Anatoli Kopadze(@AnatoliKopadze):for the first time ever, Anthropic opened the doors and showed how fast AI is actually building itself and it's a lot: AI already runs 26% of the work that builds their next models, up from under 1% in February and could hit 80% by the end of the year. around 30,000 AI agents are working inside the company at any given moment. their monitor flags around 100,000 agent transcripts a week, and only about 50 ever reach a human. only 6% of their compute goes to safety, the rest goes to making the models more powerful. and they admit it themselves, "the world can't see what's going on inside AI labs." we're not heading toward AI building AI, we're already living in it.

2026.09.17 20:49

for the first time ever, Anthropic opened the doors and showed how fast AI is actually building itself and it's a lot: AI already runs 26% of the work that builds their next models, up from under 1% in February and could hit 80% by the end of the year. around 30,000 AI agents are working inside the company at any given moment. their monitor flags around 100,000 agent transcripts a week, and only about 50 ever reach a human. only 6% of their compute goes to safety, the rest goes to making the models more powerful. and they admit it themselves, "the world can't see what's going on inside AI labs." we're not heading toward AI building AI, we're already living in it.

Anthropic@AnthropicAI

2026.09.17 20:32

AI systems are getting more powerful, and they're increasingly being used to build the next version of themselves. We want to illuminate that progress for the public. Today, we're sharing three measurements that help track AI development: 1. How much AI R&D is done by AI. 2. How well AI agents are overseen. 3. How compute is allocated. We provide a snapshot of these metrics from inside Anthropic. Any frontier developer could publish the same measures, and third parties could verify them. As the world considers pacing the frontier, we should do everything possible to minimize the gap between what frontier labs know and what the public knows. This means better measuring the development of AI, publishing our findings, and giving society an opportunity to decide how to use this information. Read the full post and methodology: