The work was the cost side of operating companies. Workforce models and capacity planning. Offshoring delivery to move margin. Pricing, from ad-hoc quotes to a standardized process. Acquisition integrations, including the unglamorous parts, chart-of-accounts transitions and scrubbing an acquired company's forecast until the synergies were real rather than assumed.

What that work has in common is that almost all of it was people reading documents and keying what they said. Every efficiency I found eventually ran into the same wall: you can make a person faster, and you cannot make them unnecessary, because the tooling always left a human in the seat.

I studied economics at Cornell and played football there.

Two systems, built to remove that wall rather than work around it.

Opus compiles a described workflow into a deterministic artifact that runs with no model in the loop, so the same input produces byte-identical output every time and every past decision can be replayed and proven. Probity governs the actions that do need a model, evaluating each one against policy while it executes and producing a tamper-evident record that can be reconstructed later from storage alone.

Together they are the foundation of Parity, which is what I work on now.

The papers came out of building this, not the other way around. They exist because the claims are the kind nobody should take on trust: that a system is deterministic, that it never fails open, that a decision from months ago can be reproduced exactly. Each one is a controlled study with the run counts stated and the failures reported alongside the results.

If you want the short version, the compiled execution study is the one that matters most.