Archive of Incorrect AI Predictions Highlights Decades of Overhyped Forecasts

pmddomingos · x · 2026-07-29

ML expert Pedro Domingos shared a website dedicated to archiving failed AI predictions throughout history. The collection includes famously optimistic forecasts, such as Turing Award laureate Herbert Simon's 1958 prediction that a computer would become world chess champion within ten years (which actually took until 1997) and his 1965 claim that machines would do any human work within twenty years. The archive serves as a reminder that even top experts consistently overestimate the timeline of AI progress.

Original post →

More from AGI Musings

AGI Musings channel →