@teortaxesTex I guess these things are not like EUVs. The US should have better ways to produce power than to use these machines to make silicon ingots and [omitting multiple steps]. At the very least, they could import solar modules, which would save them many years.
@phl43@teortaxesTex Seems to show that compute is important but not that important. If it's log(compute) on some metric, a 10x compute advantage keeps you a constant distance ahead.
@atelicinvest Perhaps intelligence was never that important and was not the bottleneck in most cases. Look around, most people never developed the human intelligence that is now being matched/surpassed by AI.
@TheStalwart Financial assets are claims on future production. If you expect future production to be distributed to future producers or the general public rather than asset-owners/rent-seekers, financial assets do poorly.
@TheStalwart If by intelligence-intensive you mean very smart people (like mathematicians) make a huge difference. Well, there aren't many. Just look at how well very smart people got paid before LLMs.
@itsolelehmann Just take the second example, cheap medical imaging. The cost in China is a few percent of the cost in the US. Do you think technology accounts for most or even a meaningful fraction of the difference?
@anri_oyuki_tshi@miwakovamp Trump seems to like and respect all three of those leaders better than the Japanese one. I don't deny there are difficult relationships to manage, but counting on American deterrence might not be the best approach.