Writing on AI agents and human oversight.
Why we built an approval layer for AI agents and what we think it should look like.
