The generative market hypothesis: price as a residue
Efficient-market theory treats price as a summary of information; the generative view treats it as the residue of interaction. Three cases where the model was right and the world had changed.
Notes
Writing it down
Writing it down
is what makes an idea refutable
Essays carry a full argument; fragments record intuitions that have not formed yet. Everything is open to read — if a judgment does not hold, write and say so.
Long form
Four arguments taking shape
The Edge: A Gap Left by Partial Observability
Translating “edge and generation” into a problem that already has a name: the state is hidden, so you maintain a belief; when the belief moves, the policy is reset. With an architecture figure anchoring each of four concepts to a mathematical object.
The Wind-Change Index: reading uncertainty as a number
Volatility describes how hard prices shake, not whether the structure is being replaced. Entropy, dispersion and network reconfiguration — plus the three cases where the index fails.
Why I stopped selling forecasts
Consulting sells certainty; clients need an order of action under uncertainty. After one workshop I replaced the whole deliverable — including the model I had built for six years.
Fragments
The parts not yet formed
Short, and not necessarily right. Their job is to mark what I have not worked out.
2026 / 11
A metric is a promise
Once an indicator enters performance review, it starts producing the behavior it measures. If the Wind-Change Index became a KPI, it would probably stop working within two years. A tool for uncertainty must not promise certainty.
2026 / 10
A model earns its keep by being disproved
People ask whether the model is accurate. I care whether it makes a specific, falsifiable claim that gets refuted within three months. A model that can be refuted gets used again.
2026 / 09
The question chain of an agent
A good agent does not answer faster; it breaks a large problem into verifiable smaller ones and changes route after being refuted. On that measure, human researchers are still the better teachers.
Discuss
Disagree with one of these claims?
Write about what you are working on and where it jams. I reply to problems I can push forward; when I cannot, I will say why.