Critics slam Anthropic for backing bill that lets AI labs define their own safety thresholds
HankYeomans · x · 2026-10-18
Reposting Zack Korman's critique: Anthropic warns AI could cause human extinction while backing legislation that effectively says "let the labs decide what counts as safe." The bill text (SEC. 4) requires large frontier developers to write, implement, and publish their own safety framework within 90 days — including self-set capability evaluation thresholds and mitigations — with the developer itself reviewing results to decide whether to deploy frontier models. Critics argue this hands safety discretion entirely to the labs being regulated.
More from Companies & People
- Oracle Health data breach exposes medical records of nearly 20 million patients — ChuckDBrooks · 2026-10-18
- X Records 13.43B Website Visits in Q3 2026, Up 6% Quarter-over-Quarter — XFreeze · 2026-10-18
- Dongyi Robotics shows 100x efficiency gains with 4-layer embodied AI stack at Canton Fair — 智东西 · 2026-10-18
- Trend analysis: Anthropic's junior technical hires projected to hit 0% by mid-2027 — Confident_Salt_8108 · 2026-10-18
- A Data Scientist's 10-Book AI Learning Roadmap: Where to Start Based on Your Background — mdancho84 · 2026-10-18
- 10 books to learn AI engineering, ranked: start with Automate the Boring Stuff — mdancho84 · 2026-10-18