Anthropic co-founder Jack Clark has pulled forward his forecast for when artificial intelligence systems could fully improve themselves. Where he previously anticipated full recursive self-improvement (RSI) arriving in late 2028, he now expects it in 2027.

Clark announced the revised timeline on October 12, 2026, in his Import AI newsletter. He did not cite a single breakthrough as the reason for the change. Instead, he pointed to a sustained run of benchmark data showing rapid gains in areas such as mathematics and automated research and development.

What Recursive Self-Improvement Means

In technical terms, full RSI refers to autonomous AI systems capable of enhancing their own design and capabilities without human intervention. Put simply, the model would no longer wait for people to build its next version — it would build that version itself. That definition is also why the timeline matters beyond Anthropic: at that point, the pace of AI capability gains would no longer be set by human research teams, which is the scenario Amodei's safety warnings address.

The Numbers Behind the Shorter Forecast

Several data points underpin the accelerated timeline. In August 2026, AI agents built on Anthropic's Claude model led 26% of core AI research and development tasks. More than 90% of the tasks in which the agents participated reached at least a collaborative level of involvement, classified as AL3 or higher.

According to the research findings, approximately 30,000 AI agents are running concurrently at Anthropic. Safety monitors overseeing those agents reportedly show low override rates.

Vals AI's RSI Index projects that full automation of machine learning engineering work at Anthropic could arrive around July to August 2027. Taken together, the internal research figures and the third-party index give readers two reference points for judging how the forecast holds up over the coming year.

A CEO Urging Caution

In September 2026, Anthropic CEO Dario Amodei published a post calling for frontier AI development to be paced, acknowledging that the risks tied to RSI are accelerating. Amodei has advocated for coordinated industry efforts to manage those risks. The sequence places Anthropic in a notable position: a co-founder at the lab is moving his RSI estimate closer in time while its chief executive asks the industry to pace itself.

Implications for the AI Industry

For policymakers, the shortened timeline compresses the window for action. Regulatory frameworks designed around a late-2028 arrival look different once the expected date shifts to 2027. Amodei's call for coordinated pacing is, in effect, an invitation for governments and industry to move before the forecasts become the present.

Several concrete indicators are worth monitoring from here. The first is whether the 26% share of agent-led R&D tasks continues to climb in Anthropic's future disclosures. The second is the Vals AI RSI Index, whose July-to-August 2027 estimate for full ML engineering automation at Anthropic stands as a dated, checkable prediction. The third is how override rates evolve as the agent count grows — approximately 30,000 concurrent agents operating with little human correction is a remarkable operational fact.