Hacker News · AI· yurivish·· 37 min agoSelectedAI score65
METR Report Analysis: AI R&D Is Sharply Accelerating, Anthropic May Have Hit RSP Risk Threshold
AI is dramatically accelerating AI research
AI guide
An in-depth investigation into Anthropic by METR reveals that AI-automated R&D has nearly doubled the rate of its model capability gains over the past year, potentially crossing the 'sharp acceleration' threshold outlined in its Responsible Scaling Policy (RSP).
Why it matters
Combining the investigation report with economic modeling, the article quantitatively breaks down the multiplier effect of AI automation on frontier R&D efficiency as well as potential safety inflection points.
Source:Hacker News · AI · planned-obsolescence.org