One week in AI: RSI goes mainstream, Anthropic IPO filing touts model-built-model loop
DavidSKrueger · x · 2026-11-04
A CeSIA newsletter recap of a single whirlwind week in AI:
- RSI is now the plan: Google's CEO told investors to expect "continued acceleration because of RSI" over the next twelve months — turning what was once a red line into official strategy.
- Anthropic filed to go public: the filing describes each Claude model helping build the next one, a process it calls "closing the loop" — what researchers call recursive self-improvement, which Yoshua Bengio has argued should be illegal.
- MIT researchers had GPT-6 Astra perform "brain surgery" on a smaller model: changing just 0.016% of weights yielded a 6x improvement on a math benchmark.
More from AGI Musings
- Don't Tie Your Identity to Your Job Title—AI Will Reshape It Within Years — KevinNaughtonJr · 2026-11-04
- Canva restructuring hits 1,271 staff as nearly a quarter face layoffs or redeployment — CtrlAltDwayne · 2026-11-04
- Anthropic's RSP update says it may 'imminently' hit AI-accelerated R&D threshold — thlarsen · 2026-11-04
- As new code cost nears zero, engineers stop collaborating and rebuild the same systems — alex_frantic · 2026-11-04
- Domingos: Gemini itself agrees CoT and RL are a wasteful detour for reasoning — pmddomingos · 2026-11-04
- Anthropic's AGI threshold: AI substituting all researchers or doubling R&D pace — DKokotajlo · 2026-11-04