The big (possibly wrong) argument against this is that China is still behind in tons of fields despite, frankly, having just obscenely more smart people in them, to the point you can think of it as ≈2 years of a gap in AGI. How many IQ-point-years do you think went into ASML's EUV? Likely LESS than what they're throwing at research every year now. And yet. Domain knowledge is deep, expensive and slow to acquire. What "iterated chip designs"? You think Jensen will hand his moat over to Anthropic and pay for Claude tokens on the way? Rationalists routinely assume the supreme power of first-principles thinking, Yud made a lot of noise about how AlphaZero rendered human Go expertise obsolete. Between learned human heuristics dulled with lossy compression into explicit transferable knowledge for accelerating generational gain, and an end to end learned heuristic that scales with available compute, it's no wonder the latter wins. But there's a difference between naturally self-contained logical domains that humans tile over with heuristics because that's all we meat-brains can do, and inherently deep and wide domains the dynamics of which are very far removed from their underlying generative process; and we don't know exactly where the boundary lies for any particular field. We can make an educated guess, however. I'd say that I wouldn't be surprised if Gemini Flash tier model that ingests Google's TPU development data and software will do better at "iterated chip design" than Claude Fable despite the latter's crushing intellect and more profound understanding of electrical engineering as publicly known. And GLM 5.2 level Nemotron customized by Nvidia will likely do better still. This is not an academic debate; this is a question of the extent to which incumbents will allow frontier labs to convert their current profits into a lasting vice grip on the economy and eventual elimination of said incumbents, to improve their own reports for the next few quarters. If they have any business sense, they'll aspire to keep this extent limited.
put simply, i think this claim is incredibly false, and this is what drives a lot of my understanding and assumptions around how all this will play out. i think viewing models as slowly replacing individual tasks and functions and "locking in" once they achieve sufficient capabilities there is deeply myopic. we will not have "the prior economy except with models doing the work". in fact what will happen is the same thing that always happens. new capabilities will lead to *new categories* of work, done by models not humans, and create huge swaths of value that was previously untouchable and incomprehensible. when you can pay for frontier++ intelligence to loop and automatically discover 3 new world-changing drugs per month, people will pay for this. in fact they will saturate spend on this, because the value of these opportunities is so so high. when you can pay for frontier++ intelligence to fanned-out run entire companies as mini experiments, you will do such, because it gives massive competitive advantage and scale in every possible niche. or maybe *you* won't, but others will, and they'll be the ones who remain economically relevant while you're having GLM 5.2 rewrite your emails and update your SaaS landing page. when frontier++ models are capable of iterating on chip design and distributed software architectures we currently view as only possible with decades of effort, countless corporations will pay the costs, because they'll generate economic returns at scales orders of magnitudes above what models are doing now.
the intelligence waterline keeps marching up. so you solved health insurance claims review with a fine-tuned Qwen that achieves 100% perfect accuracy at optimal cost without frontier models? awesome, yeah honestly that will make you a bunch of money for a while especially given regulations are gonna be slow to change. but *relative to what will happen elsewhere*, your slice of the pie will shrink to irrelevance, because other newer areas will be so so so much more incredibly valuable.