AI
Is Big Tech’s AI slowdown a safety pact or a cartel?
Over the weekend, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google DeepMind cofounder Demis Hassabis, and SpaceX head Elon Musk loosely agreed to slow
Key takeaways
- Sam Altman, Dario Amodei, Demis Hassabis, and Elon Musk loosely agreed to slow down AI development.
- The proposal calls for embedding third-party auditors, regulating domestic labs, and reaching a global slowdown agreement.
- Recursive self-improvement is starting to happen across the industry and could occur as soon as early 2027.
- The voluntary slowdown agreement is a cartel-like move to stop competitors and avoid real legal safeguards.
- AI leaders are bowing to pressure from safety researchers and employees who have called for a slowdown for years.
Over the weekend, OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei, Google DeepMind cofounder Demis Hassabis, and SpaceX head Elon Musk loosely agreed to slow down AI development to "pace the frontier." The three-step proposal, originally detailed in an essay by Amodei, calls for embedding third-party auditors, regulating domestic labs, and reaching a global slowdown agreement. The push is heavily driven by concerns over recursive self-improvement (RSI) - where AI models train and advance themselves without human involvement - which Anthropic warns could occur as soon as early 2027. While some safety researchers welcome the move, critics argue the voluntary framework is a form of regulatory capture designed to block open-source competitors, while others point to geopolitical competition with China as a major barrier to any real slowdown.
The voluntary agreement highlights growing industry panic over AI models achieving recursive self-improvement.
In their words
“The industry as a whole needs new champions.”
“Jacob is correct here - we really do earnestly believe AI could kill all humans!”
By the numbers
- 170M+
- Views on X for Jacob Coxon's resignation story
- 1,000+
- AI lab employees signing a July public letter
- 10%
- Evan Hubinger's estimated chance of AI killing everyone
How it unfolded
- Over 1,000 AI lab employees sign public letter.
- AI CEOs loosely agree to slow down development.
- Anthropic estimates recursive self-improvement could arrive.
Turn stories like this into views
Ravenclip finds the AI news, makes the video, and posts it before attention moves on.
Common questions
- What happened with Frontier AI leaders?
- Sam Altman, Dario Amodei, Demis Hassabis, and Elon Musk loosely agreed to slow down AI development.
- Where can I read the original report?
- Read the full report at The Verge.
