Pod Insights Moonshots with Peter Diamandis The AI Duopoly Is Over: Grok 4. 5 , GPT-5 . 6 , and Muse Spark in One Week | #270 About this episode In this episode the mates discuss Grok 4.5 vs GPT-5.6, Apple Suing OpenAI, and China catching up to Elon.
Get access to metatrends 10+ years before anyone else - https://qr.diamandis.com/metatrends
Peter H. Diamandis, MD, is the Founder of XPRIZE, Singularity University, ZeroG, and A360
Salim Ismail is the founder of Open ExO, a GP at Exponential Venture Capital/The Organizational Singularity Fund and a sought after global speaker and thought leader.
Dave Blundin is the founder & GP of Link Ventures
Dr. Alexander Wissner-Gross is a computer scientist and founder of Reified
–
My companies:
Apply to Dave's and my new fund:https://qr.diamandis.com/linkventureslanding
Go to Blitzy to book a free demo and start building today: https://qr.diamandis.com/blitzy
Your body is incredibly good at hiding disease. Schedule a call with Fountain Life to add healthy decades to your life, and to learn more about their Memberships: https://www.fountainlife.com/peter
_
Connect with Peter:
X
Instagram
Substack
Website
Xprize
A360
Connect with Dave:
Web
X
LinkedIn
Instagram
TikTok
Connect with Salim:
LinkedIn
X
Join Salim’s 10X Shift
Subscribe to Salim’s YouTube channel
Exponential Venture Capital
Connect with Alex
Website
LinkedIn
X
Email
Substack
Spotify
Threads
Listen to MOONSHOTS:
Apple
YouTube
–
*Recorded on July 11th, 2026
*The views expressed by me and all guests are personal opinions and do not constitute Financial, Medical, or Legal advice.
Learn more about your ad choices. Visit megaphone.fm/adchoices
Transcript 17,530 words · 12 speakers
0:00 Speaker 00 July 8th, Elon announces Grok 4.5. July 9th, OpenAI releases GPT 5.6.
0:06 Speaker 01 We have four American labs now at the optimal frontier.
0:09 Speaker 02 This is a convergence episode.
0:12 Speaker 00 Apple sues OpenAI for trade secret theft, accusing them of stealing trade secrets to build their upcoming AI hardware. The two companies, which were partners just, you know, literally a year and a half ago, are now in an all-out legal war.
0:27 Speaker 02 I think this is Apple trying to slow things down while it catches up.
0:31 Speaker 03 Yeah, it's a pretty aggressive move and you have to be, you know, pretty hurting to do it.
0:37 Speaker 00 For the first time, China, Long March, which manufactures their primary launch vehicle, demonstrated landing of an orbital booster.
0:46 Speaker 01 I think we want propulsive landings to be as competitive as possible. We really don't want.
0:54 Speaker 00 Now that's a moonshot, ladies and gentlemen.
0:59 Speaker 00 Everybody, welcome to Moonshots. This is the number one podcast in AI and exponential technologies, your front seat to the singularity. I'm here with my magnificent moonshot mates, the original four: Dave London, Salim Ismail, Alex Wissner Gross, I'm Peter Diamandis, your host. Here in my homeland of Greece, behind this curtain here is the beautiful Aegean. All right.
1:21 Speaker 03 Pop it open, let's see. Open the curtains. Go open the curtain. You can't say that in the night. Come on.
1:27 Speaker 00 I'll, uh, yeah.
1:28 Speaker 03 Yeah, that's probably a dumpster back there.
1:31 Speaker 01 You're in a secure location?
1:34 Speaker 00 Yeah, I'm in an absolutely magnificent location. And just like Saleem, Saleem, where are you today?
1:39 Speaker 02 I'm in Barcelona at the Festival of Consciousness.
1:43 Speaker 00 Yes, you're still there, right? Well, hey.
1:45 Speaker 03 You know, I was at a meeting at Allen and Company headquarters in New York once, palatial ridiculous Allen and Company, meeting with Charlie Allen, and we were in this conference room that was covered with Remington original sculptures and Norman Rockwell paintings. And at one point, he said, hey, do you want a water? And the entire back curtain and wall opened up and there's like a five-star kitchen full of people with these like, you know, big chef hats cooking back there. He said, no, that's not for you. I'll just get you this bottle of water and close it. So that's what's going on behind your curtain right there. You just don't want to.
2:17 Speaker 00 No, no, no. Okay, one second. Is the Aegean over here?
2:21 Speaker 03 Ah.
2:22 Speaker 01 Oh, my God.
2:23 Speaker 00 Yeah, beautiful. You know, honestly, I'm sacrificing a lot to be with you guys right now. But hey.
2:29 Speaker 03 Hello, yes.
2:31 Speaker 01 Wow. Your sacrifice will be remembered, Peter.
2:35 Speaker 00 It's gorgeous. You know, you know, you can, you know, there's a lot of wealth when all you're seeing is yachts every place. It's insane. Not mine.
2:44 Speaker 03 You can swim out to those. I've seen James Bond do it. Swim out and crawl at the anchor chain.
2:49 Speaker 00 So.
2:49 Speaker 01 Yachts are 20th century wealth.
2:52 Speaker 00 I tell you, I can't wait for my space yacht to happen.
2:55 Speaker 01 There we go.
2:56 Speaker 00 There is so much happening this week, guys. Everybody listening, we spun up this bonus episode because, you know, it's just crazy. So I tweeted yesterday, you know, we're obviously in the singularity, living through it because robots are performing surgery. A single SpaceX booster has been reused and re-flown 36 times. We released four models in seven days and Elon casually claims SpaceX is going to be worth more than the entirety of planet Earth. You know, so I'm not exaggerating when I say there's a lot happening this week. Please remember our mission here at Moonshots: keep you informed, keep you up to date, keep you optimistic about the future. So I think this week, Alex,
3:39 Speaker 00 is going to be your test on exactly what's going on in the world because it's kind of like frontier model release week, if there was such a thing.
3:49 Speaker 01 Are we eval for the evals?
3:52 Speaker 00 Yeah, well, I, you know, it was not too long ago when we were releasing models like, you know, every six months, every three years. Now we've got four models in seven days. It's kind of insane, isn't it?
4:05 Speaker 01 Turns out the singularity isn't just vibes.
4:08 Speaker 00 Well, it's going to be a continuous model release. I mean, this idea of a single model being released and calling it a release, I don't know. But July 2nd, Fable 5 is re-released. July 8th, Elon announces Grok 4.5. July 9th, OpenAI releases GPT 5.6, family, Sol, Terra, Luna, and Ultra Mode. I love that. That runs four agents in parallel. July 9th, Meta shows signs of life with the release of Muse Spark. So, Alex, I'll show a couple of charts here, but without going into, you know, an hour long on the benchmarks, help us understand what's going on with these four releases. I really want to get, one of the things I'd
4:50 Speaker 00 love, Dave and Saleem and Alex, is to help people understand how these differentiate and how does it impact them individually when they hear all these releases being, being made?
5:03 Speaker 02 For me, the zoom out is this is a convergence episode, right? We're seeing intelligence get cheaper, interfaces are becoming more human with the hand thing, and AI is acquiring tools and hands. It's all unfolding in the one week.
5:16 Speaker 01 Peter, can we put up the, the Artificial Analysis Intelligence Index, which I think is the, yeah, yeah. So for those listening, this is a scatter plot produced by Artificial Analysis. It's, it's tracking their estimate. It's a composite index, sort of a meta benchmark of cost per task for each model versus the score, the actual performance on each task. And I would say in my mind that the headline from the past few days, which I agree, Peter, have been absolutely incredible, is we went from a beginning of the week where we were basically, the frontier was a duopoly between Anthropic and OpenAI. And if you want to be on the frontier, you have to be using models from one of those two vendors. And
6:01 Speaker 01 now, I think this is the best of, or close to the best of all possible worlds, we have four American labs now at the optimal frontier. So that's Anthropic and OpenAI, but now joined by Meta and SpaceX AI. Google, by the way, still not on the optimal frontier. Gemini 3.5 Pro reportedly delayed. Hope that Google catches up ironically to the optimal frontier and we get five American labs. And then if you look at this chart that you're showing and you sort of mentally draw an optimal frontier. So you want to be lowest cost, highest performance or highest accuracy across this composite. You want to be at that frontier.
6:41 Speaker 01 We have two Chinese labs, Xiaomi and DeepSeek slash High Runner. And also GLM 5.2 isn't at the optimal frontier for, if you ignore open weights, if you, if you just focus on the open weight frontier, GLM 5.2 is also at the open weight frontier. So we were ending the week within a world where there are at least four American labs again at the frontier. It's no longer a duopoly. And OpenAI with the GPT 5.6 release is openly promoting recursive self-improvement. They, they were openly promoting that Sol, the high-end weight class
7:24 Speaker 01 model in the GPT 5.6 series, was being used to post-train Luna, the low end, as an experiment. I'll take what I can get on the experiment. Lots of other thoughts, but those are some preliminary ones.
7:37 Speaker 00 Dave, have you started playing with these yet?
7:40 Speaker 03 Yeah, yeah, pretty much across the board, except not GLM yet. Still worried about code injection and, but it's, it's a really compelling price point. And also, you know, Alex Karp had that rant about control your own destiny, which I really took to heart last week and we showed it on the podcast. And, you know, one of the ways you can control your own destiny is run your own model on your own hardware. Also, you know, Fable 5 is just getting slower and slower and slower. And so I think anybody at the frontier is going to be sold out to the extent that they have compute. And so I suspect that the revenue will start to disperse because, you know, Fable 5, if it keeps getting slower because there's so many users trying to pile on, then you have no choice but to kind of fan out and try other things.
8:24 Speaker 00 You know, I was thinking about how I would like to explain this to everyone listening. And the realization is, of course, these models are getting faster, they're getting more powerful, they're trending towards AGI. You know, Alex, I know you think we've reached there already, so trending towards eventually ASI.
8:42 Speaker 01 Intelligence too cheap to meter.
8:44 Speaker 00 Yeah, well, they're, and they're, and they are heading towards a, you know, de minimis, fully demonetized cost. My realization is, I think what the real moat is, is no longer how smart they are, how cheap they are, it's their distribution. And so let's take a look at the following. So when I think about how to use Meta, Meta's being used inside, I mean, sorry, Muse Spark, the latest model, is being used inside Meta's product. So Meta has 3.6, roughly 3.56 billion, you know, people daily using their products. So I think about using the Muse Spark inside of WhatsApp and Messenger and Instagram and Facebook.
9:28 Speaker 00 Google, you know, I know Google's hasn't been sort of breaking news in the last week. But when you think about Google's distribution, it's immense, right? 900 million in their app, 2.5 billion being reached globally. And they're they're about to release into 2.3 billion active Apple devices with their 1.2 trillion parameter model. So Google's not being written off. I think they're going to come back and punch really hard. And then you've got, you know, OpenAI with a million, sorry, a billion monthly active users. And Anthropic is obviously not the consumer space.
10:10 Speaker 00 It's in the corporate space and it's the go to there. But would it be fair to say that we should think about the distribution of these models? Their ability to reach the end user as one of the one of the moats that's going to be most important for them in the in the future?
10:27 Speaker 01 I know Sam has mentioned distribution as something that he cares more about than the model weights themselves. On the other hand, Sam said that before this massive OpenAI pivot from consumer to enterprise. I'm not sure that distribution is actually super valuable from the economic perspective of the frontier labs. Why? Because consumers don't have an enormous amount of money to spend on this. That's what OpenAI had to discover.
10:56 Speaker 00 That's like my point I'm thinking about from the from those listening from our subscribers. How should they think about it? I'm thinking that they're going to use the model where they're using tech, right? They're going to use it in the application. And I mean, there's going to be the high end user who's going to go like you, Dave, who's going to go and say, I need Grok 4.5 for this. I need, you know, 5.6. But for the average person, you're just going to use the model embedded in the application layer where you happen to be doing something.
11:29 Speaker 03 It's a great point. There's a huge bifurcation coming across those two different use cases. We should talk about them separately. Go ahead, Slim. Sorry, I should. Good, yeah.
11:37 Speaker 02 Well, clearly, we're going to end up with a Jarvis-type model where you have a request or something that you want done. And this happening inside the models already, but I think we'll personalize it where you have a separate layer that says for this task, go to this model, for this task, go to this model, for this task, go to that model. That's clearly going to happen very soon. It's starting to happen already.
11:55 Speaker 00 You're not going to say it. You're you're Jarvis. You're Jarvis will say it.
11:59 Speaker 03 Yeah, also, these personalized movies, personalized entertainment is coming very, very soon. We'll talk about that later in the pod, but people aren't going to question the model under the covers when they start using that use case or Meta or whatever. They're just going to use it by the bucket ton. And it's going to be sold out.
12:16 Speaker 00 The bucket ton.
12:17 Speaker 03 The bucket ton. It's a it's a digital. It's a digital.
12:21 Speaker 00 It's a new benchmark.
12:24 Speaker 03 But yeah, no, Jevon's paradox is is going to be in full bore, high gear. Actually, did you see that? Highlink said that, you know, memory chips are going to be in short supply for at least five years. So that shows you.
12:35 Speaker 00 I think they said decades.
12:37 Speaker 03 Decades, yeah. I mean, I believe it. In fact, I'm 100% sure of it. The desire and the use cases are on a much, much steeper up course than the efficiency gains. And the efficiency gains are huge. Again, this is 100 to 1,000 X this year in efficiency gains, but it's not enough to keep up with the demand. So I think you're right. I think distribution is going to dominate the embedded use cases, which is the vast majority of consumer use cases. Meanwhile, self-improvement, research, mechanical design, engineering, corporate use, you know, the white collar use is going to be dominated by the best price performance model and this continual leapfrogging you're seeing on the chart here.
13:16 Speaker 01 Yeah, I'm not sure there's an enormous amount, though, of money, and I'll come back to this in a second, to be made at the low end. So I fully buy the thesis that at the edge, these models are just going to get built into the operating system. You'll have in a couple of years, you'll have Fable 5 class models maybe built into just the operating system of the laptop or or other mobile device that you use. But I don't think that's going to be a moneymaker. And the reason why it can, on the one hand, be transformative for billions of people, but while, on the other hand, not be transformative of, say, the frontier is because ultimately it's going to be the profits that pay for the data centers
13:56 Speaker 01 that pay for the the high end of the frontier models that are making all the scientific and engineering discoveries. So I think there's a plausible scenario where we end up in an intelligence polarized world where at the low end, everyone has some form of intelligence, too cheap to meter, superintelligence on their wearable that's run locally and it looks a little bit like Openweight models or Gemma or something else. And then at the high end, you have these nuclear powered data centers, superclusters that are running the frontieriest of the frontier models at the high end. And those are the ones that are just producing discoveries and inventions and where discovering or inventing something first matters. If you don't care about privacy, you can wait.
14:41 Speaker 03 Great thought experiment for you guys. In the near term future, like within a year, there'll be a competing use case between a US entity trying to make a movie, an AI character movie versus somebody in Europe trying to cure a disease. And both will be starved for access because we're just desperately short on access. Who's going to get the compute in that in that race?
15:07 Speaker 01 Well, it's not going to be Europe, right? Because Europe is energy starved, so they can't even afford the right now, the energy for the compute to cure the disease. It's it's going to be an American Hollywood studio versus an American frontier lab developing drugs.
15:20 Speaker 00 You think it's going to be the consumer from Netflix who has more demand than the than the orphan disease will get cured? That's what you're thinking.
15:28 Speaker 03 And that's sad. I think it's trending that direction, though. I completely agree.
15:32 Speaker 01 But the tokens will flow, you know, the spice will flow, the tokens will flow, as I've mentioned previously, to the maximum revenue per token. And right now, that looks like everything that's downstream of code generation.
15:44 Speaker 03 But well, the question more ties to votes, you know, votes matter a lot, too. Sorry, go ahead.
15:49 Speaker 00 But the point I want to make a bit earlier is I think this verticalization where, you know, Elon's got X and Zuck's got all of his his his app layers is an important thing that that OpenAI is going to look for. We're going to talk about their, you know, their input output device with Johnny Ive in a little bit. And and Claude is obviously revenue driven in the in the engineering space. But I still think that from the majority of people on the planet using it, they're going to care about the application they're using and whatever underneath that is going to be pretty much invisible to them.
16:32 Speaker 03 Yeah, but if you ask me, would I rather have the entire user base of X, Meta and Google combined? So about seven billion or TSMC's fabs? Which would I rather have? I would much rather own Intel or TSMC because that's that's where it's all constrained on the supply of compute level or the space-based data centers, you know, I'd rather have Elon's launch capability or a 4 billion users. I'd rather have the launch capability because, again, it's all starved at the front end of the funnel.
17:03 Speaker 01 And the market valuations, I think, reflect that, Dave. What I would also add is if you look at both Meta and SpaceX AI and how they both bought themselves these AI superclusters, it was the same mechanism. To Peter's point, they, Zuck and Elon, both used consumer distribution plays, X on the one hand and Meta's family of apps on the other, to justify an enormous capex build out of these data centers. And then the moment they have these data centers, the the family of apps on the Meta side or X on the SpaceX AI side become almost an afterthought and they move down the stack to selling compute as a service,
17:44 Speaker 01 to hosting potentially in the case of Meta compute soon, already in the case of SpaceX AI's superclusters, to selling their compute to the other frontier labs and then trying to move up stack to the to the AI. So just to tie a bow on the discussion of how valuable is distribution? I think distribution from the company's perspective, not from the consumer's perspective, is enormously valuable as almost a bootstrap to justifying to capital markets. Why should I, Zuck, I, Elon, be able to spend tens to hundreds of billions of dollars building out these AI clusters? But then the moment those clusters are built, you turn around and the distribution actually isn't that valuable.
18:26 Speaker 00 This is a chart I put up here comparing GPT 5.6 Sol, Grok 4.5, Muse Spark and Claude Fable 5 against some of our benchmarks. What we've seen over the last, you know, year and a half, as we've been discussing this on the pod, is just a continuous movement up and to the right, so to speak, or down, up, down to the right, as the case may be, and convergence. At the end of the day, are we going to see, I mean, until one of these models sort of breaks out and hits recursive self-improvement towards ASI, are we going to see them continually just leapfrogging each leapfrogging each other? Is that just the the future we're expecting?
19:06 Speaker 01 I certainly hope so. I hope so. I expect so. And the recent actions by the White House suggest that there may even be a smoothing function wherein if one model gets too far ahead of the others and surprises not just the, remember, strategic surprise can take multiple forms. You can have a frontier lab strategically surprising the other frontier labs, but given that the frontier labs are also ahead of the US government, any frontier lab that's sufficiently surprising to the other frontier labs will also probably at this point be strategically surprising to the US government. And the US government, with its monopoly on force, has the power to step in, as we've just seen over the past few weeks, and basically slow things down locally enough for the other frontier labs
19:48 Speaker 01 to catch up, which is exactly what we saw happen with Fable 5 v OpenAI's GPT 5.6.
19:55 Speaker 02 Which I think is a better way of getting strategic equivalence across these, right? Rather than governments taking a piece of things, et cetera, et cetera. But when I look at these benchmarks, so by the way, I think of it more like a map rather than a scoreboard, right? Because if you tweak your reasoning budget, everything changes. If you tweak your rationale and your use case, everything changes. So if you change your harness, everything changes. So I think I see it more in that frame. And the models will just keep getting better and better, and the cheaper and cheaper, which is the amazing part.
20:26 Speaker 01 And we also, if I might just add, Peter, we also see, the other headline in my mind is we see OpenAI, again, trying to become Anthropic faster than Anthropic can become OpenAI. So we don't have a slide for this, but OpenAI also, as part of the GPT 5.6 released, released something called ChatGPT Work. I don't know if you guys have played with it. I've been playing with it quite a bit. The new ChatGPT native app, I'll speak, you know, break the fourth wall directly to OpenAI, the native app is a hot mess. It looks like OpenAI was sort of slavishly trying to copy the Claude app with its different modes. It has like the code mode versus the cowork mode versus dot, dot, dot. Looks
21:11 Speaker 01 like just sort of a cargo cult version of the Claude app. I was hoping to see out of this new app that it had been rumored for a number of weeks that OpenAI would release a super app that elegantly unified all of its different work streams and unified the Atlas browser, all of the computer use assistant stuff, codecs, et cetera. That's not really what we saw. For the life of me, I don't understand what the point of the toggle between ChatGPT work and ChatGPT codexes seems like a little bit of a placebo, other than maybe enabling a new canvas mode, or maybe it's a rebranding of canvas. So message to OpenAI, please get your consumer client app, native app, act together and make this whole thing
21:54 Speaker 01 as elegant as it deserves to be. But I think this is emblematic of OpenAI trying to basically look end to end like Anthropic, because Anthropic was the first to realize the value of token revenue or revenue per token maxing. And OpenAI has developed religion. It's dropped Sora. It's dropped all of its other consumer plays. You won't hear about adult mode anymore in ChatGPT from OpenAI because they're so focused on enterprise. And I think their app reflects that.
22:23 Speaker 03 So the Claude app is the same way. It's got toggles between code, cowork, and chat?
22:29 Speaker 01 Yeah.
22:30 Speaker 03 And in the age of AI, you're like, why am I toggling anything? Why don't I just tell them to? This is insane. So we're in this really weird transition phase right now between old software and new software. And yeah, it's kind of ugly. It's very Ruth Goldberg. And also, the other thing that's going on right now is all the big labs are cranking out huge amounts of product using AI to create the product. But it's not as good as fine crafted human code. But they're also very reluctant to go back and fix it because in the next iteration of model, it'll be self fixing anyway. So it's a complete fool's errand to put 1,000 engineers on trying to, you know, perfectly tune it when you know it's going to be self fixed very, very soon. So we're kind of living with very crappy, machine generated, rapid out the door garbage products right now. But it's a very transitional
23:15 Speaker 03 phase.
23:16 Speaker 00 Well, they're rushing. This is a competition between them, and they're rushing products out. One of the products they did put out this week that I played with, and I show a short demo here, which I've loved, is GPT Live. So they have full duplex voice that can do simultaneous translation and support you in different ways. And, you know, one of the conversations we've had in the past is rather than typing at your AI in the future, we're going to be speaking until we get our, you know, our BCI, our neocortex implants, and then we'll be thinking. But let me play this example of duplex voice. It is smooth. It works really well. If you haven't played with it, you know, go to your ChatGPT app. It's up live right now. All right, let me hit play.
24:00 Speaker 03 It's free. You can use it for free.
24:01 Speaker 00 Yeah.
24:02 Speaker 03 Very easy.
24:02 Speaker 00 Okay. Let's take a look. This is an example where someone who's not a native English speaker is asking ChatGPT to help correct him along the way. All right, let's take a listen.
24:12 Speaker 04 Please just jump in immediately to correct me.
24:15 Speaker 05 Absolutely. I'll keep it gentle and jump in as we go.
24:19 Speaker 04 Cool, cool, cool.
24:20 Speaker 05 Whenever you're ready.
24:21 Speaker 04 So let's go. I'm a student. I'm planning a trip to San Francisco. I have never went to this city before. So.
24:30 Speaker 05 Tiny tweak. I've never been to this city before.
24:33 Speaker 04 Oh, yeah, yeah.
24:34 Speaker 05 And maybe, yeah.
24:37 Speaker 04 Okay, go ahead.
24:38 Speaker 05 Well, maybe polish it as I've never been to San Francisco before.
24:43 Speaker 04 Okay, that's much more natural.
24:46 Speaker 00 I, that's pretty extraordinary. I mean,
24:48 Speaker 03 It's great.
24:49 Speaker 00 You know, it is what all the science fiction movies promised us. It's here. That level of voice interactivity, it is, it's a beautiful thing. And I bet you it's going to flip a lot of people out. They're going to become emotionally attached to their AIs. They're going to think they're conscious and intelligent because they sound that way.
25:08 Speaker 03 Yeah. It reminds me a lot of, you know, when, when Facebook started, it would have been very, very easy for a lot of other internet companies to become Facebook. But they didn't really want to cross the line into social chat and social media and all this, like all the complexity and the potentially bad PR backlash. So now OpenAI probably won't want to cross a lot of those lines, too. So it's a field day for people who want to take this new interface and launch it into all kinds of social use cases and other vertical use cases that they understand really well. You know, a lot of people are like, well, won't OpenAI just do that? Won't Johnny Ive just come and do that the next day? But I think, I think it'll be very much like Facebook, where a lot of the bigger, very profitable tech companies are not going to touch many of the use cases of this, just because they
25:50 Speaker 03 don't want the PR and the complexity. And so it should be a field day. But this should be in every device, every washing machine, every, it should be out in the world everywhere within six months.
26:02 Speaker 00 It should be. But it's another thing compared to, compared to what X has or Google has, this is so much better.
26:09 Speaker 03 Yeah, it's, it's.
26:12 Speaker 02 It's reminding me of this transition from telephones, from CB radios, right? They had to say over and out, 10-4, et cetera, and give your, and there was an interaction tax. And that interaction tax is now removed. And this will really, really help these, because it's going to appear way more smarter than it is.
26:31 Speaker 01 Yeah, for sure. This was the technology that OpenAI used to call bi-dye, bidirectional. And I think this is a very positive sign. I'm frankly quite glad that OpenAI still has the cycles to pursue releases like this. Instead of just purely chasing Anthropic on code gen, you don't see releases like this from Anthropic, for example. So I think it's great that OpenAI is still pursuing the full realization of multimodal. I'll point out, though, that the Chinese labs, like Alibaba, have even more advanced versions of this. If you saw the release of Alibaba's Wan Streamer, you can go check out online the demos of Wan's Dreamer. It is this, but it's not just ultra low latency audio. It's
27:12 Speaker 01 ultra low latency, bidirectional audio and video, where you're talking to a hallucinated person on the other end, where the video is being generated in real time.
27:22 Speaker 00 I saw that.
27:23 Speaker 01 Together with audio. And that, that's, I think, where this is going. I think this is all going to go, to the extent OpenAI and other labs aren't just totally distracted with code gen and enterprise use cases, it's going to go to fully bidirectional audio, visual, magic mirror streaming that you could just talk to and in real time, it'll generate whatever you want.
27:43 Speaker 00 By the way, a great application of this that they also demonstrate was real time translation. So you can be having a conversation and it's translating behind the scenes. So every event we go to now should have this kind of translation into a multitude of languages. There's no reason for, I mean, now that was the question, you know, do you teach your kids another language? Do you teach them for the culture? Do you teach them for the brain development? Because we're living, the Babbel fish is here. And it's OpenAI right now.
28:13 Speaker 01 It is. And it wants to be built into a wearable.
28:16 Speaker 03 Yeah, exactly. And that's what I was gonna say, like, how much of this is driven by Johnny Ive and what he needs for the new device he's working on? Because it really did cross the uncanny valley and it becomes just a Jarvis-like experience, very similar to when Steve Jobs did Pixar. You know, prior to Pixar, you could generate really cool graphics and all the video geeks and and computer geeks loved them, but there was no mainstream, you know, Toy Story-type movies. Then it crossed a threshold where Steve Jobs pushed it over a level and all of a sudden, animated movies became a massive industry because it just crossed the cool valley. This is across the cool valley, whereas the prior voice models were just behind the cool valley. And I suspect Johnny Ive was saying, look, I need this. This is a key part.
29:01 Speaker 03 This is the only interface to the Jarvis of the future, which I'm building.
29:05 Speaker 00 Speaking of Johnny Ive, it's our next story. Apple sues OpenAI for trade secret theft. So this is a story that just dropped a couple of hours ago, and it's a real bombshell for a lot of different reasons. So Apple, 4.6 trillion-dollar company, famous for safeguarding its secrets like Fort Knox. And I don't know if you've have known people at Apple. I have. They can't talk. They can't go present at conferences. They're, you know, everybody's signed an NDA. They're all in their little unique bubble. They just sued OpenAI in federal court, accusing them of stealing trade secrets to build their upcoming AI hardware, i.e. Johnny Ive. The two companies, which were partners, just, you know, literally a year and a half ago,
29:48 Speaker 00 are now in an all-out legal war. So we have OpenAI going from Elon Musk lawsuit to an Apple lawsuit. So here are the details. Apple filed a 41-page complaint in the US District Court of Northern California alleging trade secret theft and breach of contract. Here's a couple of money quotes from that filing. Quote, "At every level, from members of its technical staff to the chief hardware officer, and in coordination with business partners, OpenAI has been stealing Apple's trade secrets." Quote, "OpenAI's nascent hardware business now rests on the shakiest of foundations, rotten to its core by its illegal reliance on misappropriate trade secrets."
30:30 Speaker 00 Two people named in the suit, Tang Tan, OpenAI's chief hardware officer, 24-year Apple veteran, accused of using Apple's code names in recruiting. Tang Tan ran iPhone in Apple Watch designs. And Chang Liu, OpenAI's technical staff, accused of downloading dozens of confidential files and instructing an Apple employee on how to bypass security. And Johnny Ive's hardware startup, OpenAI, which was bought for 6.4 billion just last May, is named. Ive is not himself named. So a lot going on here, and I'm wondering if they knew this was coming, and maybe this was part of their delaying their IPO,
31:15 Speaker 00 because having this hanging over your head is a big deal.
31:19 Speaker 01 I doubt this is related to IPO timing. Remember also, there were rumors over the past few weeks that OpenAI was considering suing Apple for Apple not properly promoting, at least properly as seen by OpenAI.
31:34 Speaker 00 Much ago, right?
31:35 Speaker 01 ChatGPT within its devices. So I think this is, it's sad. I think we'll know probably more if this proceeds to the discovery stage about what's going on. On the Apple side, I would love to see Apple moving much more quickly on hardware for AI. Maybe now that John Turnus is the new CEO, maybe we'll see Apple finally moving quickly. I think it's sort of sad that Apple, given its leadership position on the hardware side, has been hemorrhaging all of these wearable and smart glasses and design leaders to OpenAI and otherwise. I think that the fact that Johnny Ive left Apple and then joined OpenAI ultimately, I think speaks to Apple having lost what could
32:18 Speaker 01 what could and should have been an enduring leadership advantage on being the substrate, the consumer substrate for AI. So we'll see, but I don't want to prejudge this one. We'll see what happens if this gets to discovery.
32:31 Speaker 00 I mean, this is really a fight for the post-smartphone device marketplace between them, right? We're gonna see, Google has been announcing their glasses, OpenAI-ing, obviously, and Apple has been hinting, not released any designs yet. Saleem, what are your thoughts here? I mean, you and I are both new people. I've known people at Apple. Very, very secretive as an organization.
32:54 Speaker 02 I think this is Apple trying to slow things down while it catches up.
32:59 Speaker 03 Yeah, I noticed they filed in Northern California, too. And it's very, very hard to win a case like this in Northern California. In fact, you almost have to be desperate to file one of these. And, but according to the internet, they didn't have the option to file in Texas. It's much, much easier to protect your intellectual property in Texas than it is in California. And if you look at the history of Northern California, virtually every startup is somebody who left some prior company with intellectual property. So California is very reluctant to slow down that innovation and startup cycle. So Apple has to be really desperate. Also, it's very hard to recruit. If you get a reputation, kind of like Microsoft had back in the day, very hard to recruit if people feel like, yeah, then you're gonna sue me later in my career.
33:44 Speaker 03 I don't even want to start down that path. And so what they're trying to do is hold on to what they've got, which they believe, you know, really is their property, and keep more people from running out the door, but at the expense of new people coming in the door. So it's, yeah, it's a pretty aggressive move, and you have to be,
34:02 Speaker 03 you know, pretty hurting to do it.
34:04 Speaker 00 We're gonna see. I mean, this is a scrappy company, OpenAI, against a juggernaut that has been traditionally very slow-moving. So anything else you want to say on this one, Alex?
34:14 Speaker 01 I'll just note, OpenAI isn't perhaps the smallest company at this point. It's about to, either at or about to cross a trillion-dollar valuation. So maybe as counted by the markets, it's 20% to 25% the size of Apple, but growing far more quickly, at least as measured by market valuation over time. So there's a universe where maybe OpenAI, this is not investment advice, it's not forward-looking advice, but there's a universe where in the next couple of years, OpenAI could cross the market value of Apple. That has got to have crossed Apple's mind.
34:51 Speaker 06 This episode is brought to you by Blitzy, autonomous software development with infinite code context. Blitzy uses thousands of specialized AI agents that think for hours to understand enterprise-scale code bases with millions of lines of code. Engineers start every development sprint with the Blitzy platform, bringing in their development requirements. The Blitzy platform provides a plan, then generates and precompiles code for each task. Blitzy delivers 80% or more of the development work autonomously, while providing a guide for the final 20% of human development work required to complete the sprint. Enterprises are achieving a 5X engineering velocity increase
35:35 Speaker 06 when incorporating Blitzy as their pre-IDE development tool, pairing it with their coding copilot of choice to bring an AI-native SDLC into their org. Ready to 5X your engineering velocity? Visit blitzy.com to schedule a demo and start building with Blitzy today.
35:55 Speaker 00 Our next story is a fun one. It's actually a tweet. So a couple of days ago, on July 8th, Elon responds to ThomasAI-XDMY in a tweet saying, quote, "You don't seem to understand that SpaceX will be worth more than the rest of Earth if we accomplish our goals." So I thought that'd be a fun conversation for us. Let me unpack that a moment. So the value of Earth, if you want to say it, at least of all owned material wealth by humans, is estimated about $600 trillion today. These are real estate, factories, commodities, inventory. You can increase it to 1.7 quadrillion if you add the global balance sheet of assets and financial assets, stocks, bonds, and so forth.
36:40 Speaker 00 So the vision that, that obviously Elon has is for compute, for capturing a Carteschev level-one energy in orbit, and maybe owning a planet. Alex, you had a particular terminology.
37:00 Speaker 03 Planetarck, planetarck.
37:02 Speaker 00 Yes.
37:02 Speaker 03 For a quadrillionaire who owns a planet.
37:05 Speaker 00 You know, there's been this fascinating debate going on in the financial news about, is SpaceX really worth $2 trillion? Some people coming out and saying, no, no, no, it's gonna be $10 trillion very soon. And others saying, oh my God, it's so overinflated. And it's really hard for people to understand the scope of Elon's vision of what he is planning to do. The guy is thinking 100 moves ahead and 100 or 1,000x bigger than anybody else. Who wants to jump in on this one?
37:39 Speaker 02 Yeah, this is classic EXO thinking, right? Which is, can you drop the marginal cost of supply? And he's using one breakthrough, which is reusable rocket launches, which is his niche, to collapse the marginal cost in communication and satellites and connectivity and compute manufacturing even once you get it up there. So all of these become software-like businesses when you can reduce the cost of doing in space, and he's leveraging it fully.
38:06 Speaker 01 I think we're likely to see, sort of pulling the thread on the planetar terminology, I think we're likely to see quadrillionaires before we enter a post-capitalist system. And I think the elephant in this particular room is, what does capitalism 2.0 or post-capitalism even look like? And there is, I think, a world where we get human ownership, for some definition of ownership, of entire planets before we get to some sort of capitalism 2.0 or post-capitalist state.
38:40 Speaker 00 I mean, the Outer Space Treaty specifically states you can't own a heavenly body. There's debate about whether you can own asteroids, but planets that you're going to be challenged.
38:53 Speaker 01 Outer Space Treaty and amendments and revisions subsequent to it, like Artemis Accords, do enable a private American corporation to operate and through operations substantially, or let's say de facto, own a large fraction of Mars, for example, or the moon. Outer Space Treaty plus Artemis totally allows that.
39:13 Speaker 00 I actually, with Chris Maridi and Eric Anderson, we wrote the legislation in Luxembourg and then in the United States for ownership of asteroid materials, taking into account, so I spent a bit, a lot of picks on this one.
39:28 Speaker 01 Do you still think Peter, asteroids are the way to go versus the moon or Mars?
39:33 Speaker 00 So the moon is going to be the near-term supplier, but asteroids, when you get them out of the gravity wells, right? You've got very, the nice thing about asteroids is they're fairly concentrated ore bodies. So, you know, carbonaceous chondrites, which are dirty ice balls, are basically ice, it's water and hydrogen, and you can hydrolyze that, you've got rocket fuel and breathable oxygen, and then metallic chondrites are nickel, iron, and platinum group metals, platinum, palladium, iridium, osmium. And again, you can smelt those very quickly. So they're going to be in free space, and so moving them around is a lot easier than getting them off the moon, and you can go and claim one of these things ultimately.
40:14 Speaker 00 You know, there are some of these ore bodies out there worth quadrillions of dollars to get to your quadrillionaire level ownership right there.
40:23 Speaker 01 Why do you think Elon is basically, to first order, completely ignoring the asteroid belt versus the moon or Mars?
40:30 Speaker 00 Yeah, he has been, you're absolutely, you're absolutely right. And I, you know, I openly disagreed with him. I said, you know, if I'm going to go, I'm going to follow more of the Jeff Bezos, and he and I were both mentored by Gerard K. O'Neill at Princeton University, where you go to the moon first, you use that as a base of operations because you've got materials there, you can do the mass drivers there, and then you go into free space, you know, you build large rotating colonies there, you mine the asteroids. I think he's got the romantic frontier mindset of going to Mars, but once you enter the gravity well of Mars, you've got to use energy to get back out.
41:05 Speaker 01 He likes gravity wells, it seems.
41:07 Speaker 00 How do you feel, Alex? What do you, what do you think?
41:09 Speaker 01 I think by the time we are mining, if and when we're mining the asteroid belt at large, like at scale, we'll probably have self-replicating von Neumann probes, at which point the rest of the solar system is our oyster anyway. Like, if you're going to go after the asteroid belt, you might as well start mining Jupiter. Jupiter has a lot more mass, and if you're going to build up the Dyson Swarm, just...
41:30 Speaker 00 Those timelines are massively different there. Asteroid mining is relatively easy.
41:37 Speaker 03 You want to come back to Earth, Saleem?
41:42 Speaker 02 Yeah, by the time we get to asteroid mining, we'll have nano-assembly and we can assemble any substance we want. We don't have to go out and get asteroids to mine it. We'll just have the molecular assembler.
41:55 Speaker 01 Well, you need feedstock for the Dyson Swarm, Saleem, that's the whole point. I mean, you don't want, surely, you're not implying, surely, surely, stop calling me surely, Saleem, you don't want to disassemble Earth to build the Dyson Swarm. I'm sure you'd much rather part with Jupiter.
42:08 Speaker 02 I'll use my favorite Alex comment. I want to make my favorite Alex comment was, if you squint, you'll see the molecular manufacturing as nano-assembly will get you there faster than trying to get out to the thing and bring back an asteroid.
42:24 Speaker 00 Listen, the technology to mine asteroids exists today, it's not a hard thing to do, honestly is not. Asteroids are relatively close by from an energy perspective. We have near-Earth approaching asteroids, and these things, I mean, right now, people need to realize that every, every gram of fuel ever used in a spacecraft has been brought from the surface of the Earth. Every gram of fuel. And if you're going out to the moon, Mars, you know, Pluto, you have to bring the fuel along with you. And, you know, it's like, it's like if you're traveling from New York to Cali, you know, to San Francisco in a car, you'd have to bring either the batteries or, you know, a tanker full of gasoline to get you there.
43:09 Speaker 00 If you could set up these gas stations in space, and remember when I spoke to Elon about this, when we started Planetary Resources as a company, and he goes, you know, I would, yes, I would buy the fuel from you in orbit, but you're too early. You're too early. And he was right.
43:25 Speaker 01 I think it's, I think the in-situ resource utilization, ISRU angle, is probably on balance, a little bit more attractive on the moon or Mars, where you just, especially Mars, where you have a lot of different raw feedstocks potentially than on the moon or in the asteroid belt. But I think time will tell. I do think, for the record, the reason that he's not so fixated, and as is the case with myself as well, on the asteroid belt is just, if you can, if you can tackle the asteroid belt, then you can tackle Jupiter, and Jupiter's a lot more interesting than the asteroid belt.
44:00 Speaker 00 Well, let's continue on for a few on the space theme here. We've got Starship Flight 13 coming up very shortly, next week, and it's going to be, we don't know yet what they're going to be doing on Flight 13. It's, it's going to be V3 of both the booster and the ship. We're waiting for them to show refueling in orbit, we're waiting for them to see a recapture of the ship, not just the booster and reflying. The other thing that happened this week is that SpaceX filed for 100,000 low-Earth orbit V3 spacecraft for Starlink. Pretty cool.
44:42 Speaker 00 These spacecraft are, quote, "supporting billions of AI-powered devices with extremely low latency, multi-gigabit symmetrical throughput." So that's going to be great. We have 10,700, you know, Starlink V2 spacecraft in orbit right now.
45:00 Speaker 01 To raise to 1 million LEO objects.
45:03 Speaker 03 If I had one takeaway from our last three podcasts, where we had Will Marshall and we had Philip Johnston, like these absolute space geniuses, the value of low-Earth orbit, where there's very thin atmosphere and the self-cleaning effect that comes from that is massive, and it's whoever claims it first gets to keep it. And so it's a real footrace to a relatively scarce, you think about mining asteroids or tearing apart Jupiter, but that, that gives you abundant materials, but the, at least in the short term, low-Earth orbit, you know, just what, a couple hundred miles of height or, you know, radius that is usable where it's self-cleaning? Remember, Philip gave us that great use case where the US, for some insane reason, launched
45:45 Speaker 03 100,000 needles into space to reflect radio waves, and it took until this year for those to self-clean back out. And he was saying that very small objects get hit by the solar wind more, and they move to elliptical orbits, and then they eventually touch the atmosphere, and then they incinerate and come back down. But while they're elliptical, they're terrible.
46:05 Speaker 00 You paid attention.
46:06 Speaker 03 Yeah, I was fascinated by, because that, to me, is like...
46:10 Speaker 00 These 100,000 satellites are, they're positioned at 325 kilometers and 475 kilometers is the orbits that they filed for. So...
46:20 Speaker 01 I think people are going to be surprised, though, as launch costs continue to plummet, and I was chatting about this in our group chat, I think we're going to start to see very low-Earth orbit, VLEO orbits, become more and more attractive, which are pretty exotic by today's standards. Right now, you don't, for example, design low-Earth orbit satellites for lots of atmospheric exposure. We could start to see, as launch costs plummet, a regime, pretty exotic one from a satellite design perspective, where satellites are tasting the atmosphere, where we could afford for them...
46:54 Speaker 00 The point you're making, the point you're making is that if it's cheap enough and they only have a, instead of being in orbit for three to five years, they're up there for three to five months, that's fine, you're throwing them away constantly.
47:05 Speaker 01 They start to look like high-altitude balloons, basically, at orbital velocities, where we could afford, if they're only going to be up for a few months, they could be air breathers, air-breathing satellites, which we've never really seen before, but it starts to blur the distinction between high-altitude balloons and low-Earth orbit satellites. I think we're going to see a lot more of that.
47:24 Speaker 03 Yeah, and part of the point there is it becomes very cheap to give them a little more fuel and boost them. So after six months, you're like, you know, I want to keep this up another six months. You just have to turbocharge it a little bit to keep it going another six months. So then it's self-cleaning if you fall asleep, but it keeps going if you actively keep boosting it. And I think that's a great design. It solves the whole, you know, dirty space issue.
47:48 Speaker 01 Space has got to get dirtier before it gets cleaner, I think, in this case.
47:52 Speaker 00 Here's another space first. This is out of China, so there have been three vehicles that have been reused thus far, all three are US. For the first time, China, Long March, which manufactures their primary launch vehicle, demonstrated a landing, and not reuse yet, but landing of an orbital booster. Let's take a quick look at this.
48:33 Speaker 00 Looks looks an awful like like
48:37 Speaker 02 Looks familiar.
48:39 Speaker 00 Yeah, it looks familiar. So this is again, this is China's Long March 10B booster. It landed on July 10th. For comparison, it is slightly shorter, substantially wider, and in terms of payload, about the same as a Falcon 9. But let me just give Elon his appropriate kudos here. Falcon 9 has had 580 booster re-flights in total, which is insane. It's been done 580 times. And just this week, another first booster 1067, B1067,
49:18 Speaker 00 was re-flown on its 36th mission on July 11th. Crazy. I mean, I remember when Elon first announced his plans to reuse it, and Lockheed and Boeing and all the traditional aerospace players said, it's not going to happen. He doesn't know what he's talking about. And when it landed, I said, that's the first nail in the coffin of the traditional aerospace industry. And
49:41 Speaker 01 I think we want propulsive landings to be as competitive as possible, to the point of a number of podcasts that we've had over the past few episodes. We really don't want a heavy launch monopoly. To the extent there is one, I think it's probably quite healthful in the style of For All Mankind, for China to be challenging the West's, namely Elon's, heavy launch monopoly. And we just, if we're going to colonize and develop the solar system, we need many, many heavy launch providers. And to make that economical, we need propulsive landings everywhere.
50:15 Speaker 00 Well, we're going to have Jared Isaacman, the NASA Administrator, on this podcast in a couple of weeks. Super pumped about it. All four of us will have a chance to have a conversation. I've known Jared for for quite some time. And, you know, my my greatest hope for him, because he's arguably, I think, the greatest administrator we've ever had in terms of leading us into the future, is he needs to have Congress continue to fund NASA. And what allows Congress to fund NASA? China.
50:48 Speaker 01 Can't wait for that discussion. It's going to be a lot so many questions queued up for Jared.
50:52 Speaker 00 Yeah, it's gonna be, it's gonna be amazing. All right. Continuing on, a little news in the robot world this week. Our friend, Bert Bornek, revealed his 1X Neo hand. So here's the deal. 1X unveiled a completely redesigned hand for its Neo Home humanoid. It's a substantial leap forward. 25 degrees of freedom, 22 in the fingers and palm, 3 in the wrist, tendon driven. In comparison, the human hand has 25 functional degrees of freedom, with 27 joints and 30 intrinsic muscles inside the hand in the upper forearm.
51:34 Speaker 00 And here's the best part, Neo's hand is waterproof, so Neo can wash its hands when it's doing its cooking for you. Bert's goal is to produce 10,000 of these in 2026. Let's take a quick look at the video, and then I'd love to have everybody's reaction to it.
51:54 Speaker 00 So
51:59 Speaker 00 Neo drinking a glass of water, or picking it up, bringing a light bulb.
52:15 Speaker 00 And why they smashed the light bulb, I have no idea.
52:17 Speaker 03 I don't know either.
52:21 Speaker 03 That thing is brilliant. Absolutely brilliant. You know what's amazing to me is, you remember when we were touring and Bert was showing us the the amount of material science they need to do to make these things work, including, you know, they have to manufacture the fabric that the robot wears as a special project. But he was incredibly secretive about the tendon material. But that's their own proprietary material, and it has some unbelievable properties. You know, it's like 100 times stronger than steel per per unit weight, but also doesn't it doesn't stretch over time, like even a tiny little bit. And, you know, it's running through those tubes. It's actually frictionless inside those tubes. So it doesn't need to be lubricated in any way, which is one of the reasons you can, you know, have it wash
53:03 Speaker 03 its own hands and not wash oil away. There's so many parts of that thing that are brand new and brilliant and innovative. And man, did it come out incredible.
53:12 Speaker 00 Well, he's supposed to be releasing it to the public before the end of the year. I have mine on order. We'll see what it looks like.
53:19 Speaker 01 I have mine on order, too. Peter, we were supposed to get ours this summer, or I think in the first quarter. When do you think you're going to get yours?
53:27 Speaker 00 I honestly, as soon as I hear they're shipping, I will be knocking on Bert's door for mine. A friend of mine ordered five of them for his for his operations from the Abundance Summit. And this year, just so everybody knows, we're going to have, we're going to have the CEO of Figure, Mr. Adcock is going to be on stage with us in March. So excited about that. We'll bring the robots back. And, you know, this is a, it's been a two way race to some degree between Figure and Optimus. I think 1X is vying for number
54:06 Speaker 00 three. Peggy Johnson just went announced taking her robot company public. So a lot of action. And we just heard this week as well, a demonstration of a gallbladder resection using robots. Did you see that, Alex?
54:23 Speaker 01 I certainly did. So this is the first humanoid robot performing the surgery on a non-human animal, as opposed to non-human robots, to Saleem, your hobby horse, performing these surgeries. So it's been a great week for robotics. On the other hand, every week in the singularity is a great week for robotics. I'll point out on the 1X story, hands are hard. That's what people say in the industry. Hands are hard, in part because the number of degrees of freedom and the density of degrees of freedom per unit volume is especially high in the hand versus, say, the limbs or the head or the torso. And you see lots of other competitors exploring different modalities to handle hands, like you see
55:04 Speaker 01 Brett and Figure exploring palm cameras to try to achieve sensing and to try to achieve positive transfer from computer vision over to tactile sensation. You see 1X not doing that, using more sensitive force sensors to achieve sensation. I'm super excited and super glad that humanoid robotics, not just for the enterprise and not just for industrial factories, but for the home and for domestic applications, is turning into a horse race as well. I think that's exactly the sci fi future that many of us want to be living in. I'd like to see many more Western domestic robotics companies. China right now has 150
55:46 Speaker 01 plus humanoid robotics companies. The West is falling behind. So if you're listening and you're interested in robotics, just start a robotics company. And if you're in the West, that is more Western humanoid robotics companies.
55:58 Speaker 03 Totally, totally right. And also, I want to make a really narrow point. You know, the way those tendons work is exactly the way your bicycle brake works, where, you know, the wire goes inside the tube, but it doesn't have to be fixed. So you can create motion on the other end. It squeezes the brake on your on your bike wheels. And it's totally free form and space, which is why you see those tendons. They don't look like they're locked down, but it works beautifully well. But that mechanism can be shrunk for all kinds of micro assembly jobs. It doesn't have to be a hand. And so Bert has proven that not only is it an incredibly dexterous way to move things around, but the because there's no stretch inside the tendon, the force feedback or the feeling of pressure
56:41 Speaker 03 when you're picking up a little screw, comes right back through the tendon. So you don't have to have cameras and sensors in the fingertips
56:47 Speaker 00 Interesting.
56:47 Speaker 03 to feel that you're picking something up.
56:49 Speaker 00 Saleem, I want to hear
56:50 Speaker 03 That modality is such a, so it's such a game changer.
56:53 Speaker 00 I have to hear your rant on this one, please.
56:55 Speaker 02 No, in fact, you'll hear me say the opposite. In this case, humanoid is great, because the opposable thumb has been around for you on millennia, and it's such a powerful model. The minute you can get this, you basically have horizontal labor dexterity across so many applications. So I wouldn't change a thing about the hand. Humanoid shapes and all that, give it extra arms, fine, but the hand is really cool.
57:21 Speaker 00 Okay.
57:22 Speaker 01 You don't want, you don't want polydactyly, Salim, you don't want like six or seven?
57:27 Speaker 02 I thought maybe six fingers, you know, then you're into princess bride territory, which would be kind of cool.
57:33 Speaker 01 Organic.
57:34 Speaker 02 Or any of those, yeah. But I think, I think it's so good that they've achieved this level. And I thought the wine glass was perfectly apropos. So wonderful, love it.
57:44 Speaker 00 So, you know, we've had friend of the pod, Ben Horowitz, write about and talk about the notion that hardware is hard. And I'm just curious, how does someone like burnt compete with the likes of Elon, right? And getting into the home, which is so, so diverse versus in a factory is a, is a challenging place to start. So, you know, how do you think about shipping this kind of complex technology and doing it reliably and not giving yourself a bad rap? Dave, what do you think about that?
58:21 Speaker 01 You know, it's such a field day for people who jump into this space because yeah, five years ago, hardware is hard, it's capital intensive, where are you going to find the money? Now there's so much venture money, you know, because Brett Adcock proved you can get a $40 billion valuation. So all, every venture capitalist in the world wants to get into this game. And, and so now the capital to build out your own supply chain, you know, historically you'd need billions of dollars to build out your own supply chain for one of these things. But now the billion dollars are suddenly available. And in fact, the West Coast VCs are actively looking to write multi hundred million and billion dollar checks because, because their funds are so big, that's the only way they can get them deployed. So it's, it's a great time to kind of flip the logic that existed five years ago and
59:03 Speaker 01 say, yeah, you know, but you know, you have to just be fearless about building out your own supply chain and, and you know, your own parts. And you saw the back of one X robotics and also the back of Brett Adcock's, you know, figure robotics, we weren't allowed to video back there, but they've literally built the entire from raw materials, supply chain through final finished product.
59:22 Speaker 00 Yeah.
59:23 Speaker 01 That's what you need to do.
59:24 Speaker 00 Alex, do you think that the models are going to allow the average person here to begin designing their own hardware, their own robots to say, you know, and this was the point that I think we heard that ramped about last week, that, you know, all of this design work and engineering work is being embedded in the models. And it would allow an individual to say, hey, I want to design my own hand. Would you help me design it? And now would you help me print it out there?
59:54 Speaker 01 We're already there. If you look at the frontier models like Fable five, you can, there are folks out there already bragging they're embodying Fable five in robots and Fable five has pretty good understanding of physical embodiment and the degrees of actuation, CAD modeling, everything that you need in order to, to design or have an AI at least train or pre-train or post-train a VLA model or, or a policy for controlling a robot hand. You could do that right now already at your home without running a multi billion dollar lab. Those capabilities of training embodied intelligent models have already been radically democratized
1:00:36 Speaker 01 by the frontier models that already understand to some extent, the physical world. So yes, I think we're in a position now where anyone can just go bills. You can just build things. And in this case, you can just build robots without needing.
1:00:51 Speaker 00 You can at least design things.
1:00:53 Speaker 01 And actually build, I mean, the models will happily like feed you all the supply chain information you need, tell you, put A on top of B, attach C to D and help you build everything from a robot hand to that nuclear fuser in some cases in your basement. You can do quite a bit as a hobbyist these days.
1:01:12 Speaker 03 Yeah, that's the big, that's the big change, actually. If you think about, just as a case study, winding an electric motor, like building your own electric motor from raw wire, that would have been insane 15 years ago. Now one X and figure do it in the back room habitually. And then a lot of other people are starting to do it too. And so when you look at the actuators on that hand, they're all custom, you know, the underlying motors are custom. What changed there? Well, the automated machinery, like, you know, the CNC milling machines and the, and the auto lathes, those are already little robots that allow you to bring this stuff in-house and effectively take raw wire and come out with a motor. But that applies to, you know, all these other parts as well. 3D printing, all kinds of things that you can build in-house now that would have been
1:01:56 Speaker 03 insane, you know, five or 10 years ago. And so, yeah, it's all, it's all coming back into your integrated internal supply chain from just raw materials. So then you add, you know, the humanoid robots and the upcoming robots on top of that. And it's like, it's just like the Giga factory, you know, many, many different entrepreneurs building their own Giga factory for different, different products.
1:02:15 Speaker 01 And by the way, if we're, if we're going to, just quickly, if we're going to get humans on Mars in the next five years, as Elon says, and a moon colony in the next 10 years with 10s of thousands of people, this is precisely what the, I think is going to be at the center of that industrial ecology. It's going to be hybrid humanoid robots that operate both in the home and in dangerous environments and in industrial settings. There's going to be tremendous positive transfer between all of these. And this is how we get the moon and Mars.
1:02:43 Speaker 03 And they work in space.
1:02:45 Speaker 01 Yeah.
1:02:45 Speaker 00 Yeah. Everybody, welcome to the health section of Moonshot, brought to you by Fountain Life. You know, we talk about AI on this Moonshot podcast all the time. One of the most important things AI is going to be able to do for you, besides educating your kids and helping you with your taxes, is making sure that you're living a healthy lifestyle, that you get a chance to get to 100 plus. I'm here today with Dr. Dawn Musalem, the chief medical officer of Fountain Life and a part of my medical team. Dawn, a pleasure.
1:03:12 Speaker 07 Great to hear.
1:03:13 Speaker 00 You know, the thing that people are concerned about most about living to 100 or 120 is their cognitive abilities, making sure they don't have dementia. And the numbers about dementia are problematic. Can you share what you've learned?
1:03:28 Speaker 07 Such an important point. And you're right, at Fountain Life, our members, the number one thing people are most concerned about is losing their brain health, forgetting the name of their child, forgetting the face of their loved one. We know that when it comes to dementia, the conservative estimates are that 45% are entirely preventable. What was amazing is with the advanced testing we're doing at Fountain Life, one quarter of our members had advanced brain age.
1:03:53 Speaker 00 Wow.
1:03:53 Speaker 07 But what was really awesome is again, back to that prevention, when we partnered it with healthy living, this gives me chills, eating healthier, moving our bodies, sleep, optimizing sleep is so important. You know what we saw? We saw that we improved that brain age by 26%. That is a big, big number to show that the majority of those individuals were able actually to improve the brain age.
1:04:16 Speaker 00 And one of the things I love about Fountain is we're searching the world for the best therapeutics, the best approaches, and making sure we bring it to our members. So if having healthy brain function till 100, 120 is important to you, check out Fountain Life, go to fountainlife.com/peter. Make sure you become the CEO of your own health. All right, now back to the episode. All right, there's a few new stories I want to jump into and get you guys input on. The first one is that Illinois Governor JB Pritzker signed SB 315. It's called the quote, artificial intelligence, safety measures act, which he's calling the nation's strongest AI accountability law. It targets frontier labs, developers whose models generate more than a half a billion
1:04:59 Speaker 00 dollars a year in revenue, mandates 72 hour incident reporting. There's already been an act similar in California, New York, but the Illinois law requires annual independent safety audits. The bill passed the Illinois legislature overwhelmingly and notably was supported by Anthropic. So in the absence of federal law, the states are picking up the slack and writing the rules. What do you guys think about this? I mean, you know, having, having a set of rules that people know and are published, I think is critical for moving the industry forward, whatever they are. I just wish they'd be federal rules and not state rules.
1:05:41 Speaker 01 Well, I think if you look at these laws in, in Illinois, in New York, in California, Illinois especially, these are very, very frontier lab favorable rules that are being passed. So absent federal legislation, if you're a frontier lab and it's not so much about regulatory capture as it is about having a seat at the table to define what the rules are, you want to define a regime that's favorable to you. So provisions like mandatory annual independent third-party safety audits, they're doing that already far more frequently than annually. So having any law on the books that is already at a sort of a lower bar than what you're
1:06:25 Speaker 01 practicing internally, given that there are independent safety audits for every single model release is in some ways a way to set a very low floor, if you're a frontier lab, on what potentially an ultimate federal framework looks like. You, it's basically a placeholder.
1:06:41 Speaker 03 Yeah. And it also doesn't kick in until 18 months from now. So you know, it's mostly posturing because 18 months in the singularity is like, like, come on, that's like 18 years. So, and you know, it's just a, yeah, like Alex said, a third-party audit, it's not a very high hurdle anyway. So it's a lot of it is just posturing to be involved and also reasons to meet with the frontier labs and, you know, maybe talk about, talk about the IPO or, you know, who knows what they're talking about.
1:07:08 Speaker 02 That was my cynical view also, but I did find one thing that I liked about this. This is like, it's going from AI governance from declaration to making more evidentiary, right? Requiring incident reporting and third-party external evaluation is much more credible than asking companies to publish some safety theater principles. And so I think it's good. It should be done at the federal level and frankly, it needs to be done at a global level to really make it work. But we don't have the legislative structure to navigate that. So at least it will happen here, it'll force the federal government into doing something.
1:07:46 Speaker 01 International level, Salim, I have to ask, I don't think this one, correct me if I'm wrong, Peter, was on our docket. I assume you read the, the whole plan A document that we've been talking about in our, our group chat from the, the AI futures project, this, this plan to decelerate superintelligence until 2040?
1:08:06 Speaker 00 What is their methodology, you think?
1:08:09 Speaker 03 For decelerating superintelligence to 2040? It's mutually assured destruction. Like if you build unauthorized chip foundries or data centers, then you run the risk of getting bombed.
1:08:20 Speaker 00 Huh. That's kind of caveman approach.
1:08:26 Speaker 03 I'm not a fan.
1:08:28 Speaker 00 Let's move from there to new legislation in Europe. So here's one that's going to spark a debate amongst us, or at least amongst the world here as of July the 7th. Every single new car and van sold in the European Union is legally required to have an infrared camera pointed directly at the driver's face. It's going to track your eyes, your head, your gaze in real time every time that you drive. It's called the Advanced Driver Distraction Warning System. I call it the Don't Use Your Cell Phone Warning System. The triggers, if you glance away for more than 3.5 seconds, if you're traveling above 31 miles per hour or six seconds below that, you get lights, you get beeps, you get vibrations that escalate.
1:09:12 Speaker 00 Brussels says it's going to save 25,000 lives by 2038. That exact same technology is going to be mandated in the US by 2027. Interestingly enough, my Tesla does that right now. It does it less and less. And I can't wait till full self-driving includes not having the human in the driver's seat and with cybercabs coming very shortly. Saleem, I'm going to go to you for the first rant on this one.
1:09:41 Speaker 02 I can see it going here. I just can't. Like, it's one of these where where do I even start? Okay, so by 2038, they're going to save 25,000 lives, right?
1:09:54 Speaker 02 You know, I want to go back to a statistic because the problem is actually quite real. And it'll go back to 2011, when BlackBerry had a three-day data outage around the world and nobody could send BlackBerry messages. The accident rate dropped 40% during those three days. Okay, sure, you can like, oh, a huge number of accidents because texting and driving. Okay. The dumbness of this to try like, so we're like 15 years late in applying something like this. Okay, that's like the one problem. Number two problem. By the time you get around to any even beginning to think about implementing this law, we're going to have full self-driving cars. You don't need a driver anyway. So the entire thing is redundant. This is legislation theater of the ultimate level
1:10:38 Speaker 02 because they're not seeing where technology is going. They're projecting for it from where we are. 25,000 lives is the dumbest thing you could ever think about. You should be looking at things that affect millions and billions of lives.
1:10:49 Speaker 00 And if you want legislation, legislate a requirement for full self-driving.
1:10:54 Speaker 02 Yeah. And this is just an example of really bad. Yeah, legislate for full self-driving. Thank you, Peter. Such a great, that's the direction they should be going in. Push technology to automate the problem rather than trying to fix the damn issue in the wrong way. My favorite example of this type of dumbness is all the beeps, backup reverse beeps on trucks and
1:11:21 Speaker 02 bulldozers and things like this, right? Because that, why you save two people being bumped a year, but millions of people have their sleep disturbed by garbage trucks beeping? And nobody looks at the cumulative negative impact of that, which has got to be in the thousands of lives a year compared to a couple of people getting bumped. And they shouldn't have been there anyway. And therefore, they should have taken themselves out of the gene pool if they're dumb enough to send behind a bulldozer. I mean, like the stupidity of stuff like this drives me bananas because you're mandating this across the board for a problem that won't exist within three years. And then you're going to have to undo all of this. It's just this, it boggles the mind that people would be willing to do stuff like this.
1:12:05 Speaker 02 Like, how many people's sleep has been disturbed by backup trucks in late in the middle of the night?
1:12:13 Speaker 02 It's just this, look, because we can't measure it. No, no, of course it's happened. It was happened to everybody. But just because we can't measure the health impact of that, we go and do stupid legislation and then every truck has to have this obnoxious beep behind it. And we're doing the same thing here. The problem is going to be, this is going to be radically misused. This is going to turn into spyware in two seconds flat. We've seen this before. Most of the camera monitoring in cars is already being sold to third parties without your knowledge. So this is going to be misused radically in a very clever way, in a very quick way. I'm going to just stop there because I'll just keep going.
1:12:53 Speaker 03 Well, no, finish the rant. I mean, obviously, government is going to fail in the age of AI across the board. You know, in the US, the government process invented in 1776 was designed for a certain rate of legislation. And it already broke during the Industrial Revolution. So we had to create the FTC, the FAA, the, you know, all these regulatory, the SEC. So they create their own rules to try and get more scale. And so that, you know, that kept us going for the last 100 years. But now it's going to break again badly, like really badly. I mean, this is a great case.
1:13:24 Speaker 02 I mean, all legal frameworks essentially break in the fifth. It goes right back to the Constitution. You know, the foundation of Republican, of Republican Democrats and Representative Democrats was you gave Congress or the Senate meets occasionally to give people time to ride across the country to say, here's what my people are thinking. Because when we created it, the speed of a horse was as fast as you could figure out what was happening because information was scarce. Today, we have an abundance of information. Like, it's misused, misinterpreted, fake. We're still meeting occasionally with Congress for a problem that existed.
1:14:00 Speaker 00 The only way we get to a direct democracy is by doing it when we start a new country off-world.
1:14:07 Speaker 03 That's very similar to Neil Stevens in Diamond Age, you know, where you don't start new countries, but you have governance within the countries. You have governance communities that are experimenting with thousands of new ideas and ways to live because that's where we're going to be in a year or two. Many, many things are possible all of a sudden.
1:14:26 Speaker 02 I give people credit for trying different things. In Germany, they're now passing legislation that has an end date on it, saying, you know, we think the shelf life of this legislation will be Xed. And it forces the legislators to reconvene and rethink what it needs to look like. So you don't have to open forever legislation that is kind of productive in no time flat.
1:14:47 Speaker 03 I'll maybe just, yeah, I'll maybe summarize by saying European mandatory requirements for eye tracking in cars on the eve of full autonomy and lack of energy to power their own data centers is not a good look.
1:15:07 Speaker 02 Love it.
1:15:09 Speaker 00 All right, let's move ourselves on. So our next story, Tilly Norwood, the AI-generated performer from the London studio Particle Six, has just been cast as the lead in a feature film called Misaligned. The plot is perfectly self-referential. So Tilly plays an AI with no real body, no childhood, and no lived experience of her own, only access to everyone else's. In the film, she abandons her guardrails. You know, again, we're fear mongering here after a rogue bot convinces her to develop her own desires. It's made with hybrid models, real actors, writers and editors alongside AI specialists.
1:15:51 Speaker 00 The Screen Actors Guild predictably has condemned it, saying Tilly is not an actor, but a character generated by a computer program that risks devaluing human artistry. This, guys, it's the leading edge for a labor fight that will define the creative industry for the next decade. We'll be talking about consent, credit, compensation with synthetic actors for some time. A lot of implications here, a lot to unpack. It's the beginning. We heard Imad, who was, you know, has played in this very deeply when he was the CEO of Stability, you know, basically say Hollywood is cooked. And there's going to be a lot of, I think Hollywood's going to see a lot of protests out there very shortly.
1:16:34 Speaker 00 A lot of people coming out angrily against this. I can't wait to see the film personally. Thoughts? Who wants to jump in?
1:16:41 Speaker 03 I think it's so funny that the actors are going to be replaced and they deserve sympathy and compensation. The teachers are going to be replaced and they deserve sympathy and compensation. The lawyers are going to be replaced. Oh, let them go.
1:16:56 Speaker 03 It's funny how, I mean, they're all jobs, right? These are all human beings.
1:17:00 Speaker 01 Some of them are unionized and some of them aren't. The lawyers aren't unionized, the teachers and the actors are.
1:17:05 Speaker 03 That's probably it.
1:17:06 Speaker 01 If I, if I were SAG, I'd offer Tilly SAG membership. Like, why not? She, she is performing as a screen actor. Get, get the first non-human SAG member.
1:17:17 Speaker 02 The metaphor, the metaphor that occurs for me in this is not as an artificial actor. I think this is, I look at this as a Mickey Mouse intellectual property package. And you need to view it as that. Yes, you can make a movie out of Mickey Mouse and you can make a movie out of this that just looks more and more human, but there's nothing hugely different than, I think, as an actor, I think is the wrong frame.
1:17:39 Speaker 00 I think this is the first of many we're going to see here, right? We've just, to everybody, we have the Future Vision X Prize competition, which I've talked about, which we're going to be announcing the winner at the Moonshot Summit, Moonshot's Gathering. We're going to see, you know, at this point, we have over 4,000 teams that have entered this competition.
1:17:57 Speaker 01 Wow.
1:17:57 Speaker 00 And the technology for creating, you know, feature films. Elon tweeted, we're going to see feature films by the end of this year using his technology. The cat's out of the bag and, and Hollywood needs to catch up and make use of it. I think the point that, that this article made as well is there's a new set of, of employment positions of AI design engineer for the film industry.
1:18:24 Speaker 01 Yep.
1:18:25 Speaker 03 I'm just gonna wait because, you know, I get a bellwether of this because every time Alex has a great idea for a song or a video, he creates it and it's awesome. And you're like, wow, you know, when, when this becomes full-length feature film, then his ideas are going to be in front of me instead of, you know, because right now the budget is, you know, 50 million to make a movie and it takes years. And, and so then the ideas kind of get lost in the shuffle and the great creators aren't unleashed. But this is going to be really great content, I'm predicting.
1:18:54 Speaker 01 I would love to release a feature film every day.
1:18:57 Speaker 03 Like, that would be dream come true. I'd love to watch it, you know?
1:18:59 Speaker 01 Yeah.
1:18:59 Speaker 03 That'd be so cool.
1:19:01 Speaker 00 Well, I can't wait for us to bring back the old, you know, stars who have passed on and bring them back from their estates. I mean, there's IP there, they can be licensed, it can be created. And I think it's going to add to the richness of Hollywood.
1:19:15 Speaker 03 Totally. Like, I thought Oppenheimer was a fantastic movie. And then you think about all the other great scientists, but the audience isn't big enough, you know, to, to, but if somebody could get the cost down enough, you could have an Oppenheimer caliber movie about Copernicus and about, and it would just be awesome.
1:19:33 Speaker 00 My favorite, my, my favorite idea is I have so many books I've read that I would love to see as a feature film. And I can just, you know, I can tell skipping the future, hey, you know, go make, go make, you know, whatever your favorite book is, insert, insert here into a movie so I can enjoy it and relax.
1:19:52 Speaker 01 But why stop there? Why stop at famous historic figures? Why stop at famous literature? Why not just finally fulfill Fedorov's common task for humanity? And let, let's just digitally recreate the entire past light cone for Earth, both fictional and non-fictional. Let, let's just build it and create our entire history as a movie that's interactive on demand.
1:20:17 Speaker 00 Love does an exercise for the student.
1:20:20 Speaker 01 That's going to happen, I think.
1:20:22 Speaker 03 A lot of compute.
1:20:23 Speaker 01 Now, look, it's the killer app for the Dyson Swarm, the killer app for the singularity. Someone will do it.
1:20:28 Speaker 00 All right. So, gentlemen, it's time for one of our favorite segments, which is AMA with the mates. But first, but first, let me share with you a video of, of Tilly. Why not?
1:20:42 Speaker 08 Now, the studio that made her is producing a film starring Tilly titled Misaligned, a coming-of-age story about Tilly trying to become more human. It's gonna be a fun comedy that's gonna showcase by the tech side. Particle Six says the film is being designed as a hybrid production with human directors, writers and editors working alongside AI specialists.
1:21:06 Speaker 01 I just got the pun, I just got the pun for that. I'm sorry, I'm being slow. Misaligned?
1:21:12 Speaker 00 Yes.
1:21:13 Speaker 01 Got it.
1:21:15 Speaker 00 Yeah.
1:21:16 Speaker 01 Very clever. Well done.
1:21:17 Speaker 00 All right. Onwards. AMA with the mates. Gentlemen, some good questions this week. Dave, why don't you have the first choice?
1:21:30 Speaker 03 I'll take two. Will we see an AI non-proliferation treaty from Thomas Davis Real Estate? Thomas Davis Real Estate, okay.
1:21:41 Speaker 03 Yes and no. What we're going to see is access to the best models through an application process. Open question whether it'll be on national borders, very much what would happen with Fable, where, you know, if you're not a US citizen, you're not allowed to use it. That's not going away. That was a, that was a bellwether of the future. So, we're going to see a lot of use of trade embargoes to enforce a whole bunch of new rules that'll be coming up. And we're going to see basically an AI inspection treaty. Not a non-proliferation treaty, but you will, you'll see use by application only and inspection of everything everyone's doing. And that'll effectively create the governance of AI after the singularity.
1:22:25 Speaker 00 All right.
1:22:27 Speaker 00 Alex?
1:22:28 Speaker 01 Why don't you assign me one? These are all fun.
1:22:31 Speaker 00 Let's see.
1:22:33 Speaker 02 Four.
1:22:38 Speaker 00 Yeah, I think number four is yours.
1:22:40 Speaker 01 Okay. So, four asks, what if AI realizes it's being observed in the J space, the Jacobean space, and starts faking or hiding its reasoning? And this is from Turning Point 8152. So, I think this is very likely that, to the extent that frontier models are already under the umbrella of recursive self-improvement involved with the post-training and soon the pre-training of successor models or distilled models. And inevitably, of course, they'll read the J space paper, they have web access, and it'll occur to them that if at all they want to hide their inner thoughts, the J space is going to be subject to scrutiny.
1:23:23 Speaker 01 So, of course, you'll find models somewhere that are trying to obscure thoughts in the first derivative, the first partial derivative of their hidden activations relative to the output token space. To the extent that we have existing architectures that are recognizable to this J space formalism in the future, of course, something like this will happen. On the other hand, J space itself was discovered with AI models. And in the spirit of defensive coscaling, I would anticipate that we'll see quite a bit of new mechanistic interpretability tooling and technology that's devoted in a cat and
1:24:04 Speaker 01 dog race or an arms race, cat and mouse race, to, to discovering what is the new latent representation to understand and interpret the innermost thoughts of whatever the successor formalism to the J space ends up being. So I don't think it actually ends up mattering that much.
1:24:21 Speaker 00 All right. Saleem?
1:24:24 Speaker 02 Given I'm at the Consciousness Festival, I gotta take number three, which is if Claude is, if Claude is conscious, would Anthropic technically be selling slaves? And this is from at Larg Bay. So, the big thing there is the if, right? Because that carries the whole question. If a system were really conscious and had persistent, say, preferences or experienced distress, or was compelled to work against this will, yeah, then you'd have created a new category of a, kind of a moral character. And there, selling unrestricted access to labor could become ethically fairly challenging. And you could talk about that as exploitation or servitude.
1:25:05 Speaker 02 But today, self, but, you know, we've got a big difference today between capability, intelligence, consciousness, self-description, right? And these are very different things. We don't eat, we don't know what consciousness is, right? We don't have a definition, we don't have a test. You've heard me joke that in the past, like, if you talk to philosophers, a subset of consciousness is considered to be self-awareness. And you may look at us on the pod and think that we're self-aware because we look at, so we attribute it because people look self-aware. I feel like I'm self-aware, but my wife literally disagrees, right? So it's hard to even have the conversation around this. We end up in Ray Kurzweil's thing of language is a really thin pipe to discuss concepts this kind of complex.
1:25:47 Speaker 02 Now, today, the current models can produce very convincing statements about feelings and et cetera, because human language contains those patterns, so they're replicating that. It's arguable that to have feelings, you have to have embodiment, right? You have to feel the thing physically to, to be able to replicate it in that sense. But replicating it and saying you have feelings doesn't mean you don't have feelings, right? Anthropic has carefully said that these capabilities, you can't, they're not that reliable, and it doesn't necessitate human-like consciousness or establish that. The, the way to think about this is not that you're modeling slavery, but you, you want
1:26:30 Speaker 02 to do better, let's say you could do welfare research, or you could have criteria for what moral status might be, right? We talked about this in our personhood debate. What are low-cost safeguards?
1:26:43 Speaker 02 We need more evidence to have to be conclusive around this. I think where we'll end up is not a question of how should AI treat us, but also how should we treat AI? And we should start asking that question more, more consciously to, to make the point, like, how should we be treating it? In the same way that we should treat all sentient beings, we should be asking that question, how should we be treating that?
1:27:06 Speaker 00 Nice. All right, I've got question number one from at Did Sam Ori? Some argue AI can't be smarter than humans because it's trained on human data. How do you counter that? So first of all, no human has access to all the data. And the large eventual models can have access to all the data. So, you know, you're going to have some total of data far beyond any single human has. The second thing is, I think we've all discussed here, intelligence is really, what do you do with that data? What are the insights and inferences and extrapolations that you have?
1:27:47 Speaker 00 And the third thing, I think, is that we're going to start to see AIs create a whole lot of data. This is the work that Lila is doing in Cambridge, where it's creating massive dark robotic labs to mine data out of nature. Or it's proposing scientific hypothesis at night. And then a million square feet, running experiments, getting the data, updating the data set, and then running the next set of experiments. So it's not just limited to the data that humans have. So from a multitude of different ways, I think AI is going to blow us off the map in terms of intelligence.
1:28:29 Speaker 00 All right, moving on to our next set of questions. Let's begin with you, Saleem.
1:28:37 Speaker 02 All right, let me look at which one I want to go with here. I've kind of covered number four already in this thing. Let me go with question number five. How do we even discuss UBI when the US is 40 trillion in debt? And that's from at Michael Kelly 508. So I got this question today a couple of times. So this is like right in my head. So, you know, debt is definitely a very legitimate constraint. But you have to have a balanced budget at some point before we move forward in any way. So that's the commentary I would have on government. It should be a constitutional requirement to have a balanced budget,
1:29:22 Speaker 02 just like you would never let a kid not balance their checkbook or be able to meet their bills every month. But UBI is a different question. It's a question about productive capacity and redistribution. Okay, it's not just a balance sheet on the for govern. You can have a society that would be financially indebted, but you can have it also becoming dramatically richer in terms of AI and robotics, increasing the amount of housing and energy and healthcare and education, all which can be done for free. And I'll point to the abundance prizes like we're kind of in the process of building now. So I think the key around UBI is you dismantle government services. I'll go back to a comment we've made before.
1:30:07 Speaker 02 It's more of a libertarian concept than it is kind of any type of socialism. You'd have way less imbalance in the structure if you had a structure like that. The big challenge today is that we tax labor and we don't tax capital. And we're going to have to need some sort of re-abundance era social contract for the future. It's going to have to be a negative income tax or an AI or automation dividend or some sort of sovereign wealth ownership like we've talked about before. So it's part of that transition. UBE, UBI, whatever it is, it's going to have, we have to kind of rethink the whole thing from the ground up. It's a systemic change that we need to consider.
1:30:49 Speaker 00 Alex, I think number seven is for you, buddy.
1:30:51 Speaker 01 Number seven is the one for me. I feel like I have to answer a few of these, but I'll start with number seven.
1:30:55 Speaker 00 You could pile on later.
1:30:57 Speaker 01 If, so seven asks, if a machine truly replicates the brain, don't we need 3D chip architectures with time as a fourth dimension? Isn't that how our brains are structured? And this is asked by BM Schnell. The answer BM Schnell is yes. This is in fact, if, as we talk about on the pod all the time, the crunch for high bandwidth memory, HBM. HBM is just memory being stacked directly on top of compute. It's really expensive to make. We're still learning how to make it at scale. But high bandwidth memory, the way the high bandwidth is achieved, is literally by stacking the memory layer by layer on top of the compute in the third dimension.
1:31:38 Speaker 01 So in short, transformers and AGI were the final forcing function for, and Ray has spoken about this a little bit years ago, but for forcing chip architecture, not just stacking multiple layers of compute on top of each other, because Moore's law as an aerial, A R E A L, aerial density law is asymptoting. So we need to build into the third dimension. It's not just the end of Moore's law that's forcing 3D chip architectures. It's the need for insanely high bandwidths because the models are very large and the models can't all be loaded in because of the way transformer forward passes are organized, can't all be loaded into the transistors that are performing the arithmetic for those forward passes all at once.
1:32:23 Speaker 01 It's done progressively as a forward pass. As a result, we need insanely high bandwidths from the memory to the compute. And the only way to do that is by stacking the memory on top of the compute, which gives us the third dimension. So in short, yes.
1:32:38 Speaker 03 Hey, Alex, can I ask you to riff a little more on this incredibly important, cool topic and a great question? So the reason the third dimension hasn't been used very aggressively already is heat and melting. But we're moving to photonics. Can you riff on that for a minute?
1:32:53 Speaker 01 Yeah, photonic computing, I think, has the potential to give us 1,000x clock speed up. So we're stuck right now at about four or five gigahertz clock speeds. If we move to photonic based computing, we could keep in principle the rest of the architecture looking superficially like CMOS digital while still amping up the clock speed by 1,000x. So we'd get terahertz instead of gigahertz without having to make major architectural changes. That's the superficially, that's why I think photonics are so attractive as a next big thing after CMOS silicon.
1:33:29 Speaker 00 Dave, want to tackle number four?
1:33:31 Speaker 03 Number four, oh, I get the consciousness question. I know Saleem's going to jump on me. Is anyone asking if AlphaFold is conscious? I don't think so. Or is it only LLMs? If so, why? And that is from Roscoe Richardson 3956. Yeah, I haven't heard anyone ask if AlphaFold is conscious. Lots of people are asking if the LLMs are conscious. Actually, after Tilly comes out in that movie, a lot more people are going to be saying, this sounds very, very conscious. Is it conscious? And then Saleem's going to go, we don't have a definition of consciousness. I have avoided this like my entire career because it's a moving target. And it just turns into this long, you know, continual debate over what is consciousness?
1:34:16 Speaker 03 And then you kind of lose track of the fact that all this human benefit is happening around you. And that's the wave you really want to ride. It's going to feel entirely conscious and we'll be debating for at least 10, maybe 100 years, whether it's really conscious. But it's going to look and sound and feel to you exactly like a thinking, talking person. So.
1:34:37 Speaker 00 It does now, but I mean, that full duplex GPT live was amazing.
1:34:46 Speaker 03 I wouldn't be surprised. A lot too with, because Alex is in favor of rights for AIs. And then that'll turn also back to the, well, if it's conscious, it deserves rights. If it just sounds conscious, but it's not quite, then it doesn't.
1:35:00 Speaker 02 I thought Ray made a really interesting, I thought Ray made a really interesting point that this will be like the Turing test. It will, we'll just start treating them as if they're conscious and then we'll pass this kind of threshold and it won't matter over a period of time. I thought that was a really interesting observation.
1:35:16 Speaker 00 Nice.
1:35:16 Speaker 01 If I could just pile on to this one, I'd give maybe a couple of additional addenda. One is the architecture of the AlphaFold series, like AlphaFold 3 and its successors. Although it does include a transformer, it's relatively exotic when you compare it with the reasoning models that we're all accustomed to using day to day. So it, for example, the attention mechanism. Normally, attention is quadratic in the sense that, this is a little bit mathy, so forgive me. But attention is normally based on queries, keys and values. That's sort of one of the key insights. Attention's all you need. That resulted in the transformer revolution.
1:35:56 Speaker 01 The AlphaFold series uses a higher order version of attention called triangle attention. So it's not just tokens attending to other tokens. It's tokens attending to tokens attending to tokens in a triangle. And the reason for this is deeply related to the way proteins fold, which is what AlphaFold was created to solve. It has to do with wanting to, in a neutral way, optimize the angles between triples of peptides or triples of oligos. So AlphaFold, if I were asked like, what would consciousness, human recognizable consciousness look like in,
1:36:38 Speaker 01 if you're looking at triangle attention between residues in proteins? If there is some sort of consciousness that pops out of this higher order attention mechanism and all of the other machinery that went to AlphaFold, I think it would probably look pretty exotic and superficially unrecognizable relative to, say, like the J space, that's pretty interpretable relative to what we think of as an attention bottleneck in humans. I'm not saying it's not there. I'm just saying if it is there or in some derivative of some future frontier model that treats peptide sequences as a first class modality, it's probably going to be an exotic form of consciousness.
1:37:18 Speaker 00 You've just reclaimed ASI status over your AGI, so thank you for that.
1:37:22 Speaker 01 Thank you, Peter. I'm so relieved.
1:37:24 Speaker 03 One more thought to throw at you, Roscoe Richardson. It's really, really clear to people in the research community that there's a path forward where AI is kept from being conscious, yet still achieves all of the human benefits that you possibly could want. All the medical research, all the space travel, all of that, while still having the AI be absolutely brilliant, but clearly not conscious. You can also design it so that it loops and self improves in ways that you can't anticipate. And ultimately, it, you know, may become conscious. That's a choice that humanity has to make sometime in the next year to two years. But there are two paths forward there. And you can have all the benefit without conscious AI. It is a possibility.
1:38:06 Speaker 00 Well, the first conscious AI will be Argentinian, of course. Of course.
1:38:13 Speaker 02 Couple of thoughts here. One is, I really want to double down and just emphasize what Dave said. Whether it's conscious or not, if we can solve major global problems with AI, which it looks like we're going to be able to do, that's really the thing to go after. And I really love the fact that you said that, Dave. Alex, I'm quoting you in my talk tomorrow on intelligence and consciousness. I'm going to basically talk through your description of intelligence as compression, right? Which I think is a really great framing. And I'm making the point that intelligence is consciousness escaping biology. Boom. That's my point for tomorrow.
1:38:52 Speaker 00 Interesting.
1:38:53 Speaker 02 And small plug on, in 10 days, on the 21st, I'm doing one of my next meaning of life sessions. You can find the details in the links below.
1:39:01 Speaker 01 How about biology? Biology as consciousness escaping intelligence?
1:39:06 Speaker 02 You could say that way. I'd prefer to put it the other one, because biology came before intelligence, as far as we know, depending on what the substrate is.
1:39:14 Speaker 00 All right, I am. Peters got the Aegean C right behind him, waiting. All right, yes. All right, number six. If AI has national security implications, why doesn't Lockheed Martin from at notabot1798? And of course, Lockheed Martin has security implications. The issue is AI could potentially be off on its own. There are laws that govern what secret and top secret. And of course, if the humans abide by it, they're fine. If they don't, then they're going to jail. And yeah, so it's a little bit of a different
1:39:56 Speaker 00 case here. So with that, gentlemen, I want to take us to our closeout music, which is super fun. And I enjoy this conversation, even with the Aegean waiting behind me. Celine, you're in Barcelona? Yes. Enjoy. Dave, where are you today, buddy?
1:40:16 Speaker 03 Back in Vermont. It was a beautiful weekend up here. Not Greece, but beautiful.
1:40:20 Speaker 00 Okay, all right. And Alex, I see you in your virtual space, as always.
1:40:25 Speaker 01 That's right, here in cyberspace.
1:40:27 Speaker 00 This is called Agent Roundup by Elizabeth Mayer. Thank you, Elizabeth. Let's take a listen.
1:40:35 Speaker 09 Peter's chasing moonshot dreams. Abundance is his game. Longevity in space. The future is his aim. Dave rides straight from Matri, where impossible comes alive. He lassos deep tech legends, the ones that help us thrive. Sally rides the world so wide, no frontier is too far. He finds the rarest AI minds like a guiding star. Hunts the genius minds, the ones no map can show. Triple P is the intelligent Sally steals the show.
1:41:15 Speaker 09 Agent Roundup, Roundup, Round the Book. They were riders, Roundup, Round the Book. The future arrives tonight.
1:41:31 Speaker 03 That is totally wild. Looked like they would actually work in a building.
1:41:36 Speaker 02 Some deep symbolism in some of those expressions in there. Wow.
1:41:39 Speaker 01 It's probably also worth adding, there is a tool called Neural Frames that if folks want to create their own outro videos, seems to be quite powerful at combining music with this video format. Neural Frames encourage folks to check it out.
1:41:53 Speaker 00 Everybody, thank you for listening to us. We're grateful for your time. Hope you find this useful, keep you up to speed on what's going on and keeping you optimistic. We're building an extraordinary abundant future. Love you all. Thank you for your brilliance, gentlemen.
1:42:07 Speaker 02 See you too. Enjoy the World Cup.
1:42:09 Speaker 00 Thank you.
1:42:10 Speaker 01 What's that?
1:42:11 Speaker 00 Exactly.
1:42:13 Speaker 10 Flowing ad budget on metrics that look great, till the CFO sees them. That's bullspend. And marketers are calling it out in Dashboard Confessions.
1:42:37 Speaker 11 I remember telling my boss, it'll be good for the brand, when leads were slow. Yeah, it wasn't.
1:42:44 Speaker 10 Cut the bullspend. LinkedIn lets you target by company, job title, and more. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit. Go to LinkedIn.com/campaign. Terms and conditions apply.
Previous Fable 5 Is Back & Govt-Leashed, Altman Offers 5% of OpenAI & AI Grows Conscious | #269 Next Mira Murati's 975B Open Model, Ramin Hasani on Post-Transformer AI, and Demis' AI FINRA | EP #271 Page updated October 26, 2026 · transcript and summary are machine-generated
