And if we we will contain AI, creating a new one regulatory apparatus, which takes years for the approval of new models, then we end. Completely, as Canada and Putin. It those three people who are working. I'm just kidding. We Just kidding. We are joking. . Dario, come on. This is not Dario's fault.

But in in any case, Come to the summit. Continue. No, seriously, Come to the summit. Listen, I want to say that we we count on AI as on deus ex machina to get out from this situation. And to us I really need it. I mean, we need more such neighborhoods as Nvidia and Salesforce, which just reported. To us it is necessary that this It took 10 years to grow out of this mess . One of the positive parties is that people migrate within the country or to other countries as way to vote against uncontrolled expenses.

I lived in New York. York, Los Angeles and San Francisco. And I didn't realize how angry they were these socialists and " limousine liberals", until I got to Texas. After a few years in Texas I saw what it gives to the family and an individual the opportunity to have affordable housing. In those In cities there are people who so unhappy because cannot allow own housing. They they pay exorbitant rent, and people who already have their own Victorian houses, etc. in San Francisco, refuse build. And in the wonderful state of Texas, if you look at this schedule, housing prices from the peak of 2022 to The Austins fell, fell, and have not grown, Saxe, fell by 27%.

Since March In 2020—by 25%. It It is possible to do. Need guidance , which will declare that Housing is a priority. Therefore, we will allow you build more houses. This is already is made in Texas, in Florida. There conservative, free market conditions allowed them build Bright Line. IN they are high-speed trains running 125 miles per hour, because they stepped aside and allowed private industry to find out, how to build high-speed trains. It clear problem certain groups that want to stop progress. They hinder creating housing. AND when people have housing, they feel perfectly.

People in Texas happy to have house for 250,000 dollars. They are everywhere. These problems can be solve. Needed only leadership. AND This is exactly what we don't need. there are enough leaders, who will say, "Enough is enough already". We just let people build more houses. OK, moving on agenda. I love beautiful JCAL controversy. I just terribly angry because this is because you know... This is complete nonsense. that... it is in some sense meaningless, but we we walk around here question: who will be that a leader who will say...

You, you prepared this. in advance? No, I didn't cook, I simply... This came to you in head just like that? Yes, this is another one... Yes, actually I am hired the author Sax, so that he can correct mine texts, but if to reject this, then this an obvious problem leadership, it's too expensive, true. I will give it to him. clock, let's do it a small indentation. What What are you thinking about? Druckenmiller wrote your article for with the help of AI, it caught, and Wall Street Journal said, "Hey, this "It's perfectly normal." I thought it was leftists are engaged in " virtue signaling ", but actually I I think that's it.

a special class of people, who do it. Okay, let's go. They judge, they condemn and not see the forest behind trees. And this, probably , the loudest screamers in the gallery. Wall Druckenmiller— the most incredible our investor generation. Perhaps, sorry, maybe is in the top three or the top four. He is among them. Yes. So simple Shut up and read, what he writes. And you have thought, and you're like, "Oh "Oh my god, he has a dash." "Oh my God, he used artificial intelligence for decorations.

It's like you You're missing the point. He tells you what It has to happen. Sax, do you agree? You Do you think that copyright columns in the Wall Street Journal Does the New York Times have to write artificially intellect, not a person whose name is stands under them, even if they all still written by him? He He didn't say that it was written by him. He admitted that it was 100%. It creates queries. He shapes it. He directs it. Yes. It will be more accurate to say: his opinion, but then he uses AI to help her write.

Listen, I'm not I understand all this. demonstrative morality with about this. I have I mean, it has to be quite stupid in our time not to use AI for help in writing, and he definitely not stupid. It his opinion. This is his view. This corresponds everything he said for decades. And, actually, the article quotes many times and he is right. Yes. In the article much is mentioned cases in the past, when he spoke out about this. I think that's clear. came from him heads, and AI he uses to help this write.

The fools in X would have say: “Wait a minute. How will this affect my company? How is this? will affect my portfolio? How is this? will affect my pension program? What will do this with pensions my parents? How do I to think about risk insurance? » Instead of this conversations, you are all like this: "Oh my God, he used AI." Jal, you have it with this. problem? You are one of those who have this problem? I really think so. it's like a phonogram in writing texts for the author.

Me unpleasant when the system expresses all your thoughts, on unlike using it for research that It's perfectly normal. Use for proofreading, it's quite normally. But for It's like coming to me. Adele concert, and she under the phonogram. I I really want to hear. its. Don't use Excel. Contact only those who does manual calculations. No, no, no. That's what case. I would rather he said that Adele uses Excel. Not use Excel spreadsheets. I am not I trust those guys. I only trust those who who does manual calculations.

But if you are you going to express your opinion, and write not you, it's for me disappointment that this not disclosed. I think this is much higher standard—to say: I I'll write it myself. voice. Of course, someone else like " literary Negro", could you help me to polish it. I could check with their grammar. I could do research is everything the right ways use, but I don't like it. the idea of ​​what you are writing something with the help of AI, and I'm reading this and I don't know about this.

I like it. seems unpleasant and insincere. I would prefer, so that people immediately They said: I don't do this. wrote, but I generally agree with that This charming one gave AI box, which guesses the words. This is me. personally insulting. Yes. Interesting thought. Can you answer? Of course. I tried. to know your opinion. I'm listening to you. Yes. Continue. Wherein, in your opinion, the difference between this concept and requirement to artists to reveal, or they used Adobe Photoshop and which ones filters there applied, or to DJs who use specific software provision for creation electronic music instead of recording it manually on guitars, or to people who, for example, using Microsoft Excel to create tables, not write everything down paper?

Should I? to reveal that used Excel, calculator, synthesizer or Photoshop, instead of make it analog way? Eventually, digital tools for creativity and creative result working in parallel with human ingenuity. AND human skill and creativity are expressed, are increasing in in many cases are improved by digital tools. AND determining whether someone "uses" AI", just as blurry, like the rest of the examples: "am I using AI?" » to improve spelling, grammar, create new ones sentences from scratch or new paragraphs from scratch? After all, if I give AI a bunch of input data, I take the results and edit them, or " "Am I using AI?" This, I think, is one of fundamental problems: we treat to AI as to autonomous, simple replacement systems intelligence, and this heuristics are not accurate about how people actually it is used and how this, like all other tools, there are increasing human creativity, ingenuity and potential.

Continue. The reason is reader's expectations. Readers have high waiting, and this important question, namely so people were so disappointed in this on a large scale. It not because they are engaged in demonstration integrity, and therefore that they really are appreciate Stan's opinion and want to consume it's hers, not an artificial one intellectual useless That's why people felt themselves betrayed because What did he do? Now this is new betrayed I think they are. Yes . Because they too felt that Wall Street The Journal betrayed them, because the publication would have to check whether what submitted, is not AI- useless.

Same, as professors have view works students and to make sure that their essays are not AI- unnecessary, because we we want them to learned something. AND they want the same thing, when the teacher teaches. People too don't want to pay to the teacher who uses AI- useless So we are like we have society decide when we I need this to work. from a person—100% or by 60%—and when we are ready to trust AI. I want to hear it. real opinion. IN I have a question. Many questions.

OK . Question number one. You assume that Stan didn't read what was generated, or do you think... I'm sure he is. read. Yes. OK. So, when he read this to...Well, it's open question. Did he read? OK. I think he... Well, he defended it. Yes, he read. I hope so. OK. So, you have claim to that the fact that he I read this and thought: "Yes, that's what I think" . Boom. And pressed "send" button. It Is it treason for you? I think that non-disclosure—this is there is betrayal.

If you want write if you Do you want to give me yours? opinion, interpreted through magic box AI word guessing, you owe me about to say this. I understand. I think it's difficult. to inform everyone. I understand. Now I I just want to understand I want to do confession. Er... Oh, listen, what are you... Permanent warning for all future publications that I I release, including tweets: I try not to publish nothing, without missing it first through AI for fact checking, editing and help in forming the best argument .

So it's permanent. warning: if I wrote something and published, this, probably passed through AI. I am completely I agree with the correction. I completely agree with editing. That it was clear: proofreading, checking facts—wonderful ways of use AI. Audit grammar—wonderful use of AI. Now you arguing about the limits of how much he checked the text. These are two different things. I I understand that you don't understand this concept because you don't professional a journalist like me. But I I can say that there is big difference between proofreading and when someone writes for you or plagiarizes .

And that's exactly what it's about. it is said. That's why people deny it. Question. Is Stan there? Duckenmiller his next article? Or he is allowed use computer? Of course. Of course. OK. OK. OK. Allow continue. Or he is allowed use text editor, or just Notepad? I mean, you Do you understand that he can use any of these tools. But how about what he had would you like to report this? Wait a minute, what about Grammarly? Is it allowed? him to use Grammarly for I encourage him.

use Grammarly. I am a shareholder. Grammarly. Jacob, how are you? What do you think his was like? the process of working on this post? You think, he was just given ready material, and he said " publish"? Or That's exactly what the software said. for AI detection: what this is 90% written artificial intelligence. I would be shocked if this was his process. Here is my guess. This is just an assumption, because I don't know for sure. This is just a guess. I I don't know. My opinion: in he had an idea, and he probably voiced it AI—written or oral, or, you know, basically dictated it to his assistant, who introduced this into AI.

Then he probably got the result, made adjustments and edits, and that was looks like his work process. I don't believe that. it was just spat out like AI garbage, and he said "print". imagination that happened. He is sitting in at its office in New York York. He sits down. in front of the browser. Opens ChatGPT and writes: "profitability "has grown—that's bad." Gets all this, printed and sent an article by email. if you can't to assess what the essence has to leave him/her If you can't evaluate the analogy with phonogram, this How can you open mouth to the soundtrack?

Okay, here's the problem with analogy about phonogram: he criticizes what just happened happened. It was politics. share buyback, right? He criticizes his/her former colleague about something very specific, what did you just do Minister of Finance. What another phonogram? A phonogram is when you sing along to the song, which was performed a thousand times. That's it something new and unique . Another testimony why he had write it yourself. It's about respect. audience. You must respect the audience enough to find time to write this personally, not allow...I was part of the audience.

You can have another... I...I thought, "Oh, this useful information for me regarding... absolutely normal. Just a reminder." Hey, listen. Some people difficulties with profitability. Wait, you can. to ask if you have read article and do you have opinion on its content? Yes. Well, I wanted to to find out if there were they are coordinated, and Was it some type of... did you know that generated by AI, to how someone wrote about it in tape? No, it appeared in my feed At first it was like...no. That's right, I read it.

this is the first. Wait, you gave me question, please at least for me answer. This appeared first. I saw plagiarism, AI- garbage, as if it were people not called in journalism, like this this...let me to say, I am read. Wait. I read this when It just came out. I didn't see it in the video messages about what is software provision of alleged discovered this. Tape me ahead of. That sounded quite agreed with many other cases, when I heard the doctor Miller speaks of debt problem, you know , in podcasts, written versions or somewhere else.

I didn't know. So, if he is in plagiarizes someone, then he plagiarizes oneself. Yes. Listen, plagiarism— that's not quite correct analogy here. She is not perfect, because it's something new, but this is appropriation thanks for... Wait, give it to me. to finish. He takes credit for writing AI. AND I would prefer so that he finds the time and respected the audience, having written it myself, because I respect I want to understand him. his true opinion, rather than interpreted through AI. Same, when I hear someone sings a song, I want know that they are not use autotune and not subjected her post-production to absurdity.

This is mine. position. I believe that the artist must to give one's true self art, to express one's the real position, not using AI. It just my opinion. Even the best, even the best performers use autotune and all these others tools to to do one's own thing sounding the best. AND in a similar way, we we know that AI gives to people superpowers. Dr. Miller said: "I'm not the best." writer". What makes it interesting, so that's what he is thinker. He has own position. It his old position, he is even ready to criticize oneself a former colleague, and I assume that and friend, because it is contradicts him worldview.

And he says he uses AI to make your text is better. I do this. I don't understand. I'm not...this... I have no problem with... tool for typing. Me just not I like the idea that he wrote this completely, namely this assertion. Now we We don't know for sure. We Let's talk to Stan, when will he be on west. I'll tell you, I I prefer to such performers as Mark Knopfler, who he really writes the songs himself and plays guitar riffs. I prefer so that my art, thoughts or novels were created a human, not an AI.

It my choice, and that's it I expected this. audience. But this new... Now that you know, that Duck will continue use AI, because he himself said so said, you just ignore him words? No, but I will be less respect him. You do you mean that the essence of his words is more will not have for you importance, are you less Will you consider the content? No, you just said, that I will devalue content. I will to ask yourself what his true opinion. But what you say, comes from personally from him.

You will trust the greatest macro trader of the world less tomorrow than today, just because, what did he use tool for writing text? I... I...I'll think about it. that's really his opinion and how much work is he performed, and I believe disrespectful that he didn't do it myself. Here What I feel. I will doubt. It will make me to wonder if it is really his own opinion, because I believe that He works hard, but I I don't like it when people they are sloppy. That's all 800 words.

Write it yourself. Work with proofreader. Quite to spend normally research like this in a way. Listen, we let's stay with our own thoughts on this. CIA Director carried out unannounced visit to Moscow. John Ratcliffe carried out unexpected visit to Tuesday and the same day flew back. It first visit of the incumbent CIA Director with November 2021. Then Bill Burns rode to warn Putin not to invade to Ukraine. In three months he broke in. As Wall Street reports Journal, Ratcliffe was there to warn Russia should not be attacked NATO country.

By messages, new US intelligence indicate that Putin can plan test one's resolve NATO limited attack in the near future several years. Anything: from cyberattack to small invasion by land etc. Poly market says it is only 21% chance of ceasefire between Russia and Ukraine to end of the year. Friberg, You were very active. in a group chat with about this. What are you? Are you thinking about what's here? is happening? There are wild ones assumption. Wall Street Journal and others reported that it was transferred message: no invade NATO and do not use nuclear weapons against Ukraine.

So such wild ones appeared assumptions about What does this mean. It seemed that something is happening significant. But, how Sachs noted in our chat, it's just assumption. None no evidence that What did Putin do? More than that. The President seems to called the show, to essentially say, what is this story false, but contact was ordinary. AND then Russian the party reported about the same. So, the truth is, what we don't know about that there was a meeting. Um, but, you know, quite dramatic. I think, I think that this all assumptions.

I know what this is about. was. Oh, it's started. Wall Street Journal note, created by with the help of AI. Yes. He warned, he warned Putin is not use AI for writing your next Putin warns about your next child- AI. We flew there to to convey very harshly message: we we expect using tools without AI, because we— journalists at heart. Yes, I think so. maybe Jal, that trained journalists were wrong about this, and we resort to the wild assumptions and hyperbole, actually not knowing what it was about meeting.

Well, I mean that wild guesses—this what we do here. This is a podcast. Well, in this and that's the whole point show—speculate that, Damn it, here is happening. First, and American, and Russian side refuted these dramatic speculative statements about what Ratcliffe was supposedly there to warn the Russians not invade Europe. I actually believe this. refutation, because for there would be no Russians sense to invade NATO countries. This is the one the most common type of bloating the threat that we we have been hearing from for years neocons about the alleged imperial ambitions Russia to conquer all Europe.

For this there is simply no reason. And, frankly, The war is coming for the Russians. not bad. She, of course, it continues much longer than anyone would expect. AND Russia paid high price in the form of losses. I don't know whether numbers are like you said Friberg, but I think if you look on most sources, what is currently being covered war, they say that Russians do slow, but steady progress. AND exactly Ukraine suffers the most. First of all, Russians have an advantage in in the air.

In Ukrainians practically exhausted anti-aircraft defense. Therefore Kyiv regularly undergoes bombings. Russians can destroy almost everything that they will wish. So Ukrainian infrastructure undergoes enormous losses, and this on the eve expected enough cold winter, and Ukrainian power grids are being disabled. Moreover, Odesa and Black Sea ports Ukraine were bombed and actually closed. So their economy now pays incredibly high price. They are not can export grain. So, their economy in fact is destroyed, and in them running out of money. Zelensky is running and says that he need 35 more billions of dollars, otherwise they have the money will run out.

The US is not funding this, but he is trying receive these funds from Europeans. And theirs too losses, although we do not we know numbers, too huge. And now stories appeared about what they are potentially mobilize women because they are catastrophically not There are enough soldiers. Keep in mind that population of Russia in four or five times more than in Ukraine. So they can to hold significantly larger armies. Therefore, situation for Ukraine doesn't look very good good. And, frankly speaking, Zelensky it was worth it to listen when President Trump said to him: "You have There are no trump cards in this one.

war. You should to make a deal. Zelenskyy didn't want to to do this. He is not wanted to make a deal. He still doesn't want to. bargain. He should have made a deal. And, to be honest, I don't I think that Ukraine will emerge victorious in this matter. Chim, do you want something? add, shall we finish? Yes. Thanks Meta for the solution my biggest problems— installation restrictions on Instagram. Yes. Good. We even didn't get to this point topics. This is an important topic. I'm tired of playing " "Catch a mole" from with my children, I don't I can handle this.

They have fake ones accounts, secret accounts, this and that, yes what, please thanks, it was worth it do it earlier, but I'm glad you did done. Yes, I will give it to them. due for that Yes, they were forced to make the right one thing, but TikTok and YouTube, I I hope TikTok and YouTube will do the same. And that's what life will do. parents much simpler. This will allow us really educate children, not just to follow using devices. I would like to see this. on Android.

AND it's exhausting to be child with a device. This is exhausting, and it's terrible for young girls— sit in these Instagrams. We we we held on until 16 years old, but at 17 we lost the battle. Where you were the one holding it defense? Regarding children and social media? No, my children don't allowed until they will turn 16. But not That's the point. Simply, Look, we are already talked about it. I was very consistent in on this issue. It could happen much earlier. I I'm glad it happened.

I applaud Meta for... they did. Hopefully TikTok and YouTube will do the same. Such restrictions appear in Australia, Canada and Great Britain. They probably will appear in others as well places. It physiologically and psychologically true useful for children— have such restrictions. Therefore, Yes, these apps for parental control is not that great and wonderful. They They work well. But not great. So how Father, I am always in defense, and this It's frustrating, I would rather wasted time on others dealing with children than to shout at them to they postponed phones.

Yes, tools difficult to implement. Hey, Friberg, what's up? with the summit? as we We are holding a summit. Well, glad. to inform that to All-In Summit all that's left fortnight. This year we have incredible sponsors who help to implement big ideas into life. Iron performs again main sponsor All-In Summit. Big thanks to our friends from Iron. Necessarily visit Iron House at summit. Always great We spend time. Necessarily visit Iron House at summit. This is a lounge area, where can relax, to rest and to communicate between sessions.

And the new thing this year, novelty, Do you remember this from Napa, boutique gifts, presented by merch.com. Each participant fills the bag All-In merch, plus special offers from sponsors. Come in, you go out, you get a full bag of goods— you will be very satisfied. Oracle returned. They they will bring a simulator F1 races to the summit. If you ever wanted to know how it is to enter a turn at a speed of 180 miles for an hour, a little similar to the debate with JCAL here on the All-In podcast, without getting into hospitals.

This is your chance. And EY takes over business, bringing your modules for meetings, it's quiet space where you can go out, to call, to conclude an agreement or to hold a meeting that, by the way, often happens on summit. People conclude agreements. This is just the tip. iceberg. We also have other important news from our wonderful sponsors. This is just part, the rest we we are still holding on secrets. On you waiting for more interesting. We are in foreboding. See you all. you at the All-In Summit.

This will be wonderful. party. Hey, Fred, what's up? is happening with this a cancer vaccine and shares that tripled for the last two weeks? We didn't have time discuss this last week, but I I want to hear about it, and sure that the audience too. Yes. You probably talk about stocks Modern, right? Yes. It seems that theirs market capitalization increased from 20 to 60 billion thanks to positive the results of this thing called a cancer vaccine. And I I think it's worth it. explain how it works this system of therapy cancer.

This dates back to the 90s. years when it arose the idea of ​​creating so called neoantigens. These are small peptides or proteins that mimic a peptide or protein unique to cancer in your organism. So this one antigen that is immune the system perceives as invader, forces her start producing cells that destroy this protein. Entering this protein into the body, you activate the immune system system so that it found and destroyed this protein in others parts of your body. So, the goal is to in order to determine protein unique to specific cancer, to make it, to introduce into the body, and then the immune system will destroy the cancer itself.

This immunotherapy, this system of use antigen, designed for fighting cancer in your body, exists, let's say, since the late 90s. There was a big revolution in DNA sequencing, when we realized that can you take samples cancer, put them in DNA sequencer and determine gene sequence and proteins, unique for this cancer, to create personalized neoantigen, because all cancerous tumors mutate, they are different, and two identical crayfish does not exist. How are they find this "key" personally for me, but not for you? How are you? you get this information?

They take a sample. your tumor or cancer cells. Therefore, if we take, say, melanoma, then the reason, why is she one of the biggest goals for immunotherapy, is that melanomas have very unique mutations DNA through influence ultraviolet radiation on skin. Ago specific cancer that you get when melanoma, it will be very unique compared to with the rest of your body. This will be very unique mutation. So, they take a little melanoma, extract DNA from it, and when do they DNA sequencing, they reproduce this DNA millions or billions of times.

You you put all this DNA into the sequencer, it is crushed and is read. And then data is coming in sequencing, which they say, "This is what it does "This cancer is unique." This is this small one. DNA sequence. AND this DNA sequence causes cancer to produce unique protein, which, in essence, is key, certificate or a copy of this cancer. Then they take this DNA and they make protein from it . This protein is injected into your body, and now your immune system goes crazy and starts attack this protein, because it reacts: "Oh, antigen, antigen, we need it destroy".

Now that these cells activated, these immune cells are configured on this protein. Eventually, they will go and destroy this cancer. So, that was it. successfully. There were hundreds tests of this therapy based on neoantigens for many different types of cancer, and all of these trial tried to find out, which one is correct delivery method neoantigen. Or we create protein and we enter it into your body, or, as in the case of with Moderna, we create mRNA , we enter it into organism, and mRNA makes your own cells produce this protein, push him in blood, where your immune cells are activated.

So, instead of ex vivo protein creation, that is, outside your body, you do it in vivo, and now your body it produces this white. So the idea mRNA concepts is that you can you make one injection, this injection enters the body, and now your body produces the same protein, how we did it with the COVID vaccine. Yes. Your cells will produce a certain protein segment, which activates your the immune system, and she will go kill this one cancer. There are other questions. about which there are others adjuvants for which types of cancer can use this system?

And also combination therapies. Are there others? antitumor drugs with which can use this in combinations? And everyone test that are currently ongoing, specifically related this mechanism delivery method, which used, cancer types, dosage and how all this right to adjust. What interesting about this, is this that it is rather a method, than medicine in traditional understanding. This is a process It's a process, right? Therefore, this process adapts to every person. Therefore there is certain expenses for execution for each human. That's right. But cost, because I I am doing DNA sequencing in to your company for days.

Getting that The sample is very cheap. Obtaining DNA data very cheap. Actually, the execution many preparatory The work is very cheap. Theoretically mRNA production or there should be protein too quite cheap, but there is human trials and human safety. Therefore, there are many questions about how make sure that everything safe, and what correct combinations so that people don't have hyperactive immune system or side effects. So that's all setting this up. But I think that most of all me upsets about everything this is what everyone praise it as some kind of a unique breakthrough, special and powerful, but that's not true, because many people go to places like Montana, where are they?

print peptides for their specific cancer sequences, they create own neoantigens, introduce them into your organism and are cured of cancer. You can to pay someone 50,000 dollars so that they did it for you already today. THERE ARE many clinics that do this will do, and it's not some special medicines, FDA approved. To yours thoughts, Jake, this process. Process is to take a sequence The DNA of my cancer, create protein, to return it into my body, and my the immune system goes and destroys cancer in my body.

Great idea. Then why does Moderna say that is going to take over Is that $500,000? And this what I think, It saddens me. all this. All everyone is praising it very much I'm glad about that, but I actually very sad, disappointed and saddened by this. I am not I think that this method, which to a large extent was developed for several decades of iterations, research and developments that mainly funded by the NIH and other state funds that led to to this incredible concepts neoantigenic cancer immunotherapy, which can eradicate or theoretically eliminate a lot various types of cancer, must be now patented, FDA approved and to cost half a million dollars for treatment.

Yes. So, what do I not I like this one history, that's what private pharmaceutical the company saw how its market capitalization increased from 20 to 60 billion, because they passed through this regulatory process, to create what, in my opinion, is a form regulatory delight, claiming that everything this is for safety, tests, patents and everything another. Whereas, according to In essence, there is a basic technology that was financed by many state research dollars, and I think what it should be more open, more ubiquitous, and every hospital must be trained because how to conduct this process for treatment of patients with cancer; this should happen much more common phenomenon.

Do you think people I have two. question. We can end it here, but do you think that they will be able to succeed protect this patent in the United States, and also in everything world, because the world obviously global, and people are really now traveling for treatment. Medical Tourism exists. So if it's worth it here half a million dollars, is it possible to do this for 50 thousand in South Korea, Japan or others places? Yes. And in- the second thing that happened to mRNA and vaccine against COVID, and for many concerns that were in people, we still don't received an update, but this podcast, obviously started right during everything this.

And that was the first large-scale mRNA unfolding- technology if I don't I'm wrong. What are you think about this reason? Yes. Over the years, Moderna worked on this mRNA method cancer treatment. They they call it vaccine, but actually it is immunotherapy, because vaccine means that you prevent disease. While in this case you actually get rid of that what you already have. Ago I don't like it. the word "vaccine". But Moderna has been working on these years and still not was able to get away with this market, and then COVID appeared, they very quickly reoriented and released this vaccine against COVID.

Moderna has a bunch of patents for methods and the tools they used for creating mRNA to make sure that she does not integrate into the human genome and not uses traditional system replication. So you don't you can actually replicate this mRNA, and this means that she will not start to spread in your body because your cells usually can start creating RNA copies, if it is gets inside. So, there was a lot Moderna's developments, which made theirs specific mRNA- technology patented, safer and so further. Therefore, I believe that they have a strong patent protection for its specific tool kit for mRNA.

But, as I described earlier if I take the sequence. your cancer cells, I look at her DNA and I see what exactly makes it unique. There is such a small sequence, and now I'll take this one DNA sequence and can I insert it into bacterial cell Escherichia coli in small tank, which is called bioreactor, and this bacterial cell will start producing this protein, because I took DNA and placed it in intestinal cell sticks. So, now intestinal cell rods encode this DNA into RNA and then into proteins, because it is like this sequence, right?

Yes? DNA turns into RNA, and that—to proteins. So, in the cell Escherichia coli is billions of copies of these cells. They create billions copies of this protein. You secrete protein from bioreactor and you can inject it into organism, and when you inject this protein into body, your immune system the system is activated. My opinion is that mRNA is not mandatory necessary, although it really effective way to do it. It extremely cheap, because you don't need to produce all the protein out organism. You can literally simple print mRNA, to give one injection—and ready.

But this through-and-through the system has a very low real cost. And this is one of those things about which we must put fundamental question: is it worth it to give away a monopoly for the treatment of cancer of the of the company that passed FDA process and filed a bunch of patents for certain methods? But general theory, general thesis, general therapeutic process well described, good studied, generally understood and must be available everywhere. And I I think that during the next few years we will see how this system cancer immunotherapy will begin to spread, and you will see how it is is happening at very cheap abroad.

This is very effective, and in We have a mutual friend, who is trying organize a business, to do it very cheap. In Montana, where you have "right to try" and where they can to make these

Go deeper

• All All-In Podcast episode guides