48–50 Future & Frontier

48 AI AGI & ASI

Daily AI News Updates on AI AGI & ASI: project 48 of 50, in the Future & Frontier group (48–50).

Latest updates · 21 October 2026

An Anthropic co-founder told the IMF that the singularity has begun: Jack Clark's Per Jacobsson lecture in Bangkok argued that AI is starting to speed up science itself.

Research, industry & policy

Import AI 477: In which I tell the IMF the Singularity has begun, Jack Clark, Import AI · 19 Oct 2026: The text of Clark's 17 October lecture cites Humanity's Last Exam scores rising from 8% to 67.7% and FrontierMath from 2% to 94%, and calls for rules, oversight and restraint.

Journals & preprints

Nested Shared-Operator Recursion on ARC-AGI-1: An Empirical Study of HRM, TRM and URM, arXiv 2610.26666 · 19 Oct 2026: Recursive reasoning models refine persistent internal states by repeatedly applying learned operators.

Why it matters

Claims that AI now accelerates research itself shift the AGI debate from timelines to governance of the transition.

Poster: AI AGI & ASI, 21 October 2026 — Download poster

Events ahead

5 upcoming events for AI AGI & ASI.

See events for 48

All issues for AI AGI & ASI

21 October 2026: An Anthropic co-founder told the IMF that the singularity has begun: Jack Clark's Per Jacobsson lecture in Bangkok argued that AI is starting to speed up science itself.
20 October 2026: A Times of India essay argues that 'superintelligence' is not yet science, three weeks after a US executive order told federal departments to stop saying 'artificial…
19 October 2026: AGI research this week looked at self-improving orchestration and AI reaching the limits of computability, while the US task force on superintelligence began meeting.
18 October 2026: Superintelligence has moved into policy documents: Nature examines its science, and the White House has issued an order on the 'era of super intelligence'. There are no…