When AI builds itself, reading Anthropic’s lengthy official article was somewhat shocking. Authored jointly by Anthropic co-founder Jack Clark and head of internal research Marina, the piece for the first time publicly disclosed a set of previously confidential internal operational metrics: as of May 2026, over 80% of code merged into Anthropic’s codebase was written by Claude; compared to 2024, engineers’ daily code merge volume has increased eightfold; employees estimate their own output has quadrupled since using the latest model, Mythos Preview, compared to when no AI tools were used. More critically, Anthropic introduced a concept that has unsettled the entire AI industry: “recursive self-improvement”—the ability of an AI system to autonomously design and improve its successors without human intervention. This stage has not yet arrived, but “could occur within the next two years, or even sooner.” If a system can fully build its own successors, the ways we protect, monitor, and shape its behavior become vastly more important. The article’s core message is Anthropic’s urgent call to slow the pace of this technology’s development, buying more time to address its profound implications and allowing societal structures and alignment research to catch up with technological progress. Is Anthropic genuinely concerned—or is this merely a publicity stunt for Mythos? Or perhaps both? 🤔