Knowledge
Essays on evidence, accountability, and Verifiable AI.
Arguments we are willing to defend, on what AI accountability actually requires and why most of the tooling around it answers a different question.
-
May 9, 2026
The continuous compliance shift: why annual paperwork is dying
SOC 2 was designed for an annual rhythm. Buyers, regulators, and scoring tools have all moved to continuous. The audit calendar isn't going away. It's becoming the floor.
-
May 8, 2026
What buyers actually buy when they buy compliance software
Nobody writes a check for SOC 2 software because they want SOC 2. The artifact is evidence. The decision is the product. Compliance vendors that miss this ship the wrong thing.
-
May 7, 2026
Hallucination is a category error, not a tuning problem
The AI industry treats hallucination as a knob to turn. We think it's a wall to redesign around. Here's why the framing matters.
-
May 6, 2026
What Verifiable AI actually means
We build what we call Verifiable AI. Here's what the term means, what it isn't, and why an AI you can't check is an AI you can't trust.