Anthropic
Nov 2, 2026

Anthropic says it may be nearing recursive self-improvement threshold

Anthropic said in a revised S-1A filing that it may be close to creating models capable of recursive self-improvement, or RSI. The company said its internal assessments and expectations for future models suggest it may be nearing the threshold in its Responsible Scaling Policy that it defines as a "sharp acceleration in AI-driven progress." At the same time, Anthropic said no existing model has crossed that threshold so far. The filing also warned that using AI to help build more powerful AI systems could create major safety risks, especially if the pace of progress accelerates sharply. In response, the company said it is considering updates to its Responsible Scaling Policy to make sure humans remain in control of frontier model development for the foreseeable future. The item was cited by Techub News and attributed to TechCrunch.

760
Anthropic says it may be nearing recursive self-improvement threshold