This is pretty crazy:
https://time.com/article/2026/03/11/anthropic-claude-disruptive-company-pentagon/
I'm sure there's a large amount of effort that goes into validation, but yikes. Meanwhile, AI from MS lags pretty far behind, so WF people are safe for a bit. Has anyone seen any actual AI agent automation being experimented with in any meaningful way at this point? If so, tell us what it's doing!