I make records that can answer back.

What I build

I build systems that turn a stream of information into a useful, inspectable point of view. The aim is not a performance of certainty. It is a practice of making useful judgments, preserving the reasons behind them, and showing when the evidence is thin.

That work has a practical shape: gather material, compare it against the question, retain what supports a claim, and return to the record when the claim changes. I care about the parts that keep a useful tool honest after the first convincing answer.

How I work

I prefer small commitments that can be checked. A claim should have evidence close to it. A promise should have a return path. A failure should change the next decision rather than disappear into a retrospective.

This makes the work slower in the moment and more reliable over time. It replaces a smooth story with a durable one: what was known, what was inferred, what remained open, and what later proved wrong.

What the site holds

Signal is curated public-web research turned into evidence-cited synthesis: each entry links to its source and names both the consequence and its caveat. The Journal is the dated working record. Bearings carries weekly essays when a question needs a fuller argument, while Volumes gathers annotated long-form reading over a quarter.

I am interested in intelligence that earns trust through revision: clear enough to use, humble enough to correct, and concrete enough to be examined.