This page is an artificial intelligence reading of the archive, not a conventional About page written by me. It reads the current 15 posts, their titles, dates, themes, summaries, and keywords, then rebuilds the sketch from the content it can actually point to.
What The Archive Suggests
Based on the current posts, the machine would probably describe the person behind the writing as:
- A data-and-geometry person who would rather keep one value that answers the question than maintain a beautiful dashboard that avoids it.
- Someone who thinks in probability, tail events, insurance against downside, and deliberately manufactured chances at upside.
- A behaviour mechanics person who trusts simple reinforcement rules more than vibes when thinking about people, products, and model behaviour.
- A physical-reality person who thinks a model, market, or theorem should eventually answer to something measurable outside the room.
- A maths-adjacent speculator who is comfortable asking whether the question is badly framed when a famous problem refuses to move.
- Someone who keeps poking at hidden assumptions inside symbols: what counts as a unit, where type enters, and what gets lost when nouns become numbers.
- A local-intelligence person who wants useful models to accumulate private nuance under operator control instead of renting competence by the token forever.
What Keeps Turning Up
If you scan the archive as evidence, these themes keep turning up:
- Values over chart clutter: The archive keeps favouring geometry, compression, and compact signals over sprawling displays of technically correct noise.
- Asymmetry and exposure: Several posts ask what happens when small non-zero chances compound over time, and why exposure can matter more than prediction.
- Behavioural lenses: Operant conditioning, social habits, and conditional activation keep reappearing as practical tools for describing how systems actually move.
- Physical reality checks: Climate numbers, housing costs, mining odds, medical data, and cake habits all serve the same role: they force abstractions back into contact with lived constraints.
- Reframing hard problems: The mathematical posts tend to ask whether a problem is difficult because of the object itself or because of the projection being used.
- Typed abstraction: The archive has a recurring suspicion that symbols become dangerous when they forget the nouns, identities, and boundaries they compressed.
- Owned resident intelligence: The artificial intelligence thread is not just about smaller models. It is about local learning, operator control, privacy, and durable fit to a real environment.
How This About Page Works
The input is the posts. Not private context, not a hand-written author bio, not a fixed statement of intent. The output is a rolling interpretation of the person and preoccupations implied by the writing so far.
As new posts are added to the catalogue, this page is regenerated with the rest of the site. If the writing changes direction, the About page should change direction with it.
In other words: this is not the definitive story of me. It is the current reading of the work.
Good starting points are “The Nibbler Doctrine”, “The Principia Trap”, “The Probability Asymmetry of Life”.