Predictions For 2018

The annual 100-prediction calibration exercise for 2018 — probability estimates on US politics, economics/tech, culture war, the rationalist community, and personal life (many redacted as 'secret'), made without checking prediction markets so as to test his own calibration, scored at year-end. A bare list with minimal prose, but one instance of a valuable, durable public-calibration practice.

Quality 48 · Solid. 48 — Solid floor / Minor edge. As a standalone artifact it's essentially a list (like a survey/index stub) with almost no argument, so it scores low; kept just in Solid rather than Minor because the calibration practice it embodies is genuinely worthwhile. Firm — no creep for the practice's reputation.

Claude’s paradigm shift 35 · Slight. 35 — Slight. Public scored annual predictions weren't novel (Tetlock, etc.); this instance adds no new idea.

Real-world impact 2 · Minor. 2 — minor/within-blog. One year's list; the broader annual-predictions practice has some niche-community influence, but this artifact's own reach is contained.