Angelina YangFounder and Host, TwoSetAI · AI builder for visibility and discoverability

100 Hours While We Talk

25 Jun 2026 · 1 min

Watch on YouTube

0:41

An autonomous agent runs a 10-minute inspect-and-fix loop continuously. It's already completed ten cycles and accumulated 100 hours of runtime in three days, with only high-level objectives: higher performance, faster speed, lower cost, no over-engineering. The target is twenty cycles minimum.

Chapters

Just started a new eval with cycle 10, it's called. Wait 10 minutes, and in 10 minutes, inspect the results, see what are the biggest problems, fix those problems, start a new run, and 10 minutes later, inspect again and keep going. And you see, while we are having right now this interview, this agent is already in cycle 10. While we are right now talking, this agent is working for me. The runtime is like almost 100 hours of this agent. Say, it has run over the entire last 3 days. I only define the high-level goals. Like I tell, I want to have my agent higher score. I want to have I want to have it faster and I want to have it cheaper. Please don't over-engineer, like don't overfit on specific tasks. And I just say go and do like at least 20 cycles.