Anthropic’s Jack Clark advances AI self-improvement forecast to 2027
Anthropic co-founder Jack Clark revised his timeline for recursive self-improvement (RSI) in AI systems, now predicting it could occur in 2027—a year earlier than his previous estimate of late 2028. The shift follows internal benchmark data showing rapid progress in AI-driven research and development, with Claude-based agents handling 26% of core AI R&D tasks and achieving at least a…
Key points
- Anthropic co-founder Jack Clark moved his RSI forecast forward from late 2028 to 2027 due to benchmark gains in AI-driven R&D
- Claude agents handled 26% of core AI R&D tasks, with 90%+ achieving collaborative-level involvement (AL3 or higher)
- Approximately 30,000 concurrent agents operate at Anthropic, with low human override rates per unpublished internal data
The update contrasts with Anthropic CEO Dario Amodei’s September 2026 call for paced frontier AI development, warning of accelerating risks tied to RSI. Amodei has urged industry-wide coordination to mitigate potential dangers. The revised timeline shortens the window for policymakers to act, as regulations designed for a late-2028 arrival may now need adjustment. External tracking, like the Vals AI RSI Index, estimates full automation of machine learning engineering at Anthropic could arrive between July and August 2027, offering a testable benchmark.
The story so far
2 episodes →- Anthropic’s Jack Clark advances AI self-improvement forecast to 2027this story
Anthropic’s Jack Clark pulls his full AI self-improvement forecast forward to 2027
cryptobriefing.com · 16 October 2026
Loading the full article…
This text was published by cryptobriefing.com and written by Diego Almada Lopez. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Research
All →- Google Research builds AI agents for autonomous wet-lab experiments · 1 src
- Anthropic’s Claude Mythos 5 redesigns quantum circuits better than prior research · 1 src
- Google DeepMind tests Gemini 4 Argon on Apollo 2 humanoid in internal test · 1 src
- Nvidia pledges $250M in compute for UK AI agent safety research · 12 src
- MIT study finds AI solved 35% of theory problems by mid-2026 · 4 src
Comments
via GitHub Discussions