On October 31, Scott Alexander published "Against SURA" on LessWrong. He uses SURA (Stop Unilaterally + Raise Awareness) for the view that AI companies, their employees and countries should halt AI work on their own and focus on raising awareness, and argues it would hand the lead to companies that do not prioritize alignment and cost the US its bargaining chip with China. He cites AI Futures Project members' survival estimates for scenarios without a global pause, mostly in the 25%-50% range, to push back on 90%+ doom estimates. Blogger jessicat quoted it and said parts of AI safety now behave like a mass political movement with bad epistemics (about 359 likes, 63K views); Bloomberg Odd Lots co-host Joe Weisenthal summed it up as people competing to out-doom each other (about 287 likes, 38K views). This is an opinion debate, not a policy announcement from any organization.
Key Takeaways
- ✓Source: Scott Alexander's "Against SURA" on LessWrong, posted 2026-10-31 19:32 UTC; SURA stands for Stop Unilaterally + Raise Awareness
- ✓AI Futures Project members' survival estimates for three no-global-pause scenarios (C+ / C / D), as cited: Daniel Kokotajlo 40%/20%/10%, Ryan Greenblatt 60%/63%/55%, Eli Lifland 50%/45%/32%, Thomas Larsen 45%/35%/20%, Romeo Dean 40%/35%/20%; the AI forecaster FutureSearch gave 44%
- ✓He says that if the US unilaterally passed Sanders' "Ban Artificial Superintelligence Act" (endorsed by MIRI), the AI forecaster he asked put the chance of China agreeing to a pause within a year at only 7%, up from 1%
- ✓jessicat's quote post (Oct 31, 23:01 UTC) has about 359 likes, 28 reposts and 63K views; Joe Weisenthal's quote (Nov 1, 14:57 UTC) has about 287 likes, 33 replies and 38K views
- ✓Note: the "10-40%" in Weisenthal's post is his paraphrase; the estimates in the essay's table range from 10% to 63%
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
Project Links & Resources
Direct AccessIn-Depth Technical Analysis
Background
The AI safety community has long had two broad strategies: push for a global pause on AI development (Plan A in the essay), or, if the race cannot be stopped, keep alignment-focused labs in the lead and have them spend that lead on alignment research at the last moment (Plan C). Scott Alexander argues that since 2023 the community has bet ever more heavily on the first and now treats the second as an accelerationist excuse.
The argument
He groups a set of positions under the name SURA: alignment-focused companies should stop unilaterally, their employees should quit, and the US should restrict AI without a treaty with China. His claim is that each of these is net negative in a world that ends up needing Plan C: it hands the lead to companies that do not prioritize alignment, weakens alignment teams, and removes the US bargaining chip that could bring China to the table. Quote posts from jessicat and Joe Weisenthal shifted the focus to community culture, saying parts of the movement act like a mass political movement where people compete to sound most pessimistic.
Benchmarks and data
There are no benchmarks here. The only numbers are subjective probabilities from AI Futures Project members: survival estimates for three no-pause scenarios range from 10% to 63%, mostly 25%-50%, which the author uses to question 90%+ doom estimates. These are personal forecasts, not measured results.
What it means for practitioners
For people working in AI, the debate bears on real choices, such as whether to leave a frontier lab and how to view unilateral regulation bills. For the full picture, read the LessWrong essay and the rebuttals in its comments, and check the AI Futures Project source document rather than relying on the numbers summarized in quote posts.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.