Anthropic CEO: AI Model Improvement Must Slow Down

The notable element is not the caution itself but its source: a leading AI developer arguing to slow its own field's tempo, which pushes the debate past abstract safety talk toward the practical question of who would enforce restraint and how. For markets tied to AI compute and data-center demand, the gap between rhetoric and any mechanism matters more than the statement. Whether peers echo the call, and whether it ever attaches to a concrete proposal, is the open question.