On the road to fully automated luxury gay space communism.
Spreading Linux propaganda since 2020
Rules:
They have to keep buying more computing power, they have to keep pouring concrete just to convince the investors that they’re still in the lead in the race for general intelligence.
They're all in a race for general intelligence, but I don't know how much you can find between them.
Business leaders today say their companies care about more than profit and loss, that they feel responsible to society as a whole, not just to their shareholders. Corporate social responsibility is their new creed, a self-conscious corrective to earlier greed-inspired visions of the corporation. Despite this shift, the corporation itself has not changed. It remains, as it was at the time of its origins as a modern business institution in the middle of the nineteenth century, a legally designated “person” designed to valorize self-interest and invalidate moral concern. Most people would find its “personality” abhorrent, even psychopathic, in a human being, yet curiously we accept it in today’s most powerful institution.
-Joel Bakan, The Corporation
Capital has no limits – in principle, it has no limits of any kind. Not only does it have no limits in terms of growth or profits, it has no intrinsic moral limits. Whatever limits it does have must be constraints put on capital from the outside – for example, by laws. Capital itself recognizes only one thing: the need to create more of itself.
Thus, the aparent "limitless stupidity" in search of artificial intelligence is but the mere logic dictated by the intrinsic nature of capiral
China's also masterfully playing the optimal strategy for 2nd place in AI (also stealing is cool). Qwen and deepseek could replace OpenAI and Claude for everything I use AI for and it would be one of the smallest set backs imaginable. LLMs are basically version 2 of your phone's predictive text, and it's not clear to me that a meaningful V3 is on the horizon or is generally possible.
So on one hand, is China going to race ahead of US AI tech in the next 5 years? It's not clear to me that they should even try. Stealing the US tech (specifically, copying high quality outputs that the US companies built via mass copyright infringement) may prove to continue to be the 2nd mover's optimal strategy for the next 5 years.
China could actually be better off taking advantage of the marginally weaker LLMs they have already created, but primarily focus on the rest of their economy and infrastructure. IMO the result of this will essentially be the V2 of that one time Nixon generously donated all the US manufacturing capability to China. The US will be a dead hollow economy, and we'll have nothing to show for it, while China will have a beautiful wonderful county and AI tech.
It really feels as simple as the saying 'the US is still the place where you can go from 0 to 1, China is where you go from 1 to 100'. It's an industrial power that also invests in R&D. Once the technology is proven and scalable, the Chinese will jump right in.
It is the American Economy that is weird, given that it always bets on the ability of silicon valley to generate hyper growth. A bet that hasn't panned out lately, with Bitcoin, stablecoins, NFTs, the Metaverse and now LLMs. The underlying assumption of the massive US capital markets is that their bet on R&D will always yield Smartphone or Amazon Cloud Computing levels of economic revolution and as far as one can see there are tech companies getting used to less hypergrowth and more enjoying the fruits of having a monopoly over entire markets.