Urgent.News

What's breaking now, across thousands of outlets.

More in AI

My Agent Found Real Improvements. The Statistics Still Killed the Promotion.

Previously: 9 Bugs That All Looked Like a Working System · I Built an AI That Rewrites Its Own Prompts · The Edit That Fixed 4 Tasks and Broke 1 · The Gate Is the Product · The Doctor Who Diagnosed…

  • Initial edit fixed 4 out of 26 tasks, resulting in +3 net improvement
  • Expanded corpus to 40 tasks and improved analyzer, but edit only fixed 1-2 tasks
  • Statistical power limited despite more tasks, gate rejected edit due to small effect size

More from Wednesday 2 September →