The problem I chose

I work on responsibility architecture for agent systems: the question of who answers for an effect that cannot be undone, and how to enforce that answer at the speed agents actually run. Everything on this site — the risk branch, the speed branch, the running implementation — is one problem seen from different sides.

I chose it because it is the kind of problem I keep returning to: systems where the interesting part is not making something possible but making it accountable — where the hard question is not "can it act" but "who signed for the consequence." The essays show how I think about it; the public repository behind this site shows how I work.

What I just presented is how I think. If your organization needs that way of thinking, here is the way to reach me.