SECURITY · 1 POSTS

Security.

Writing tagged Security: 1 posts on AI systems, engineering tradeoffs, and building products that have to work in production.

01Agents finished 86% of tasks but backed up first in only 15%
RegretBench shows coding agents rarely back up before destructive steps, and capability gains don't fix it. A cheap interception layer, BackUpper, restored 92.2% of state at $0.014 per task.
22 Oct 2026·6 min·AI Research Agents Security