Rsi

Anthropic warned it is close to RSI (recursive self-improvement), the point where AI becomes capable of training itself well enough to trigger rapid capability acceleration. The company framed the milestone as a threshold with major safety implications.

AI Safety • TechCrunch AI

Goldman Sachs is pulling AI coding tools including Claude Code and GitHub Copilot from traders and bankers, restricting access to engineers. The bank cited productivity and risk management, reversing its earlier public praise of 'vibe coding' by non-technical staff.

Enterprise AI • The Decoder

SemiAnalysis launched GDPx, an agentic capability benchmark spanning 11 domains including chip design, cybersecurity, finance, legal and RSI. Opus 5.5 topped the leaderboard with a 12% score, per SemiAnalysis.

Benchmarks • Techmeme

Productivity

Simon Willison reports occasional confusing Python dependency errors after switching to Python 3.15, where a leftover uv.lock file pins a dependency version that lacks a Python 3.15 wheel (example: watchfiles==1.1.1 failing to build because PyO3's maximum supported version is 3.14).

A Brown University undergraduate has mapped the hidden global labor force that powers AI, revealing that many workers in the Global South are paid low wages for the work that enables algorithmic systems. The research highlights the often-unseen human cost behind AI.

AI and Society • Hacker News

adsilico • Series A

University of Leeds spinout adsilico has raised £10 million in a Series A round to expand its AI-powered simulation technology for assessing medical device performance across diverse virtual patients.

AMZN

VCs are set to make billions on Anthropic's IPO, but Amazon is positioned to make much more, according to TechCrunch. Massive note conversions reveal the true extent of Amazon's ownership in Anthropic.