Unilateral pauses risk empowering less cautious competitors, Anthropic warns
I think it's excellent & notable that the answer to "what should we do" in Anthropic's blog post on RSI is essentially figuring out ways to slowdown/temporarily pause frontier AI development.
Many users accused Anthropic of regulatory capture and monopoly-seeking in its call to pause AI development over recursive self-improvement risks, while others praised the statement as honest or prescient.
@MittRomney hahahaha mitt is a doomer
And my top quibble is that the tone reads like "RSI could happen but don't fret too much, it'll probably be fine" rather than "omfg we're possibly on the brink of AIs that make smarter AIs that make smarter AIs, society needs to *act*". But it's a step in the right direction.
Our highest and most urgent national priority should be AI safeguards. The risks of AI weapons, pathogens, mass unemployment, surveillance, and even extinction must not continue to be largely ignored.
@MittRomney Pls sir Don't fall for the Doomer psyop plz
@MittRomney
One big quibble is that they aren't thinking big enough. With sufficient intelligence you *can* predict pretty well what a drug would do over a decade of use, and/or sidestep that problem entirely by designing de-novo more-predictable biosystems from scratch, etc.
@MittRomney I actually think there's an even bigger priority, which is stopping AI companies from pursuing recursive self-improvement and superintelligence.
Safeguards are good, but we actually don't know how to make AI safe and beneficial, and that only gets harder as AI gets smarter.
@MittRomney Thank you Senator, humanity is indeed in grave danger. Your colleague Sanders met with @ESYudkowsky, @So8res, @JeffLadish, @DKokotajlo and you could do a lot worse than meeting with that same group!
@MittRomney Please sir take your cues on this from Utah’s current governor and embrace abundance over caution. Many of these risks are only abstract theoretical ones and cannot be disproven. The precautionary principle will do us more harm than far off non zero extinction theories
@MittRomney "The facebook told me the AI is gunna nuke the Shoneys!"
@MittRomney Who are you?
@MittRomney While we debate sci-fi extinction, real data shows we are dealing with tools, not self-improving minds
Independent METR studies show devs using AI took 19% longer on tasks. The actual capability of the technology is far less alarming than the hype
@MittRomney No, it should be deportations and shutting down the H-1B program.
If I can’t have my country back then I don’t care what happens with AI.
@MittRomney No, the Save Act is the most important priority for national security. If we don’t have confidence in our elections, the whole nation goes down.
@KrittinKalra @MittRomney Hi Krittin, we published an update in February where we say: Based on conversations with study participants, we believe it is likely that developers are more sped up from AI tools now — in early 2026 — compared to our estimates from early 2025. https://metr.org/blog/2026-02-24-uplift-update/#wider-adoption-of-ai-has-made-it-more-difficult-to-measure-task-level-productivity
@j_asminewang @tyler_m_john - it's crazy that this was less than 3 years ago, and the entire debate was ridiculed as completely irrelevant to what could actually happen. https://forum.effectivealtruism.org/s/vw6tX5SyvTwMeSxJk/p/3hSEQnEN2D3SSzHWn
@MittRomney We're glad to hear you say this!
We list 90+ ways the public and policymakers can help here: http://existentialsafety.org. Please consider galvanizing your followers to action.
@MittRomney In the wildest dream, would I pondered agreement with you ... respectfully, Mr. Romney.
But I do AGREE WITH YOU.
HIGHLIGHTING : "even extinction must not continue to be largely ignored."
@MittRomney AI dooming is not based on logic, data, or evidence.
@So8res Saw this here first, so linking the blog in case anyone else in the same boat: https://www.anthropic.com/institute/recursive-self-improvement
