It will never stop at 5%. It will go on and on and on. The governance will become a nightmare. Political capture will be real. And it will generate precisely no goodwill with the public. None, if they themselves see no direct financial benefit.
“What has the AI industry even done for America.”
“Well, it handed a collective $200b of itself to Donald Trump.”
Half the country instantly hates you, and even a decent chunk of Republicans will assume this is corrupt by default.
No, nobody at OpenAI is discussing a word of this with me. I don’t work there yet. And this rumor may even be importantly untrue or misleading in some way. A vehicle that distributes ownership among the people can make sense. A government stake, however, is the wrong path.
I do however invite all the people who said I have become a corporate marionette to submit their public apology forms at their leisure
• • •
Missing some Tweet in this thread? You can try to
force a refresh
I think part of it, at least vis a vis US/China competition, is that US and western chattering classes find it hard to believe that the market-driven outcome of frontier AI could possibly be right. They basically believe, in their hearts, that the Chinese system, with its “industrial strategy,” has eclipsed capitalism. So they harbor the same inferiority complex toward the Chinese system that many Americans once harbored toward the EU’s system. Their heuristic is that the industrial strategists of China have grasped the whole picture of the technological competition in a way that US industrialists, with their “profit maximizing incentives,” could not possibly have matched. And so any outcome in the economy that is not the result of “strategy” is therefore prima facie worse than what the “strategists” have concocted. They also believe the Chinese strategists possess awesome powers of foresight and the ability to evade all tendencies of financial and economic gravity, due of course to “strategy,” really it’s almost a kind of orientalism.
Meanwhile the U.S. industrialists are making new advances in math and science, and the fastest-growing businesses in history, by spending hundreds of billions of dollars on high-margin chips whose legacy is in rendering video games, cramming them underneath tents if need be, and investing generational capital into new energy generation technologies as they do it, and perhaps even colonizing space as an instrumentally convergent result. But none of that is “strategy,” you see.
The salient thing is that nobody, absolutely nobody, in the Washington DC strategic class describes “capitalism” as existing in competition with the Chinese system. It is always “democracy.” They think capitalism already lost and that we have to become like China, with state-led public/private enterprise (“strategy”). So the word we use to describe, waves hands, “our way of life vis a vis their way of life” is “democracy.” Their defense is of a political order. On average they seem to feel no particular attachment to the economic order. This is bipartisan and is why the feeling of many in DC was “of course the government can tell anthropic to do whatever they want!” Civil-military fusion, the death of the private and its assumption into politics, they ultimately cheer these things on. They love them. They only decry them when social media is framed as the driving agent behind these trends (eg “polarization” discourse). Then it is bad, but that’s just because that framing makes them feel not in charge.
The irony is that I suppose I am a part of the U.S. strategic class and the above analysis is in fact Confucian af
The AI community is laser-focused on SB 1047, but there's another AI bill in California that is equally, if not more, aggressive than 1047: AB 3211.
Like 1047, AB 3211 has passed one chamber of the legislature and is authored by a powerful legislator.
Let's take a look. 🧵
AB 3211 is a deepfake bill, officially titled the Provenance, Authenticity, and Watermarking Standards Act.
It mandates the use of watermarking for all generative AI systems, and requires all large websites and apps to display provenance information for digital content.
The bill requires that *all* generative AI systems, regardless of content or size, include watermarks. As written, it applies to everything from a nucleic acid sequence predictor made by a grad student all the way up to multi-billion dollar foundation models.
The Frontier Model Division now has a new power: the ability to raise or lower the threshold for a covered model at will. This gives the agency regulatory powers that it could deploy in unpredictable ways.
This power includes the ability to unilaterally change the threshold for fine-tunes that are under its jurisdiction.