Stop The AI Race organizer proposes daily pickets to persuade AI company employees to support a coordinated pause in developing more powerful models
Stop The AI Race organizer proposes daily pickets to persuade AI company employees to support a coordinated pause in developing more powerful models
In an essay published on November 10, Michaël Trazzi urged people to act as though this were their last chance to prevent AI from causing human extinction. He proposes further hunger strikes to show how seriously participants take the threat.
His essay was prompted by a November 5 report from the research organization METR on the AI developer Anthropic. METR estimates that, by October 2026, the rate at which its models improved on benchmark tasks had roughly doubled compared with the period from March 2024 to October 2025. Anthropic employees report that AI writes code and helps run experiments to develop subsequent models. Trazzi fears that more powerful systems will accelerate the development of still more powerful ones.
Trazzi asks readers to imagine that humanity has been wiped out by superintelligence, meaning AI that surpasses humans at virtually every intellectual task. You are given a chance to return to the present and try again: “What would you do differently?”
Trazzi, a former AI researcher and documentary filmmaker, went on a seven-day hunger strike outside the London office of the AI lab Google DeepMind in September 2025. Participants asked DeepMind’s leader to publicly commit to a pause if other developers also stopped. Trazzi now proposes daily pickets outside Anthropic. He estimates that two to four people would be enough to hold signs at its two main entrances. Employees would encounter critics on their way to work; Trazzi hopes this reminder of the risk will encourage them to support a halt to the race. The aim is to “be Anthropic’s conscience.”
Anthropic co-founder Chris Olah has said that AI labs need outside critics: developers face pressure to generate revenue and outpace competitors, alongside geopolitics, pride, and ambition. He called on people free of those pressures to be “our sincere, thoughtful critics.” Trazzi cites those words in proposing daily protests with signs outside the office.
In the discussion of the essay, AI safety researcher Vojtěch Kovařík supported Trazzi’s protests and suggested considering how risky actions might affect other participants. He considers it acceptable to risk his own well-being to bring about change at a company. When several participants take risks without coordinating, they can harm shared resources and trust in the movement. His criterion is that the potential benefit should outweigh the harm that failure could cause.
The movement calls for all competitors to pause simultaneously. Under its proposal, existing models and specialized AI applications would remain available; teams developing more powerful models would turn to other tasks, including AI safety research. Anthropic itself makes the feasibility of a pause conditional on verifying that other developers have also stopped. Otherwise, a company that secretly continued training models could pull ahead. The lab explains that agreement is needed in advance on the conditions for starting and ending a pause and on who would resolve disputes.
Marches in 18 cities have been announced for November 14. Their organizers are asking Donald Trump and Xi Jinping to begin negotiations on a treaty that would prohibit the development of more powerful systems until it is safe to proceed.
- https://www.lesswrong.com/posts/Nts7RA9D2wTfcvk56/have-you-considered-actually-trying
- https://metr.org/blog/2026-11-05-anthropic-ai-rd-acceleration-assessment/
- https://t.me/UkhvatNews/4889
- https://www.latimes.com/business/story/2026-07-22/they-built-ai-now-silicon-valley-workers-protest-its-threat-to-jobs-humanity
- https://www.anthropic.com/news/chris-olah-pope-leo-encyclical
- https://www.lesswrong.com/posts/Nts7RA9D2wTfcvk56/have-you-considered-actually-trying/comment/Gv9ApuMmgrPuWytxY
- https://stoptherace.ai/
- https://www.anthropic.com/institute/recursive-self-improvement
- https://aitreatymarch.org/