A new directive linked to Anthropic already restricts foreign nationals.
Financial Times, quoting a person close to OpenAI:
"In recent days, the [AI] industry has been working [with the USG] on ensuring foreign national researchers could continue to work on developing the most advanced models, a practice that the Anthropic directive has now banned."
This appears to indicate that the USG wants to restrict non-U.S. persons from working on frontier models generally, across the industry (not just at Anthropic).
Many users criticized the US government's move to bar foreign researchers from frontier AI models as misguided and damaging to progress, while a few welcomed talent returning to Europe or called the policy good business.
told ya
Financial Times, quoting a person close to OpenAI:
"In recent days, the [AI] industry has been working [with the USG] on ensuring foreign national researchers could continue to work on developing the most advanced models, a practice that the Anthropic directive has now banned."
This appears to indicate that the USG wants to restrict non-U.S. persons from working on frontier models generally, across the industry (not just at Anthropic).
Complete and utter self destruction
Financial Times, quoting a person close to OpenAI:
"In recent days, the [AI] industry has been working [with the USG] on ensuring foreign national researchers could continue to work on developing the most advanced models, a practice that the Anthropic directive has now banned."
This appears to indicate that the USG wants to restrict non-U.S. persons from working on frontier models generally, across the industry (not just at Anthropic).
@kristoph @Presidentlin I don't think that Jakub Pachocki is any more likely to leave for China than Sam Altman is.
Conversely, a U.S.-born person of Chinese descent could just as easily leave for China as a Chinese-born person of Chinese descent.
@ChrissGPT @AndrewCurran_ I don't even think that making these people citizens is strictly needed. The national security concerns related to model weights are real and serious, but preventing Andrej Karpathy from working on Mythos 5.1 is not the answer.
I’d bet huge amounts of money against this. It’s just libs doing their anti-Trump thing. They did it with visa processing a few week ago. Fine hate Trump for being corrupt and stupid if you want. He doesn’t want the US to lose the lead in AI and understands foreign researchers are needed correctly well
https://www.ft.com/content/f6940d59-28f4-4ae4-a569-c6fc421e52b9
@TeksEdge Note that this isn't about access for consumers. This is about banning frontier lab employees who are not U.S. citizens from working on further development of frontier AI models.
@daniel_mac8 @deredleritt3r What are they missing?
If these things are as dangerous as nuclear weapons, like Dario says, why wouldn’t they apply the same principles to the research?
@deredleritt3r Welcome to Europe
The export controls were just the lever they could pull fastest. There has never been any intention to prevent foreign researchers. That being said I am willing to bet we see how seriously the threat of espionage within labs is taken to increase drastically. No one will want a repeat of the Rosenbergs leaking nuclear secrets to the Soviets
@deredleritt3r @AndrewCurran_ Why don’t we just begin the process of making them citizens, monitor them and continue their critical work
@reliabytes @deredleritt3r @Presidentlin Anthropic has had "unvetted foreign nationals" for years, resulting in them making the (arguably) best models in the world. Same for OpenAI.
@deredleritt3r @Presidentlin The rationale here is that non-US persons are more likely to simply leave for another country (China) given additional financial incentives which US persons are less likely to do.
Also US persons can be subject to regulations non-US persons can avoid by just leaving.
@daniel_mac8 @johnennis @deredleritt3r When you have Fable 5 running full speed ahead and someone from the Trump administration throws a wrench into it….
@deredleritt3r Read that too. Seems like access will be based on proof of identity. X does this now.
@deredleritt3r @kristoph @Presidentlin @grok how many generations of China descendants get to move back to China (vs. just allowed to visit)?
@deredleritt3r There is also direct counterevidence to the broadest interpretation: The Information reports that the government is unlikely to extend this specific directive to other AI companies and that an official characterized it as particular to Anthropic’s response to government concerns
@daniel_mac8 @deredleritt3r There are work arounds for foreign nationals. Dario made everyone’s life more complicated by being stubborn.
@CalimanuLoredan @deredleritt3r Yeah, this is crazy. Lots of thinking is needed. Andrew Curran wrote an interesting piece. Last night like 1000 views. Today 1.4M views.
@kristoph @Presidentlin Totally agreed.
I think a government vetting process for each foreign (and non-foreign) employee would be fine in principle. This would allow for granular exclusion of people who actually could be a risk (not based on nationality, to be clear).
