A Four-Day Coma Couldn't Stop This Game ft. Johnny Monsarrat, Founder of Turbine (MMO)

The AI XR Podcast.

The episode opens with a critique of recent AI political theater, including Trump's rebranding of AI to 'superintelligence' and Dario Amodei's White House appearance, before pivoting to major corporate shifts like

Key takeaways

• AMD acquired Fei-Fei Li's World Labs for $8.2B to secure a foothold in physical AI modeling and compete directly with Nvidia's market dominance.
• OpenAI has delayed its IPO while Anthropic pursues it, largely due to mounting legal fears over public statements that AI could cause mass harm.

Main topics

• AI political theater and the 'superintelligence' rebrand
• AMD/World Labs acquisition and Nvidia competition

Notable quotes

Reality is South Park.

Conclusion

The episode underscores the tension between AI's rapid corporate scaling and its unresolved legal/safety

Transcript preview

Speaker 2 (0:00) Good morning, everybody. I'm Charlie Fink with Ted Chilowitz and Roni Abavitz for the AIXR podcast. It's Friday, October 2nd, 2026. Good morning, guys. Good morning. Speaker 4 (0:11) I'm in the middle of the country this week in Nebraska. Data center central. They got rid of the Speaker 2 (0:18) cornfields? Yeah, Speaker 4 (0:20) they're getting rid of the stuff and putting data centers in. Speaker 2 (0:22) No more corn for you. Yeah. So, SI. Speaker 4 (0:27) SI. Speaker 4 (0:30) You know, in grade school, Speaker 2 (0:31) in grade school, it meant something else. That's kind of appropriate coming from Donald Trump. So we're now going to say super intelligence. Is this like renaming a lake? Speaker 4 (0:42) Well, yeah, we talked before we started the podcast how many people know how many Great Lakes there are. Turns out there are five. You know, Charlie, you and I have talked about this. I don't really like calling AI even AI. I like to call it PI because it's actually predictive intelligence. That's what AI is actually doing underneath the covers. It's predicting using a corpus of knowledge, right? So adding a new moniker, calling it super intelligence, I don't even know where that even begins to have any reality. It's Speaker 3 (1:12) super. Speaker 4 (1:14) By the way, Speaker 3 (1:15) just for folks that did not see the outcome of this summit this week where Dario is standing next to the president and it's like someone had a hand in his you-know-what and saying, say these words now. And it was like watching a kind of those videos from overseas where someone's being held against their will or something. It's like, Yes, I do like this thing that I'm being told to say. And like this group of scary looking guys is standing behind him. And like, if you say it wrong, we disappear you to Gulag. It just had that feeling, didn't it? Speaker 4 (1:53) Well, you know, the interesting thing about this is like in the days before all this kind of inanity, like it's in complete insanity. The South Park guys had to work harder. to parody these things. Now, you gotta wonder, are they just sitting in their offices going, why do we even need to make our show anymore? Like, reality is South Park. Speaker 3 (2:12) By the way, what's even more amazing is I love these, like, neo-AI comedy videos where, like, Trump and JD are characters selling, like, street food in Bangladesh and they're rice farmers in China and it's just, like, the fact that they can cast them as they are. and make them say whatever they want. Speaker 4 (2:32) Was it one of you guys that shared, it was either on X or Instagram, you shared with me a really great AI parody of all of the world's most awful authoritarian leaders sitting and they were representing inside the friend's coffee shop. Oh my God. Did you guys send that to me? I sent it to you. I sent it to you. Oh, you sent that? It's unbelievable. It's unbelievable. Even better, Speaker 3 (2:57) even better. There is the Sam Altman Dario 70s sleazy computer sales guys on the everything computer. And then Trump does this huge party for China who shows up. Those are, these are works of art and genius. If you haven't seen these online, they're just the most incredible genius creative things I've seen. Speaker 2 (3:18) Let me understand this, because Trump super intelligence, he couldn't have given everybody a lot of notice. So when Trump has a summit or a photo event or whatever you call these things at the White House and he summons you, you're like, you know, an assistant comes in the room and whispers to you, you're Jensen, you're in Asia. He comes in and whispers to you, you have to be in Washington in 48 hours for a photo op with Donald Trump. And so, I mean, do you think that's the way Speaker 4 (3:46) it works? Probably pretty close to that. Yeah. I mean, I think it's, you know, it's all just sort of random lunacy. So, you know, whatever reason, here we are all talking about it and the rest of the world is talking about Speaker 2 (3:57) it. So. Okay. Well, let's talk about something else. Okay. AMD buys World Labs for 8.2 Speaker 4 (4:03) billion. Speaker 2 (4:03) Yeah, Speaker 4 (4:03) I Speaker 2 (4:03) thought this was, Speaker 4 (4:04) this is a really Speaker 2 (4:04) interesting, Speaker 4 (4:04) legitimate story. Is this the biggest Aqua hire in history? Potentially, yes. Speaker 3 (4:09) Well, Fei-Fei Li is very smart. And apparently there was a lot of back and forth. And I think they're early investors in World Labs. But it's near zero revenue. So it was a team, physical AI modeling simulations. And I think it's AMD trying to catch NVIDIA. And Lisa, look, Lisa Su, I met her. AMD is worth $2 billion. She just crossed a trillion. Speaker 4 (4:36) You Speaker 3 (4:37) got to give her credit. She's probably one of the great tech CEOs. who nobody talks about as much, but she's incredible. Speaker 4 (4:42) Yeah. No, I agree with all that. I think the World Lab play was very much a NVIDIA can't own all of it. We need to own a certain percentage of it. And it's very clear with, you know, what Jensen is signaling and how he is, you know, sort of gobbling up these pieces of the pie that they just want their slice. And this was a probably relatively smart acquisition for them. Speaker 2 (5:04) Well, it looks like it's not going to be a duopoly that is Intel. It's going to be AMD that is going to have the other half of the market. And I might add, NVIDIA, like Microsoft in the 90s, needs AMD to succeed. You know, they need AMD to have 10 % of the market or they're going to be very vulnerable to antitrust. and more regulation. So, by the way, just taking this another step further while we're on the topic of the mega frontier AI companies, OpenAI not going public, Anthropic going public with lots of caveats. You have an opinion about this, Roni? I'm not sure why one is going public and the other isn't. Are OpenAI's books not looking very good? Are their numbers not encouraging? What's up there? There's semi Speaker 3 (5:54) -educated perspective. I do have a little bit of insight, but not like people inside there. One, having taken a company public and helped other companies go public, you have all these risk disclosures and all your financials and everything else has to come out. So as auditors and legal teams and the SEC look through everything, clearly there's some significant risks that came out. hey, we're losing a lot of money, maybe more than some people thought. So it's kind of like, are we ready for that? So maybe one company is more ready than the other in terms of financial loss. And by the way, what is your turnaround time? Are you really ever going to make a profit or is it something that just keeps sinking? But the other piece is this constant, we're going to kill everybody on the planet, which never goes away. So the almost unlimited liability that you have to put in your registration. here's all the risks with the business. One of them is our own team has said in public and thousands of our scientists and other scientists are agreeing that our product is so dangerous, it may cause mass harm. Speaker 4 (7:02) So Speaker 3 (7:02) legal people are not letting that go. And that's probably the big thing. How do you get a company through the public offering process with the issue of we have this? So imagine a medtech company saying we're about to put out a pacemaker. And one of the chief scientists says, well, our pacemakers may kill one out of 10 people. Speaker 3 (7:23) That company is not going public. But this thing is so big and has so much behind it that despite the we may cause great harm and everyone in the world may sue us, someone's got to underwrite the liability of that. And I think there's a lot of that going on. They're trying to figure it out. And they're also making people walk back their statements. Oh, you didn't really mean that. Those people have been fired. They've been gagged. That was nonsense. That was a psyop. Because there's so much money at stake that lawyers who are going to sue, if anything these companies ever do causes harm, just like every lawyer in the country and around is going to sue them. Because they said it was going to cause harm and their leadership agreed with it, at least at some point in time. So the liability seems like to have infinite. Like a pharma company, if it was going public, says our drug. could potentially kill everyone on the planet, not just hurt people, but again, would not get cleared. The FDA would shut it down, wouldn't go public. So that's a conundrum here. But somehow people see this as having such importance. It overrides things that would not stop, that would stop every other category in every other sector. So it's really weird what's happening. Speaker 4 (8:37) And then there's Jensen as a, here's two sidebars. Jensen saying, well, we're going to go ahead and put a lot of money and resources into kind of trying to self-regulate this industry and make a safe version of AI using NVIDIA products to do it. Okay, that's not more self-serving than anything. One more moat. Speaker 3 (8:57) Look, let me give props to Jensen. By the way, he and Lisa Su are cousins, two of the smartest people in the world. Jensen is a adult in the room who does have a notion on how to make things safer. I'll completely give that to him. The other folks may not. And he's saying, you people are out of control. I'm going to have to put this down at the chip level because I don't know what the heck you guys are doing at the top. I'm going to have to control this at the chip level. And he's calling out the president saying, those guys are bonkers. I'm a responsible adult. Let me frame the whole thing here. And I think that's what he did. I get that. And maybe he saved the whole sector. I Speaker 4 (9:35) get that. But it is awfully. very transparently self-serving to say the only way to fix this is to do it at a chip level when he's the guy that makes the chips. Like, I get it, but there's a direct conflict of interest there. Speaker 3 (9:47) I just want to, I want to props to Jensen on a couple things. Not only is he smart and often can be right. but it also helps him and his company at the same time. Speaker 4 (9:56) And Speaker 3 (9:56) that's why he's built a $5 trillion plus company because he does all of those at the same time. Speaker 4 (10:01) It's sneaking up to $6 trillion. By the time we do our next podcast next week, it may be over $6 trillion. It's $5.7 trillion. Speaker 3 (10:09) Let's just double click on Jensen one more time. He's trying to say, tone this stuff down, use this to build prosperity and functional things. which is a very outside US worldview. Like outside the country, let's just use it for stuff, for technology. What did Anthropic do? They brought religious leaders. Did you guys read this article? Speaker 2 (10:32) Yes. Speaker 3 (10:33) They brought a whole cohort of religious leaders to Anthropic to try to explain to them that Claude was conscious and maybe waking up. And they brought priests and rabbis and imams. And Jensen's like, what are you people doing? We're selling tech things. And you guys are on some weird hippie psychedelic trip. And he's trying to bring the industry into a sense of normalcy, like it's computing. Wait, Speaker 2 (10:57) I thought you liked the hippie psychedelic trip. Speaker 3 (11:01) I do, but this is crazy what they're doing. I don't know. It's craziness. All Speaker 4 (11:06) right. Here's another little sidebar that we should talk about for a second. On the open AI side of the equation, this is something that a lot of people are talking about right now. So I want to bring it out into the public a little bit, a little public sphere. Did Sam and all of his sort of brethren at OpenAI just have a really bad whatever they acquired Johnny Ives' company, $6 billion or so, when Meta and Zuckerberg released their little Tamagotchi AI thing, their little Muse Tamagotchi situation? Is that what Sam and Johnny Ives were building? And did they just kind of get beat to the punch? That's an interesting open question. Speaker 3 (11:48) He could be crops on Mark Zuckerberg. He senses, not accidentally. He knew that that's what they were Speaker 4 (11:53) doing. He's like, we could outdo them. Speaker 3 (11:55) He's got the largest data intelligence gathering thing in the world next to Google. And he moves fast and he copies well. And he's smart. So yeah, he did. Speaker 2 (12:06) So you Speaker 3 (12:06) know, Speaker 2 (12:06) OpenAI had a dev day this week. And they introduced something called DOTS. And it's a counter to Meta's Muse. You know, we were talking about their Tamagotchi. Speaker 2 (12:21) Is Joni Ive making the same thing? Well, I've said many times, everybody is trying to get to the body, to get close to you, to get in between you and your smartphone, to be your intimate companion. So this is, you know, because once you get engaged with a language model, it creates a real, you know, switching barrier. By the way, there's something Speaker 3 (12:45) about Dots that you guys, at that same dev conference, Charlie, I don't know if you saw this, but. They announced a partnership with Mod Retro. Yes. Which is Palmer Luckey's retro video handheld thing. But it was kind of cool. You can use Codex to basically hack games into a special Mod Retro. So if you were at the conference, you got it. Which is kind of weird when you think about it. It's kind of this retro video game vibe, really friendly. But behind the scenes, you got a defense tech company. And then you have this other kind of like, we're going to kill everyone on the planet coming. So, but look at this cool thing we're doing. And dots are really friendly. Sesame street, street style. There's, there's Speaker 2 (13:27) a Speaker 3 (13:27) strange. Speaker 2 (13:28) Yes. They're trying to give them personality. Do Speaker 4 (13:31) you guys remember at least four times on this podcast? I talk about something I refer to as the seven sisters effect, which is when there's a generally relatively good idea about something in technology, you get seven versions of it almost instantaneously. And lo and behold, we're seeing it again, right? We saw it with email clients. We saw it with web browsers. We saw it with video chat clients. And now we're seeing it with these sort of AI-ish chatbot things. You can do the same thing in Muse that you can do in Dots, that you can do in Gemini, that you can do in Anthropic with Clint Claude. Speaker 4 (14:07) Soon there will be seven, and then two will exist and be the leaders, and the other five or six will sort of live in obscurity. And it's proven over and over and over again. Let's call something Speaker 3 (14:18) out that most people don't know. First, there's Clippy. Speaker 4 (14:21) Yes. Clippy is this thing, but Speaker 3 (14:24) it's Clippy 3.0. Okay, Clippy now is so old, Speaker 2 (14:29) you