Could come sooner than most institutions are prepared for
25 October 2026 (The Decoder). Anthropic wrote in June that recursive self-improvement in AI has not yet been achieved but could arrive sooner than expected.
Anthropic co-founder Jack Clark moves his prediction for self-improving AI forward to 2027
AI and safety · AI and science · 2026
Who else?
What it refers to: Recursive self-improvement, AI systems that develop their own successors without human oversight.
Other voices
- “This dynamic is called recursive self-improvement, and it is starting to happen” Dario Amodei, Builder,
- “We are pretty far from the end state of RSI” Sundar Pichai, Builder,
- “When the teacher becomes the student” Jack Clark, Builder,
- “The probability of catastrophe is higher, and the probability of dystopia is very high” Gary Marcus, Researcher,










