This is a special post for quick takes by Archimedes. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
Archimedes's Shortform
4Archimedes
7ryan_greenblatt
2Archimedes
2cubefox
2ryan_greenblatt
Limiting China's computing power via export controls on hardware like GPUs might be accelerating global progress in AI capabilities.
When Chinese labs are compute-starved, their research will differentially focus on efficiency gains compared to counterfactual universes where they are less limited. So far, they've been publishing their research, and their tricks can be quickly be incorporated by anyone else. US players can leverage their compute power, focusing on experiments and scaling while effectively delegating research topics that China is motivated to handle.
Google and OpenAI benefit far more from DeepSeek than they do from Meta.
1One of the main drivers, perhaps the main driver[1], of algorithmic progress is compute for experiments. It seems unlikely that the effect you note could compensate for the reduced pace of capabilities progress.
Both labor and compute have been scaled up over the last several years at big AI companies. My understanding is the scaling in compute was more important for algorithmic progress as it is hard to parallelize labor, the marginal employee is somewhat worse, the number of employees has been growing slower than compute, and the returns to compute vs faster serial labor seem similar at current margins. That's not to say employees don't matter, I'd guess Meta is substantially held back by worse employees (and maybe worse management). ↩︎
Compute is definitely important for experiments. The limits undoubtedly slow China's progress, but what's more difficult to determine is whether global progress is slower or not. In the toy scenario where China's research focus is exactly parallel and duplicative of Western efforts, they contribute nothing to global progress unless they are faster. More realistically, research space is high-dimensional, and you are likely correct that the decreased magnitude of their research vector likely outweighs any extra orthogonality benefits, but I don't know how to apply numbers to that tradeoff.
Both labor and compute have been scaled up over the last several years at big AI companies. My understanding is the scaling in compute was more important for algorithmic progress
That may be the case, but I suppose that in the last several years, compute has been scaled up more than labor. (Labor cost is entirely reoccurring, while compute cost is a one-time cost plus a reoccurring electricity cost, and a progress in compute hardware, from smaller integrated circuits, means that compute cost is decreasing over time.) Then obviously that doesn't necessarily mean that an AI company with access to FLOP/s compute and AI researchers has an advantage over a company with only FLOP/s compute but researchers.
In fact I think in that sense labor is likely more important than compute for algorithmic progress. And that doesn't seem so far away from reality, if you model as a US company with cheaper access to compute and as a Chinese company with cheaper access to labor (due to lower wages).
I don't think parallelism works very well among employees while it works great for compute.
I agree that labor is probably a somewhat more important input (as in, if you offered an AI company the ability to make its workers 2x faster in serial speed or 2x more compute, they would do better if they took the 2x serial speed. I'd guess the AI companies are roughly indifferent between 1.6x serial speed and 2x compute, but more like 1.35x vs 2x is also plausible.
It seems plausible to me that well enforced export controls cut compute by a factor of 3 for AI companies in china, and a larger factor is plausible longer term. This would substantially reduce the rate of algorithmic progress IMO.
Curated and popular this week