Hacker News
Dan Luu Scored Roughly 30 of Ed Zitron's Falsifiable AI Predictions and Found Almost None That Held
Dan Luu published a September 1, 2026 audit of the AI skeptic's track record, filtering Zitron's writing down to roughly 30 falsifiable claims with source links and discarding tautological or endpoint-free statements. The misses include "we're reaching upper limits of what generative AI can do" (February 2024), calling OpenAI's revenue forecasts egregious (October 2024, forecasts were exceeded), calling a 500M Gemini user target so unrealistic someone should be fired (February 2025, it reached 750M), and "models have clearly hit a wall" (July 2025). Luu used ChatGPT to assemble the prediction list, then verified each against the original posts and had both ChatGPT and Claude fact-check the piece.
↳ Follow the thread