The AI Race: Who Hits the Brakes First?

On September 12, Anthropic CEO Dario Amodei published an essay titled “We Must Pace the Frontier.”

The idea wasn’t to stop AI development, but to give safety, alignment and independent evaluation more time to keep up with rapidly advancing capabilities.

Sam Altman and Elon Musk both nodded in agreement. (A rare sight, but yes, it happened.)

And then came the next 10 days.

September 21: xAI released Grok 4.7.

A new, larger base model and xAI’s most capable model for coding and knowledge work, designed to work longer on difficult tasks and verify its own work more carefully.

September 22: OpenAI released GPT-6 Sol and GPT-6 Luna.

Sol brings GPT-6 intelligence to complex coding and agentic workflows, while Luna pushes that capability toward high-volume work at dramatically lower cost.

September 22: Anthropic released Claude Opus 5.5.

It delivers roughly Claude Fable 5.1-level performance on most work while costing, according to Anthropic, 40% less to run than Opus 5.

Turns out everyone agrees that the race should slow down.

Nobody just wants to be the first one to hit the brakes.

References

• Dario Amodei: We Must Pace the Frontier
• The Guardian: Anthropic CEO calls for an AI slowdown (12 September 2026)
• ITV News: Musk and Altman back calls to slow AI development (12 September 2026)
• Reuters: Anthropic unveils Claude Opus 5.5 (22 September 2026)
• Reuters: OpenAI expands GPT-6 lineup with Sol and Luna (22 September 2026)