China’s Endgame: ASI Timelines, US-China Relations, and the $1.7T AI Bubble With Alvin Graylin | 281

I like when AWG presses people that dodge his questions.

As a political scientist I must say that analogy between US AI investments and USSR arms investments is simply preposterous. Here is why. It is conceptually backwards to treat “strategic investment” as if the object of investment were irrelevant. But it matters ENORMOUSLY whether a state is pouring resources into nuclear warheads, tanks and missiles or into a general-purpose technology that expands its capacity to produce knowledge, software, medicine, industrial designs, scientific discoveries, logistics, weapons, education, media and eventually perhaps much of the rest of the economy. Weapons are overwhelmingly terminal goods. Their value is concentrated in a narrow range of four basic functions: attack, defense, deterrence and coercion. A strategic missile does not normally make your pharmaceutical industry more productive, discover new materials, automate accounting, improve weather forecasting, design better factories or accelerate scientific research. Even when military R&D generates civilian spillovers, those spillovers are incidental to the principal function of the asset. AI is almost the exact opposite. Intelligence is an extraordinarily general productive. Ptentially, it is an input into nearly every activity in which cognition matters. More capable AI can help produce better AI, but also better software, engineering, drugs, manufacturing processes, logistics, scientific models and, yes, military systems. It can therefore increase both civilian welfare and military power simultaneously. That difference completely changes the economics of “overinvestment.” The Soviet arms buildup consumed enormous amounts of capital, skilled labor, industrial capacity and scientific talent in order to manufacture assets whose civilian economic return was extremely limited. In a sufficiently extreme arms race, additional military expenditure can actually weaken the productive base that ultimately sustains national power. I agree that investment in AI infrastructure may be wasteful at the margin, particular firms may overbuild, and speculative bubbles are obviously possible. But that is a extremely different proposition from saying that AI investment is structurally analogous to Soviet military overinvestment. Because AI is becoming a major general-purpose technology, the compute centers, energy infrastructure, semiconductor capacity, algorithms and human capital created during the investment boom can contribute to productivity throughout the economy. The apparent “arms race” is simultaneously an industrialization race, a scientific race and a productivity race. In fact, if one insists on using historical analogies, AI may be closer to electrification, industrial machinery, computing or the development of a highly educated scientific workforce than to accumulating tanks and nuclear warheads. Sure, a country that “overinvests” somewhat in electricity, computation or productive intelligence may end up with excess capacity. A country that systematically underinvests while the technology transforms the global production function may discover that it has surrendered not merely military advantage but economic, scientific and technological leadership. An arms race is largely about relative destructive capability. An intelligence race is about the capacity to generate capabilities themselves. Whoever possesses more abundant, cheaper and more capable machine intelligence may acquire an advantage not in one particular weapon system but in the rate at which new technologies, products, institutions and strategies can be developed. Weapons are produced BY an economy. Sufficiently advanced intelligence can improve the economy that produces the weapons. Conflating the two misses the recursive character of the technology. The legitimate questions are whether the investments are technologically sound, whether capital is being allocated efficiently, whether energy and semiconductor bottlenecks are being handled rationally, and how the resulting gains will be distributed. But the Soviet arms-race analogy does almost nothing to answer them, because it mistakes a potentially universal productive technology for a stockpile of specialized military hardware. By the way, the obvious propagandistic messaging from the guest was so pervasive that it absolutely overwhelmed the analytical substance. Big credit to AWG for his vigilance and assertiveness!

Salim knows the struggle of common man glad he is always on this show

As important as this discussion is, I think this is the most suspicious I have seen Alex react to anyone on this pod. If Alex is worried and questioning what he is hearing, then I think we need to be very worried.

Well, this guest has really strong safeguards that Alex is trying to penetrate. I applaud him for it.

I knew AWG was brilliant. After this performance, I absolutely love him.

AWG, thank you for standing up for objectivity. If no one calls balls and strikes like you do, it's game over for society.

Keep the Regular Uploads! Upload everyday and I will watch! I’m a College student and I consider these videos more important than my Physics lectures

doing his best to respectfully call "balls and strikes" with Alvin. ⚾

Single most impactful podcast in my life.

This has been an exceptional episode. As brilliant as AWG is, I would argue that the silicon valley folks can be a bit out of touch even with the American public right now, and Alvin is doing a good job bringing a more balanced viewpoint. Anyone familiar with the prevailing sentiment in China can tell that he's being truthful and not a "CCP mouthpiece" as some people in the comments suggest. Whether or not his thesis holds water is a different story, but I believe that he is sincere and there is a general arrogance in certain tech circles in America that may prove to be more detrimental than one would expect.

The episode started out like a full blown CIA debriefing session😂

Did he say Chinese models remove Tiananmen square for formality? :49

This gotta be the best episode of Moonshots yet....that's what I think every month or so. It's a constant race to the top with you guys

Very insightful interview. Thank you, all!

Another great pod. Alex calls ‘em as he sees ‘em. Dave tried heroically to get a straight answer and Alvin did what he had to do. Respect to all. Excellent dialogue/debate/answer evasion exercise.

Much respect, Alvin. But as the saying goes, you can't wake someone who is pretending to be asleep

Excellent Discussion A++

thanks for the perspective of the China. it is refreshing to see the different view. most of the commenters with the biased will see through their own biased lens, well, it can also be said of your own talking point a bit too. afterall, this is part of your experience in China and they have been exposed to the thinking there. it is a different perspective than what we keep hearing all the time so I welcome the perspective.

This is exactly Salim’s main thesis. It’s not about technology. It’s about the redesign of the organization.