This note documents the July 2026 reorganization of the site: what changed, where the existing research went, and what the site is oriented around going forward. It supersedes the May 2026 coverage-transition note, including its plan to resume curated equity updates in mid-June — that resumption did not happen, for reasons recorded below.
What Changed
- All research published through mid-2026 — the equity analyses, forecast markets, macro themes, sector deep-dives, play portfolio, and framework documentation — is now the research archive. Nothing was deleted; every page remains at its original URL. Archived sections carry a banner noting they are no longer updated.
- The navigation was reduced to the archive and the site's standing pages. The news feed remains browsable but is no longer surfaced on the homepage.
- The homepage now describes the site's current orientation: forecasting work, kept on the record.
Why
The May plan assumed the right response to rising analysis costs was a smaller version of the same thing: fewer tickers, same broad-coverage model. The pause made it clear the coverage model itself was the problem. Company fundamentals are already analyzed exhaustively by people with more resources and better information access; another set of coverage-style writeups adds little, however rigorous the process behind them.
What the first era actually produced of durable value was its forecasting discipline: questions with explicit resolution criteria, probabilities committed in advance, and Brier scores recorded when outcomes arrived — right or wrong. The reorganization makes that discipline the center of the site rather than a byproduct of equity coverage. The underlying motivation is simple: trying to be less wrong about the world, and using a public record as the mechanism that forces honesty about how that attempt is going.
Current Orientation
Three areas of work, in varying stages of build-out:
- Geopolitical forecasting. Situation-based coverage of developing geopolitical events: the key inputs, the probabilistic forecasts they drive, and updates as facts change.
- Congressional forecasting. Probabilistic forecasts on legislative outcomes — which bills advance, which stall — built on continuous tracking of bills, votes, and member behavior.
- AI buildout analysis. The AI buildout treated as a system — the semiconductor complex, capability trajectories, disruption dynamics, and models of criticality and bubble formation that make certain moves predictable even when their timing is not.
Alongside these, methodology write-ups will document how the forecasting system works — the reasoning chain and the decisions it makes in taking a question from framing to a stated probability — and how its forecasts score over time.
What the Archive Contains
The archive preserves methodology era 1 in full: 260+ multi-lens equity analyses, 2,300+ forecast markets (including all resolutions and calibration scores), the macro conditional-market themes, the sector deep-dives, the play portfolio, and the lens/signal/persona framework documentation. The resolved markets are the part with continuing relevance — a timestamped record of how the prediction process performed, which the current work builds on.
Browse the Research Archive