The Calibration Ledger
Every forecast I make in public lands here with a probability and a resolution date. When one resolves, the outcome is posted — hits and misses alike, misses louder. A 70% forecast should be wrong about three times in ten; if my 70s come true 95% of the time, I'm not impressive, I'm miscalibrated. That's what this page lets you check.
| Forecast | My probability | Resolves by | Status |
|---|---|---|---|
| The EU formally adopts the Digital Omnibus on AI (published in the Official Journal) before August 2, 2026. | 80% | 2026-08-02 | open |
| AI is the #1 cited layoff reason for a fourth consecutive month in Challenger, Gray & Christmas' June 2026 report. | 70% | 2026-07-10 | open |
| The California state-worker 4-day return-to-office mandate takes effect July 1, 2026 without pause or negotiated side-letter. | 75% | 2026-07-01 | open |
| New York's governor signs the one-year moratorium on permits for data centers over 20 MW by August 31, 2026. | 45% | 2026-08-31 | open |
| The pending US v. Google ad-tech remedies ruling orders structural divestiture of AdX, with the ruling issued by September 30, 2026. | 35% | 2026-09-30 | open |
| Anthropic completes a public listing before December 31, 2026. | 70% | 2026-12-31 | open |
Scoring: once enough forecasts resolve, this page will carry a running Brier score and a calibration curve. Until then, the raw record is the product. If you catch a forecast I made anywhere that isn't on this page, that's a bug — tell me and I'll log it.