Anthropic CEO Dario Amodei: AI is on an exponential curve, policy must catch up now

Policy on the AI Exponential

Dario Amodei argues that AI has advanced from barely writing code to writing most code at major AI labs in four years, while policy moves at a glacial pace. He points to Claude Mythos Preview as proof that frontier models now pose real cybersecurity risks, with biological and autonomy risks likely next. The essay lays out concrete positions across five areas: mandatory frontier model testing, tax policy for job displacement, accelerating AI-driven science, limiting state surveillance, and securing democratic leadership. Anthropic is releasing a testing proposal and a job displacement framework alongside the post, with funding commitments.

Why it matters: Dario Amodei's policy essay is a same-day must-read: CEO-level primary source, first public confirmation of Mythos Preview's security risk tier, and a concrete four-year capability arc. Not a 95 because it's a framework piece rather than a product launch or hard-data report.