Debate: Should Kurzweil's 1999 AGI Prediction Count? Scaling Believers vs Prediction Skeptics

teortaxesTex · x · 2026-09-06

A pointed exchange on the epistemic value of early AGI predictions:

@phl43 argues that even if AGI had already arrived, someone making a confident AGI-now prediction back in 1999 shouldn't be taken seriously — the puzzle pieces making AGI plausible only fell into place much later, so early predictors couldn't have had a good argument. Prediction, he says, is overrated as an epistemic benchmark: it's easy to be right for the wrong reason.
teortaxesTex counters that this kind of 'clear-headed thinking' is why the French lag the Anglos at producing paradigm shifts: Kurzweil, Legg, Vinge, Moravec, and Butler all made bone-headedly simple arguments — the line goes up, mechanism be damned — and the naive extrapolation won.

A collision between 'predictions need mechanism' and 'long-horizon simple bets have their own value'.

More from AGI Musings