Blog
Notes on trust, provenance and physical systems
Why verifying what an autonomous system actually did is harder than collecting data from it — and what that means for operators, regulators and the people building these systems.
-
Physical AI Is Coming to Manufacturing. Platform Governance Won't Be Enough.
Microsoft and NVIDIA are right: trust is the limiting factor for physical AI. But platform governance and independent verification are not the same thing.
Read -
AI Safety Laws Are Coming From Both Coasts. Here's What They Actually Require.
New York's RAISE Act and California's SB 53 create specific compliance obligations for AI developers — safety protocols, independent audits, incident reporting, and multi-year data retention. Here's what the laws require and how to build for them.
Read -
When the Regulator Calls: Lessons from Tesla's FSD Data Challenge
Tesla's struggle to meet NHTSA data deadlines highlights a challenge every AV company will face: the gap between collecting data and producing it when regulators come asking.
Read -
New York Just Killed Its Robotaxi Plan. The Real Problem Isn't the Technology.
New York shelved its robotaxi expansion — not because the tech doesn't work, but because there's no way to independently verify that autonomous systems are operating safely.
Read