Want to dive into Scott Alexander's work and his thousands of blog posts? This fan website lets you sort and do semantic search through the whole codex. Enjoy!
| Date | ||
| Length | ||
| Likes | ||
| Comments |
Scott analyzes whether AI forecasters will plateau near human levels or keep improving, arguing that even non-miraculous progress could yield 4-12 percentage point improvements in prediction market accuracy.
Scott asks whether there's an accepted method for measuring calibration separately from domain knowledge, noting that common scoring rules conflate the two.