All-In PodcastAnthropic IPO at Risk, Meta’s Muse Pop, Token Prices Fall, Open Source Gains Share, Alignment Fails
EVERY SPOKEN WORD
90 min read · 17,805 words- 0:00 – 3:54
Bestie Intros: Favorite All-In Summit moments
- JCJason Calacanis
All right, everybody. Welcome back to episode 290 of the number one podcast in the world, it's the All-In Podcast. With me again, it's the core four, David Friedberg is here, Chamath Palihapitiya, and David Sacks calling in. I am your host, Jason Calacanis, and we have a full docket today. I'd be remiss, though, if I didn't ask everybody for their fondest memories of the fifth annual All-In Summit, produced by [laughs] Mr. David Friedberg. David, do you have any, um, thoughts? Uh, h- how do you rate your performance as executive producer of the All-In Summit?
- DFDavid Friedberg
You know what we didn't do, what we forgot to do? At the end of day two, we didn't give a big shout-out to our production team, to Kimber, to Lisa-
- JCJason Calacanis
I did. I did
- DFDavid Friedberg
... to Nick. I think we, uh-
- JCJason Calacanis
No, Jason did that.
- DFDavid Friedberg
Oh, okay. I thought we should have brought them out, but I just wanna give a big shout-out to-
- JCJason Calacanis
That's always the move, bring them out for, uh, the standing O
- DFDavid Friedberg
... the team backstage absolutely crushed it as always this year, and then I had a great time actually. You know my favorite part is going to Universal Studios. You just get to have a good time. It's great. Just great chilling.
- JCJason Calacanis
I, I am reliving my childhood, Chamath, with no trauma.
- DFDavid Friedberg
By the way, shout-out to our friends at De'Longhi. I got one of those coffee machines that we had at the, uh, event. I put it in my kitchen.
- JCJason Calacanis
You know what I got? I d- I don't know if you guys saw, but there was a water machine in the VIP area-
- DFDavid Friedberg
Wise well
- JCJason Calacanis
... that is a zero plastic water machine.
- DFDavid Friedberg
Yes.
- JCJason Calacanis
It goes through super filtration, no microplastics in the water. I have one now in our office. Thank you very much to them. I appreciate it.
- DFDavid Friedberg
Yeah.
- JCJason Calacanis
Amazing company. Sacks, do you have a, uh, favorite moment from the show?
- SPSpeaker
Well, I think the highlight of the show, the thing that we'll remember for a long time, was the president calling in during Jensen's talk, and I think Jensen's overall, the content and, and demeanor really helped calm down this national panic over AI. It felt like we were the center of this national conversation, and I think it happened very organically. You remember there's basically been this PSYOP for the last few weeks leading into the summit to convince everyone that humanity was gonna go extinct unless we did exactly what the Democrats want [laughs] and stop all AI development. And then, you know, we had first Satya came out and kinda calmed everyone down, and then Jensen did the same thing, and the president called in and did what he did best, which is seize total control of the narrative that I'm sure these Demor groups had been carefully planning for weeks and kinda thwarted that whole operation.
- JCJason Calacanis
Nerds, enough, get back to work.
- SPSpeaker
[laughs]
- JCJason Calacanis
You have a super intelligent Trump-telligence is now the term.
- SPSpeaker
[laughs]
- JCJason Calacanis
We're gonna call it Trump-telligence going forward, but I think you make a good point. A lot of people wanted to know, Friedberg, if this was planned. Was the President Trump call-in planned, Friedberg? T- c- take us behind the scenes.
- DFDavid Friedberg
Basically, Jensen was t- connecting with him right backstage and said, "Hey, call me back."
- JCJason Calacanis
Yeah.
- DFDavid Friedberg
And then he called him while he was on stage. Uh, there was some texting going on.
- JCJason Calacanis
Mm.
- SPSpeaker
No, it was completely spontaneous.
- JCJason Calacanis
It was spontaneous, yes. Thank you.
- DFDavid Friedberg
It was not like a planned thing where Jensen was like, "Hey, let's do a whole thing." And so that was pretty amazing to have this conversation that Jensen was waiting for from the president to happen live. And clearly this was the topic du jour in DC as well as in the news and the media, and it was amazing to see it all happen on stage with us.
- 3:54 – 20:03
Reacting to AI chaos: Liability, competition, and rebranding frontier "labs"
- JCJason Calacanis
perfect line. But Chamath, in all seriousness, the nerds were losing their mind last week, and, uh, the adults came. Daddy came. Elon came. Jensen came. All the adults in the room came, and they just said, "Listen, if you have software that's defective and not ready to put in the world, maybe don't put it in the world, and make it a little bit, you know, uh, as you like to say, less brittle." So your thoughts on where we are in this national discussion, which went from slow down to maybe just release non-brittle resilient software.
- SPSpeaker
Well, I think that there was this classification that happened last week between up-and-coming organizations that want to call themselves labs but are in fact companies that have P&Ls and shareholders and have to abide by the same rules that everybody else does, and then the sophisticated companies that have lived in scrutiny for many decades and have learned the hard way the process of making things robust and the cost of making things not robust. So what do I mean? It, on the one side, you saw, in many ways, it's like curious times creates curious bedfellows, right? You have the cohort is Satya, Jensen, Sacks, Zuck, Elon, the President, Lina Khan. We have product liability language.
- JCJason Calacanis
Yeah, [laughs] strange bedfellows indeed.
- SPSpeaker
We have this product liability perspective in the United States. It's meant to prevent companies from doing things that they're not supposed to do, especially if they know that those things could be harmful. And on the other side, it's folks making the thing that were being, I think, a little naive and basically saying, "We're not a company, we're a lab. Protect us." And I think the mature response is you should have internal controls and you should robustly test things. And what Zuck said earlier and what Elon has said in the past is, "We're just under a lot of scrutiny, and so when things aren't perfect, we just slow things down." Elon has done that with FSD. You said this, Jason. He could have been at FSD
- CPChamath Palihapitiya
Three or four years ago. But he has publicly stated we can't afford it because every accident that Tesla incurs is magnified 1000x than what any other company would incur, and so we just have to go slower. Zuck last week said, "We were about to roll out Muse," which looks like one of the most successful product launches in recent memory, has huge implications actually, which we should talk about, and he had to slow roll it. He and Alexandra, who I was talking to last night, had to slow roll it for a couple of months to get it perfect. So it's all within your control. It's just what do you choose to do? And, uh, I think that was the most interesting observation. We have to stop calling these companies labs. They're companies, and we have to be judged on the same standards, including product liability.
- JCJason Calacanis
Yeah, and there was a lot of coverage of this moment, Sacks. Interestingly, some of the fake news media [laughs] were, like, on a popular podcast, and then they play the clip, but they're just afraid to say the [laughs] name of the podcast. Pretty hilarious.
- DSDavid Sacks
I think Chamath is getting us into the first topic here already, which is whether we're gonna have individual responsibility as our North Star or some sort of collective action. And the thing that always works the best is individual responsibility, whether it's at the level of individuals or companies. The relevant decision to make in every case is, should we ship this product, not, can we get the whole world to agree? And the problem with talking about these collectivized approaches is that they defray accountability away from the actual decision-maker. And I, I think people sort of feel this. You know, like this week you had Dario and Sam go to the United Nations and say we need some sort of global AI governance, and it kind of reminds me of, like, the billionaires who fly to Davos in their private jets and then rail against climate change. You know, there's something fundamentally sort of unrealistic and kind of out of touch about it.
- JCJason Calacanis
Hypocritical.
- DSDavid Sacks
Hypocritical. And it's like-
- JCJason Calacanis
Delusional
- DSDavid Sacks
... guys, look, what we need is for you guys to make smart decisions about what to release. If a product is unsafe, which is to say it's unreliable or behaves unpredictably, then don't release it, and stop looking to some sort of global governance as the answer. Because if you guys don't get this decision right, it's not gonna matter what the United Nations does. And by the time the United Nations agrees on anything, we're gonna be much further along in this whole, whole journey. So there's something just very weird about the way that the principal actors in this keep talking-
- JCJason Calacanis
Hmm
- DSDavid Sacks
... because it really feels like they're trying to push off the responsibility. And-
- JCJason Calacanis
Well, there's a theory going around here, and I think it is worth us discussing here, that the frontier labs are covertly-
- CPChamath Palihapitiya
They're companies. They're companies
- JCJason Calacanis
... They're companies, yes. Let me just finish the-
- DSDavid Sacks
No, no, no, but, but Chamath made a really good point that I want to make sure we just don't miss.
- JCJason Calacanis
Okay.
- DSDavid Sacks
They're corporations.
- JCJason Calacanis
Yes.
- DSDavid Sacks
And when you call them labs, they sound like non-profit entities, and it's not true.
- JCJason Calacanis
Okay, so the corporations making these products are looking for President Trump to specifically give them what the internet companies had with Section 230, which is shield them from liability, and that there is a consideration inside the Trump administration, the 47th Administration, that maybe there could be some leverage put on these companies that they would give 10% of their equity to the, uh, sovereign wealth fund of the United States, et cetera, and then that would result in them being given some pass when it comes to their liability. That's an insane thing to do. It's like giving some plane company that gives 10% of their equity to the US government the ability to not be responsible for their planes falling out of the sky. But what is your thought on that conspiracy theory? I'm sure you've heard it, Sacks, that they're looking for a shield.
- DSDavid Sacks
Well, hold on. You're, you're, you're making it sound like this is the idea of the Trump administration. That's not true. Every-
- JCJason Calacanis
No, no, not the Trump administration. This is what-
- DSDavid Sacks
Okay
- JCJason Calacanis
... the frontier labs/corporations are pitching-
- DSDavid Sacks
Yes
- JCJason Calacanis
... to them, and that there might be, the rumor is there might be some people who are open to that in the administration. Again-
- DSDavid Sacks
Okay. Look-
- JCJason Calacanis
... this is all a back channel
- 20:03 – 29:22
Non-frontier performance and cost: What this means for frontier companies
- JCJason Calacanis
at Oholo.
- DFDavid Friedberg
Let me kick off the first topic. What I wanted to do was just take a step back and look at the insane flurry of model releases happening.
- JCJason Calacanis
Yes.
- DFDavid Friedberg
I think the pace of model releases happening on both the open source side and the premium side, [laughs] the Anthropics, Groks, OpenAI's of the world, is unfathomable right now. The list is insane, in just, like, the last 10 days. September 9th, DeepSeek 4.1 Flash gets released, $1.20 for a million tokens output if they're hosting it. Key value cache from V1 of DeepSeek. Remember what a big deal DeepSeek was when it came out? The key value cache then was, like, 390,000 bytes. Now it's 890. Insane drop. Insane efficiency improvements. On September 20th, Qwen, which is the Alibaba model, comes out, version, you know, 2.1. It's an open weights model that outperforms NanoBanana 2, which is Google's big breakthrough insane image generation model. This thing is free, open source. You can put it on your desktop computer and run it at home to generate images for free, and it's made by Alibaba. Xiaomi put out Mimo on September 22nd, two days ago. Mimo Pro performs on par with Claude's Opus 5, which is, like, this insane model that everyone thought was gonna change the world and break the internet, and GPT-5 six Sol across most benchmarks. It's a 309 billion parameter model. It is entirely open weight. You can download this model and put it on your server and run your own instance without paying anyone any money except for the cost of running the server. This model called, uh, Bonsai 2 came out on September 17th by Prism ML. This is actually a fork of Qwen, which is another one of the Alibaba models. It's a 27 billion parameter model, 98% of the performance of the big Qwen model. It's 5.9 gigs in size, and you can run it on a local computer. You can have an Nvidia chip or have a good Mac Studio computer, and you can run this insane model on your desktop computer for free. And then the big crazy models, the big closed source models, Anthropic releases Opus 55 on September 22nd. OpenAI Astro comes out with Sol and Luna on September 23rd. Grok 47 comes out on September 21. Meta Muse comes out on September 22nd. Any one of these stories would have broken the internet a year ago, and they all happened in the last 10 days. You now have an entire catalog of open weight models to generate visual information, to do VLA, where you can actually instruct it and it can control a robot. You have models far beyond the most advanced LLM from a year ago that are open weight that you can install and run on your desktop computer. We are past the point of Bernie Sanders, "Put the superintelligence away. It can never be allowed in the real world." This stuff is out. The open source open weight models are free. They are on the internet. You can go and download them, and you can run them on your computer at home. I don't know in what world we think we're gonna have a police state that's gonna come bang down your door, take your desktop computer from you, and test it to see if you're running an LLM on your local computer. The efficiency wave is here, and the pacing at which these models are coming out, now all of this stuff happened in the last 10 days, it's just mind-blowing. We're now in this kind of frontier space where things are accelerating, the performance improvements, the cost decline. AI is becoming so performative and so affordable, it is going to be ubiquitous, and we are all going to benefit from it. This is not just about the Anthropic and OpenAIs with the closed world making billions of dollars selling enterprises hosted models. There is so much proliferation right now. So I just thought it was worth taking a step back to recognize where we are, 'cause in the context of we have to stop the superintelligence, we have to shut down the data centers, you don't even need data centers to do 90% of what you can do with AI. You can download it on your computer for free today to do what was the most advanced technology in human history less than a year ago. It's all here today. And I thought it was just worth taking a beat to step back and make sure we all recognize that as we get into this debate about these, quote, "labs" or corporations. I don't know if it even matters anymore. I think open source is here, and I think that the pace of improvement is such that this is only gonna drive productivity-
- JCJason Calacanis
Yeah
- DFDavid Friedberg
... and maybe we are, maybe we are seeing the AI wave underway right now, the productivity wave, the productivity boom.
- JCJason Calacanis
Yeah, and then to consumers, they don't care what model they're using. We've spent the last three years geeking out about this, and then GrokBot comes out, and then Muse comes out from Meta, and I've been using both of these products, and it's f- the first time that a normie, a normal person in the world, it's not meant to be derogatory in any way, can get value from AI. If you look at how we all looked at this, running corporations, investing in them, we took a different approach to like, okay, maybe this is gonna retire, you know, some positions. Maybe this is, um, gonna add some positions here or there. But now you can, like, with Muse, connect to everything instantly and actually get real work done in the world. If you're a normal person who wants to get the benefit from AI, we will be turning a corner, Friedberg, and I think you laid it out perfectly. We're gonna turn a corner, I think, at the end of this year, where moms, our uncles, people are gonna start using Muse and Bot, GrokBot, which are growing very quickly, to just solve real-world problems for that, and they're gonna go, "Wait a second. I could never afford a chief of a staff. I could never afford a personal assistant, an executive assistant. Now I get a free executive assistant?" It's going to blow people's minds, Chamath, in the next couple of months when they start using these products, and the bottom half of America in this K-shaped recovery experiences some benefits. Your thoughts on what's happening here with these models. Have you started playing with it? 'Cause you were, uh, you know, originally a technologist. Have you loaded these models on your computer? 'Cause I'm playing with it on my MacBook Pro, and it's awesome.
- CPChamath Palihapitiya
Let me maybe e- give you the answer by explaining something. So I think what we're seeing is that these models are converging and clustering, which is to say that they're now within margin of error. And so the thing is, there are so many of these models that are roughly the same at different price points. So where is the edge? There's still edge, and the edge is in the harness that you use to wrap the model. So if you think about it as like you have a model that's like a brain, but you need to embody that brain with arms and legs and hands and eyes to do work. That's how you agentify a model, okay? And you make it agentic. That's what that phrase means. And there are all kinds of different harnesses that the industry has invented to make these models more agentic, and there's still a lot of edge there. So at 8090, we have a labs function, and we've been ripping and teasing apart all the different models with all the different harnesses just to see how they perform, and they are wildly variant in terms of cost and quality. So just to kind of build on Friedberg's point, you have the models that are converging. The harnesses are wildly different, and so their performance is all over the map. And so now companies and people have to sort of navigate, what are you willing to pay for? I sent this tweet to Nick, if you can just put it up here, and I don't know if this is instructive or accurate, but what this shows, and this is the case for Ant, but it's probably the case for a lot of the closed models, is you're starting to see this massive revenue concentration, which is a few folks consume all of the highest and most expensive tokens. And I think if you take what Friedberg just showed to its logical conclusion, those folks at the top of this table are going to be under a lot of internal pressure, whether it's from their finance teams or whether it's from their CFO, or it's going to be from shareholders to explain why, as this clustering happens, they are not moving down to cheaper and cheaper models. Now, you may choose to go down to cheaper and cheaper Ant models, so then Ant keeps the revenue, but there's also a risk that you may do what Friedberg just said and just go to an open source model and just host it yourself. So we're in, like, a very complicated part of the cycle, which is the first version of the AI trade was relatively simple, which is you're just selling tokens, you're wrapping the tokens, you're passing it through, and there was enough disparity that there was value. I think that that's going away. It'll force OpenAI and Ant to go up the stack. There is no choice. You can't sit there and serve a token that has diminishing value because then all the subsequent value is getting absorbed and made by people wrapping your intelligence token. You will have to do that work. And so Jason, that's why they will have to go and do cyber. They'll have to go and do law. They'll have to go and do customer support. All these things that we weren't sure whether they were gonna compete, these trends may force them to do that.
- JCJason Calacanis
Yeah.
- CPChamath Palihapitiya
And I think then that sets up a very interesting setup for their IPO, in my opinion.
- JCJason Calacanis
Yeah, and just to back all that up, Saxx,
- 29:22 – 53:11
Anthropic and OpenAI postpone IPOs: liquidity risk and open source pressure
- JCJason Calacanis
the frontier [chuckles] lab corporations, uh, Anthropic and OpenAI, they both released models this week at 50% less token prices, and their IPOs are looking like they're both gonna get delayed. Anthropic just targeting a $2 trillion valuation, OpenAI at $1.2 trillion. Sam already said OpenAI plans on doing 2027 over because of safety concerns, and Anthropic, which was supposed to file, uh, and go out in October, the report from The Wall Street Journal is that the IPO will happen in November or may be delayed even more. Here's your Polymarket. Will Anthropic go public in 2026? With our amazing partner, Polymarket. Shout out to my guy Shane. Peaked at 96% earlier this month that Anthropic would go public this year, but now it's fallen to 76%. Is this like a cell phone here, Saxx? The [chuckles] the leading model companies need to go public to get capital to do their build-out, and now they're delaying their own IPOs. Please explain how any of this makes- Sense, and if you were a major investor in one of these companies or on the board of them, what would you be doing or instructing these founders to do to get these companies to go public and get that capital they need to continue the AI build-out?
- DSDavid Sacks
Oh, well, I, I think that we know from our friends that the investors in Anthropic are pulling out their hair about what's happening right now. Forget about the fake whistleblowers who were there for, like, a few weeks and then left. You have current leadership in the company saying that there's a greater than 10% chance of causing human extinction, and their own product is unsolved in this area. So think about the risk factors that's creating. How do you IPO with your own management of the company saying that? Then you've got, I think, other evidence of, I don't know what to call it except corporate schizophrenia. You've got Dario putting out this essay saying, "We need to pace the frontier," days before they launch Claude 5.5 and set a new frontier. I mean, he must know when he's writing this essay that they're about to release a product that will extend the frontier, so at a minimum it's hypocritical, but maybe it's even worse than that. Then you have them publishing an essay about the bio risk associated with AI and how that needs to be stopped, and then meanwhile, they announce a new bio lab in San Francisco.
- JCJason Calacanis
Wait. [laughs]
- DSDavid Sacks
A new wet lab.
- JCJason Calacanis
Wait, they started a lab after telling us that there's a 10% chance of us all dying, Sax. Let's just pause here for a second. Is this a dereliction of duty as the chief executive officer of Anthropic to sabotage your own IPO? You're supposed to be acting on behalf of your CEO, of your shareholders. This to me seems like if we were the board, if the four of us were somehow magically the board of Anthropic, we would replace the CEO. We would put the CEO in the lab, and we'd find a CEO who knows how to run the company and not sabotage it. He's literally sabo-
- DSDavid Sacks
Well, I'll tell you-
- JCJason Calacanis
... taging his own company.
- DSDavid Sacks
I think there's an element of that, and I think investors are pretty nervous about giving the founders, or, or Dario specifically, super voting shares. There was a story in The Information about that, that that's an active debate right now about whether to give him super voting shares.
- JCJason Calacanis
Explain what those are, please, and, and why in this context-
- DSDavid Sacks
Well-
- JCJason Calacanis
... this is critically important. Our audience may or may not know what those are.
- DSDavid Sacks
Supposedly the founders of Anthropic only have about 2% ownership each of the company, which in the grand scheme of things is a relatively low number relative to what other founders typically have at time of IPO. It'll still be an enormous amount of money, you know, tens of billions of dollars, but that is the case. Super voting shares basically say it separates the economic shares from the voting shares in the company and gives certain people the right to control the company even if they don't have the economic ownership that would usually be necessary. It's a pretty common thing. I mean, Larry and Sergey I think maybe were the first prominent founders to get-
- JCJason Calacanis
Zuck got them, yeah
- DSDavid Sacks
... super voting shares in, in Google, and Zuck... It can be a stabilizing thing for the company because it means that as long as the founders are there overseeing things, they're not gonna be subject to, you know, like takeover attempts, things like that. But on the other hand, it requires a tremendous degree of trust in those founders because they effectively can't be removed no matter what they do. And you just have to wonder if in this case that's earned and warranted. I have to say that even for someone like me, who isn't necessarily advocating for a lot more regulation, I kinda get nervous about the idea of Anthropic opening a wet lab in San Francisco. I mean, I was think-
- JCJason Calacanis
Great timing.
- DSDavid Sacks
I was thinking about moving back, and I don't know if I wanna be... I mean, like, can you just put it a little further away? I think we've all seen this movie before.
- JCJason Calacanis
Yeah, offshore maybe. I don't know.
- DSDavid Sacks
[laughs] Yeah. No.
- JCJason Calacanis
Put it on an oil rig or something. [laughs]
- CPChamath Palihapitiya
What I would say is that if I were on the board, I would keep Dario as a CEO, and the reason is that he's clearly demonstrated that he can build a unique culture and position against the otherwise most formidable corporation competitor, which is OpenAI, and he's done that, and he's come from behind, and I don't think you should undervalue how hard that is and what that takes. So he's done that, and he's built, he's built essentially the greatest business ramp in all time. I think the thing that I would probably do if I was on the board is fix what Sax just said, which is there's now a bunch of incremental liquidity risk to the company, and I think if you take the 20-year view of Anthropic, that corporation needs hundreds of billions of dollars to fulfill their ambitions, and I would solve for getting that as risklessly as possible, and I think there's only two things that you can do. The first is you have to throw everything and the kitchen sink into disclosures. The problem is that that will make what is an otherwise dense and impenetrable S1 even more dense and impenetrable, okay? So you can kind of, like, superficially cover your regulatory risk, but the way that it actually gets translated in the market is you're just going to have to water down the expectations of the IPO sellers and create a much larger margin of safety for the IPO buyer, and that's effectively how you find a market clearing price for this, Jason, which is to essentially say... And bear with me. I'm just giving out rough numbers, right? Even if you think that they're on $100 billion run rate, what would otherwise in a very clean sheet IPO maybe would've been a $2 trillion market cap, now, you know, you're at one or less. And the reason is the hedge funds and the pension systems and the long-only mutual fund complexes who have a fiduciary responsibility to their sources of capital will justify all the risks by essentially asking for a lower and lower price, and I think that's what you're gonna see. I think you're gonna see a much more moderate and reasonable IPO motion here. And it's probably gonna clear at a much lower price than anybody thinks. And I do think that what Sacks says is the reason why, but I think it's good for Ant because it clears all this noise, it forces them to just be a corporation and then run the business, and I think that they'll do great. Look, I'll be honest with you, like, I've tried to recruit against Ant. I can't. It's impossible. The comp that they give is incredible. The business ramp is so compelling. It's like when you're in a head-
- JCJason Calacanis
Culture is great. Yeah
- CPChamath Palihapitiya
... when you're in a head-to-head with these guys, I have not won a single bake-off for folks. I wish I could say that I did, but I have not, and it's hard for a startup. So they are a magnet for talent. OpenAI is a magnet for talent, but these other risks are gonna weigh on the IPO, and the only way to flush them is to just bring the price way down.
- JCJason Calacanis
Yeah, and Friedberg, when we look at the risk factors in an S1, those are perfunctory. Those are disclaimers typically. You get the SpaceX one, "Oh, what if Starship doesn't work? What if there's competition?" It's kind of checkboxes. In Anthropic's, I think we're gonna be looking at something totally different, which is, "Hey, we [chuckles] and our employees are quitting, saying, 'This is gonna kill humanity.' We have massive regulation. We're asking for regulation. And by the way, we're doing biology work, and we're launching a lab." So do you agree with Chamath here that the S1 is gonna have [laughs] an actual risk factors and disclaimers that are material against the IPO, Friedberg?
- DFDavid Friedberg
I think this customer concentration point ties to the earlier comments I made about the proliferation of incredibly performant open weight and open source models, and the reason is Anthropic puts out the best models for life sciences today. I run a life sciences R&D organization. We use their models. They're incredible. They're better than other models, and so we wanna use them. When it comes to writing code, to writing software, to do internal workflows and run operations, we can use other products to write that code for us, and we do. And I think that that's likely the way the market is gonna develop, and I think we're starting to see that proof come out in the market, and you're seeing it in this customer concentration point. Extremely high value, technical, engineering, math, life sciences, those really hard technical problems, you wanna have these top-tier, top-shelf models, and you're gonna pay a huge premium for them, and you're totally gonna make money doing it because the value is extraordinary. And this has always been my case about AI. AI is not so much about the value of replacing old stuff. 99% of the value of AI is about enabling new stuff that's never been possible in human history, and I think that's where Anthropic and OpenAI are gonna make their money. Separately, I think OpenAI is gonna make a lot of money in consumer and Anthropic maybe to some extent as well. But in the enterprise setting, I think the money comes from creating new frontiers, discovering new enzymes, which Anthropic announced this week, that can cure human disease, and developing new technology, new engineering systems, making things that have never been possible to be made before. And I think that's the premium value, and it's the premium price you'll pay for those closed models and that premium model. But that's a different business case than taking all of enterprise. Separately, I think most of the enterprise applications are gonna be run using these open weight models, and code is gonna be written with cheaper models and so on. I don't know why you would go to a higher-end model to write code when you can do it for free, and everyone in the organization can write code basically for free. And so I think Anthropic's bigger issue is this customer concentration question, and is it really a premium product that they're selling in the market, and that's where they take the market? That would be kind of where I think the market is headed right now and would be the biggest risk factor. This whole thing about, oh, it's gonna end the world and whatnot, it's like, dude, just make sure your product doesn't end the world, and you'll be fine. That would be a good first step here, like Mark is doing at Meta.
- JCJason Calacanis
I mean, it's an incredible company. The execution has been extraordinary. The ramp of revenue, Sacks, has been extraordinary. But when you start to look at the impact of open source, we had a, a couple of, um, viral charts come out, 70, 80% of the token usage are those dark tokens I mentioned six months ago here on this pod that are not tracked, but some people can track them on routers. Those dark open source tokens are now the majority of tokens, and to-
- DFDavid Friedberg
Can I just give you a stat on that, JCal? 'Cause-
- JCJason Calacanis
Sure
- DFDavid Friedberg
... what you just said is really important, but there's a stat that I think in tw- in the last 12 weeks alone, token use has flipped from 80/20 closed versus open-
- JCJason Calacanis
Yes
- DFDavid Friedberg
... to 80/20 open versus closed.
- 53:11 – 1:07:58
Political reactions to "Pacing the Frontier": Bernie's AI ban, Trump, Bessent, Obama
- CPChamath Palihapitiya
Let me ask you guys a question. What do you think would be the method by which, like, a ban on superintelligence... Bernie Sanders proposed this bill to ban superintelligence, and, you know, the Dems take the House, they take the Senate, and then maybe in '28 take the White House, and you end up in a world where this sort of thing is so weirdly political, which it shouldn't be. I don't understand how we ended up in a political party divide on progress. It's insane.
- JCJason Calacanis
They asked for it is the reason. They asked for it.
- CPChamath Palihapitiya
It... I... No. I... There's, uh-
- JCJason Calacanis
They provoked it
- CPChamath Palihapitiya
... it's just, it's, it's, it's so awful and weird that you're anti-progress because of your political party. The whole thing is not objective. It's a bunch of weirdly distorted worldview thinking to try and create an opponency to the individuals that you don't like, but it's very weird. Anyway, my point being, even the Republicans you're hearing murmurs about, "Well, maybe we should ban data centers, ban AI." So Sacs, let me just ask you, what is the pragmatic path with all of the open source, open weight models being out there? Is there even a pragmatic path to do what everyone's stating or talking about, or Bernie's talking about, which is to ban superintelligence, to stop AI, to turn it off? Is there even a way to do this? Like, what would happen if this law passed tomorrow? What do you think would happen, is my question.
- DSDavid Sacks
If Bernie's bill passed tomorrow, everything would stop because the way he defines artificial superintelligence in his bill, you could argue we've already hit those thresholds. I mean-
- CPChamath Palihapitiya
Right
- DSDavid Sacks
... the way... You know, it's basically just saying that models-
- CPChamath Palihapitiya
Yes, we have. Yes
- DSDavid Sacks
... at a certain capability level, and I think it's arguable whether we've already reached it or not And so it's a very loose definition. And then on top of that, he's got 20-year prison sentences for developers who violate the law. So no one's gonna wanna take the chance
- DFDavid Friedberg
So what happens?
- DSDavid Sacks
So think about the chilling effect
- DFDavid Friedberg
Do you think people delete open weight models off their desktop computers? Do businesses have to go remove code from their systems? Do, do all these people lose their jobs?
- JCJason Calacanis
I mean, I, I-
- DSDavid Sacks
Listen, let me, let me tell you. The, the, the Democrats wanna do to AI what they did to crypto. They're gonna drive the whole industry offshore. They're gonna basically drive all the innovators out of the country and-
- JCJason Calacanis
Goes offshore
- DSDavid Sacks
... goes offshore and-
- JCJason Calacanis
That was literally where I was going with it
- DSDavid Sacks
Yeah
- JCJason Calacanis
... is, like, just move your company to Singapore, move it to Zurich, move it to a geo where people-
- DFDavid Friedberg
'Cause that's where, that's where my, it-
- JCJason Calacanis
... are embracing open source and are embracing super intelligence. It's totally-
- DFDavid Friedberg
But this is literally where my head goes. I'm like, "You can't stop AI. You can't just tell people a class of software can't exist." In my head I'm like, "Okay, if the US passed a law that said, 'This class of software cannot exist,' I would move to another country."
- JCJason Calacanis
Yes.
- DFDavid Friedberg
Other people would... Like, w- like, isn't that the only-
- JCJason Calacanis
People would give up their citizenship and move to Singapore. That's the answer. Peop- literally you would be s- taking Google's AI-
- DFDavid Friedberg
Did you guys see that Xi, Xi's, Xi spoke at the White House today, and he said, "I'm gonna invite 100,000 young Americans to come and tour China"?
- JCJason Calacanis
100,000 young Americans. Yeah.
- DFDavid Friedberg
"Yeah, come and see what we're doing here." Like, it's, it's a very interesting moment where our people are saying, "Let's stop progress, technological, economic prosperity," the opportunity for everyone in that K-shaped economy to lift themselves up because open source AI enables everyone to increase productivity. It is not AI for the billionaires. It is AI for the everybody. And then we're gonna turn that off. And then meanwhile you've got China saying, "Come over here. Come and see what it's like over here, where everyone gets to use AI and everyone gets to build."
- JCJason Calacanis
Which was our playbook, Friedberg brilliantly stated. It was our playbook to get the top Chinese scientists, the top European scientists, the top scientists and thinkers and mathematicians in India, in Germany, and invite them to come to the United States. That was our playbook. China's taking our playbook, and then they're gonna try to get investment will be the, obviously the next shoe to drop. Here is President Trump, 30 seconds at the United Nations renaming AI to super intelligence.
- 1:07:58 – 1:27:13
Meta launches Muse, AI "alignment," Oracle's Force Majeure
- CPChamath Palihapitiya
and I think if we're gonna look at the light here, perhaps people's big gains, we thought maybe ChatGPT would be that, but people just use that as, like, glorified search. But man, agents, as expressed through Muse, feels like the first time tens of millions of Americans are gonna get some demonstrable value from AI, you know, things that will s- make their lives better. Muse hit number one last Friday in the App Store. Stock, Meta stock was up 10% after this new agent, uh, Muse was released. Meta said the design, heavily inspired by OpenClaw, which we talked about in January when everybody got one-shotted by OpenClaw, but obviously, that product was not easy to use, and this one is free. And it's been downloaded three million times in essentially 10 days. This product is extraordinary. I've been playing with it, comparing it to Grok Bot. It's not as freewheeling, but it's really good at getting [beep] done. Has anybody played with Muse, and does anybody have thoughts on Americans feeling really like they're gonna get something out of this AI revolution and not be left behind? They put me into the test flight a couple weeks before it was released, and I started to play with it. It's really quite excellent, I have to be honest with you. Because it simplifies a lot of these more complicated technical capabilities into a very simple, usable interface that can solve the basic problems that people have. I have it triage my personal inbox. It does it relatively flawlessly. If you ask it to book a flight, it does that. If you ask it to book a hotel, it can do it. So what are we learning? Both Grok Bot and Muse show that, A, what you said, Jason, this is just software.
- DFDavid Friedberg
[laughs]
- CPChamath Palihapitiya
And the scaled manifestation of software is utility, and what Friedberg is saying. It improves the lives of everyone. This is not about rich people getting richer. This is just simple, useful capability that now is available at everybody's fingertips for free. F-R-E-E is a pretty great price here, David, and it's-
- DFDavid Friedberg
Pretty great price
- CPChamath Palihapitiya
... pretty great price, and it does [laughs] everything, you know, that, uh, let's face it, OpenClaw with 10 hours, 20 hours of technical setup, or Claude Code at $200 a month would cost. It's now taking that experience, which then leads me to believe, wow, how did Google miss this? Man, Google's gotta get in the game, but... And I, I understand they have this product ready to go. It's the rumor on the street, and they've also got a frontier model they're gonna drop. That's the other rumor on the street is that they're gonna be dropping some hotness soon. But man, how did they miss this, Sax? And, and what are your thoughts here on maybe the bottom half, uh, in the K-shape recovery here, getting some efficiency dropping to their daily lives and turning around the negative toxic AI sentiment in the country?
- DSDavid Sacks
I think that product releases like this will help a lot. I mean, Mark Zuckerberg said in his essay that he thought that AI capability should be decentralized, and here he is walking the walk and creating a product that's very easy to use. I mean, look, ever since OpenClaw was released, it was the most obvious business opportunity in Silicon Valley is to take the concept of Claw, but make it very easy to use. You know, solve the usability issues, solve the security issues, make it reliable, make it predictable, and that's what they seem to have done. Grok Bot did it, too, but now they're kinda going, I think, one step further. So yeah, look, I think if a billion people start using, you know, personal AI agents as their digital assistant just to help make their lives more efficient, you save an hour or two a day because your AI agent's doing all these tasks for you, it's gonna create a much more positive impression of AI or super intelligence than, you know, what the media portrays. I think ultimately that's the-
- DFDavid Friedberg
Yeah
- DSDavid Sacks
... that might be the solve here to the public sentiment issue is just, like, the more people use these products and like 'em, it'll improve their approval rating, but more importantly, it'll demystify them-
- DFDavid Friedberg
Yes
- DSDavid Sacks
... and people will be less afraid of them.
- CPChamath Palihapitiya
People love awesome products, number one, and also people love cute products, Jason, and the logos and the iconography for both-
- DFDavid Friedberg
It's certain, yeah
- CPChamath Palihapitiya
... Grok Bot and Muse is lovely. It's delightful. It's not stressful. You don't feel like there's a lab leak [laughs] coming around the corner. You feel like, wow, this is nice, and I like it.
- DFDavid Friedberg
And they're, and they're not trying to take your jobs. They're bots and assistants for you.
- CPChamath Palihapitiya
[laughs] Right. Exactly right.
- DFDavid Friedberg
They're not going to kill you. They're not murder bots.
- CPChamath Palihapitiya
Exactly right.
- DFDavid Friedberg
Nor are they trying to steal your job.
- CPChamath Palihapitiya
[laughs] Exactly right.
- DFDavid Friedberg
They're trying to make you better at your job. I always tell founders there's three ways, three ways to make money in the world. Save people money, make them money, or entertain them. And you look at this bot, I started using it. I made a commerce bot both on, on both platforms, and I said, "You know, I... When I wanna buy something, I wanna buy this anchor thing, you know, this new hub for my, for my, um, Mac Studio coming." And I was like, "This thing costs $400. Just for giggles, um, tell me where I can get it at a better price." And it was like, "Absolutely
- JCJason Calacanis
You can, if you go to the Anchor website and you get, uh, if you're a first-time customer, you're gonna save 85 bucks. Now, I don't need to save 8- 85 bucks, but I was like, "Well, that's intriguing." And it then routed me around Amazon, and I'm like, "Okay, I'm gonna buy it from Anchor as a first-time buyer and save 25% on this product." That's why Amazon this week said, "We gotta block these things," and they've been taking action over and over again. First they did Perplexity. I think now they're going after Muse and other bots. This is gonna save people money. It's gonna go into their email and say, "What services are you subscribed to that you're not using? Let me unsubscribe it for you. Let me do your returns for you." That's gonna be absolutely amazing for everybody to have a chief of staff executive assistant house manager who fixes stuff for them and saves them money.
- CPChamath Palihapitiya
Both Grok Bot and Muse do two things which I think are very important. One is it'll force all the big companies who have resident services to essentially block all these third-party services. They're not gonna allow these things to happen, and the reason is exactly what you said. There's price discovery and transparency, and I think that that's bad for opacity. And if you benefit from opacity and leakage and breakage, this gets rid of that, and I think that that's, that's bad for a whole host of companies and services. The second thing, though, is really interesting. Things like Grok Bot and Muse really put the App Store and its 30% rev share on notice.
- JCJason Calacanis
Ooh, good one.
- CPChamath Palihapitiya
Why? Because all of a sudden, what you do is you force many of these services to exist headlessly, right, where the UI is less important, and really what you want is to be able to transact and navigate f- on behalf of agents that come because a consumer has deployed them to your website. In that world, Jason, I don't think that there is any reasonable claim that any of the App Store owners can make about why they should get a rev share. And so I actually think that's the biggest thing that, that this starts to question in my mind is, wait a minute, like, games could be constructed differently, right? Like, for example, if you said to Muse or to Grok Bot, "I wanna play a game," it finds you an experience, and it can just serve that to you.
- JCJason Calacanis
Amazing.
- CPChamath Palihapitiya
And, and you make in-app purchases. There's no re- and now you have Stripe integrated naturally. There's no reason why the App Store actually gets that flow of funds. So on the surface, it's utility, but underneath the waterline, I think that this is a really important moment for some apps.
- JCJason Calacanis
Oh, such an understatement, Chamath, and I love when we get to real examples here because s- tons of people pay for Wordle. You could just go to Grok Bot, "Hey, give me a Wordle. Uh, give me another one."
- CPChamath Palihapitiya
Yeah.
- JCJason Calacanis
And, and I-
- CPChamath Palihapitiya
Yeah
- 1:27:13 – 1:34:10
Anthropic’s bio research and "wet lab" in SF
- JCJason Calacanis
Freiberg, uh, you know, after the Wuhan thing and, and all of this, uh, hand-wringing around the e- the end of days, sh- what does the scientific community think of Anthropic launching a lab, and what is the purpose of them having a physical, biological lab?
- DFDavid Friedberg
There's hundreds of labs that can do things like testing proteins and testing enzymes for function in this area, in the Bay Area, and these labs are not making viruses. They're not making pathogens. That's not what, what's going on. Basically, if you look at the paper that they just published, they did a preprint. They took large amounts of DNA data, and they sent a bunch of Claude agents to try and analyze the data to try and identify novel enzymes, novel proteins, that hadn't been characterized before by just looking at the DNA data, and the system effectively found what looks like a really interesting enzyme that looks like a CRISPR-type enzyme. These are the things that allow us to do things like gene editing to repair genetic defects or help people with certain diseases, and that this discovery that they made may actually yield a number of therapeutic pathways. So the question then is, okay, well, we discovered this. Like, how do we test it? How do we show that it is what it is, that this protein actually has some function? So the predominance of the work that goes on in a lab like this is actually taking DNA, putting it in a bacteria to make a protein, and then measuring what that protein does. Does it have a specific function? And so protein discovery is the, the baseline of all antibody discovery, which is the majority of the therapeutics industry today, is discovering novel proteins that we can use as a therapeutic agent. So that's the sort of work that's going on. It's BSL1, BSL2, which means, you know, these are bio-safety levels. The lab that they set up-
- JCJason Calacanis
Yeah
- DFDavid Friedberg
... they've publicly talked about the lab. So the lab itself is sort of like a low-level research lab where they can make simple proteins and test them in the lab to see if the AI is doing a good job discovering or postulating protein folding or these sorts of theories, and to actually run really rapid testing and experimentation cycles to see if these proteins are what they're predicted to be.
- JCJason Calacanis
Okay.
- DFDavid Friedberg
Uh, so I would say that-
- JCJason Calacanis
So they shouldn't run, like, 4,000 agents against it and do a hugging face with proteins, yeah?
- DFDavid Friedberg
No, I mean, I think what they're really doing is just having experimental proof that what the software is predicting... Remember, go back to AlphaFold. AlphaFold was predicting protein structure, three-dimensional structure of a protein from the DNA sequence that codes for that protein. And so in order to prove that, you have to basically predict what the structure of a protein would be from a DNA sequence, make the protein, and then look at it and see if it actually looks like it's supposed to look like. And so in the case of some of the other things that they're doing, they're trying to discover function. I wanna find a, an enzyme, which is a type of a protein, that degrades a certain molecule that can be used to degrade, for example, you know, a therapeutic target or, uh, something in the body that we wanna get rid of. And so they'll start to kind of make these predictions in software, but then you need to test them in the lab to say, okay, does that target actually do what it does? Now, there are literally hundreds of labs like this. These are low-level research labs that are not doing, like, gain of function research. They're not making viruses. They're not making pathogenic things that can escape. That's why they're BSL1, BSL2. They're bench-top labs where automation can help kind of reduce the cost and increase the throughput of the testing of what the software is making predictions around. And ultimately, the objective here, I think, is for them to get some proof that their software is actually doing what it's, uh, predicting, and they have an incredible life sciences function. I think at this point, I don't have as much insight on where Gemini's at with the new models with Google, but they have the best life sciences models at Anthropic, and I think this is the fast track to basically seeing if these models can be used by therapeutic companies, by pharma companies, by government labs to help discover new therapeutics and so on. I, I don't think that the world should be scared away from discovery R&D, research and development- into new therapeutic modalities as predicted by AI because we heard the word wet lab in Wuhan and everyone's like, "Okay, any lab is bad." We still wanna progress the frontier of therapeutics, of human health, of discovery, and this is a really important aspect of Anthropic proving that their models can add value here. I think that's the summary of it.
- JCJason Calacanis
All right. Thank you so much to our friends at IREN. They were the presenting, they were the presenting sponsor for this year's All-In Summit, our fifth. IREN's AI Cloud Lounge was packed for all three days. Did a great meet and greet with IREN on Sunday and got to take some selfies and met many of you there. Thanks again to IREN, I-R-E-N, one of our great partners. Niagen. Did you catch Niagen in the Wellness Hub for IVs? Tons of, uh, great gifts and supplements being given away. Great experience, great product. Great job to our friends at Niagen for doing the lounge again. Niagen is the official NAD partner for the All-In Pod. Yes, Niagen Bioscience created the NAD category. Now hosts 45 plus clinical studies. And PayPal, my favorite way to settle up. They had a great lounge on site at the summit where everybody got to hang out and get those beautiful embossed luggage tags, and they hosted the payments and fintech dinner, which was a huge hit. Let me tell you folks, All-In Summit 2027 will sell out again, and the applications are open now. Go to allin.com/events and apply for your chance to be in the room where it happens. What are you waiting for, folks? allin.com/events. All right. There you have it, folks. Episode 290, the All-In Pod. Four. David Sacks-
- DFDavid Friedberg
Fantastic
- JCJason Calacanis
... Chamath Palihapitiya, and your Sultan of Science, Freebird. I am the world's greatest moderator. We'll see you next time. Bye-bye.
- DFDavid Friedberg
Love you, boys.
- SPSpeaker
Besties are back.
- DFDavid Friedberg
Bye-bye.
- SPSpeaker
We'll let your winners ride.
- JCJason Calacanis
Rain Man David Sacks.
- SPSpeaker
And it said, we open source it to the fans, and they've just gone crazy with it.
- JCJason Calacanis
Love you, Wes. Ice Queen of Kin Law.
- SPSpeaker
Let, let your winners ride. Let, let your winners ride.
- JCJason Calacanis
Besties are back.
- DFDavid Friedberg
That's my, uh, dog taking a piss in your driveway, Sacks. [laughs]
- JCJason Calacanis
We are not friends.
- DFDavid Friedberg
[laughs]
- SPSpeaker
Oh, man.
- DFDavid Friedberg
My habitat, your room, meet me at Glens Falls.
- JCJason Calacanis
We should all just get a room and just have one big huge orgy, 'cause they're all just useless. It's like this, like, sexual tension that they just need to release somehow. Wet your feet. Wet your feet.
- SPSpeaker
Wet your feet. Wet-
- DFDavid Friedberg
We need to get merch.
- JCJason Calacanis
Besties are back.
Episode duration: 1:34:23
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode cvP_1jmnkmM
