Overview
Few topics generate more confident predictions and less reliable data than the timeline to artificial general intelligence (AGI). This report doesn't pick a date — it synthesizes what surveys, scaling research, and forecasting history actually support, and explains why obsessing over the date is the wrong focus.
What the experts say
Large surveys of AI researchers tend to produce median estimates for human-level machine intelligence somewhere around the middle of the century — but with massive variance and frequent revision. The honest summary: experts strongly disagree, and aggregate medians have shifted as capabilities surprised people. Treat any single confident date with skepticism.
The definition problem
Half the debate is semantic. "AGI" means different things to different people — matching humans on most economic tasks, matching the best human on any task, or genuine general reasoning. Without a shared definition, timeline arguments often talk past each other. Progress is real and measurable; "AGI by year X" mostly isn't.
Scaling laws and their limits
Recent leaps came largely from scaling — more data, compute, and parameters produce predictable capability gains. The open question is how far scaling alone goes before hitting diminishing returns or needing new architectural ideas. Reasonable researchers disagree, which is precisely why timelines are uncertain.
Why forecasters keep missing
AI forecasting has a humbling record: experts have been too pessimistic on some milestones (Go, protein folding, language) and too optimistic on others (self-driving, "AGI soon"). Complex technological timelines resist precise prediction. Confidence intervals should be wide.
What this means for you
Don't plan your career or business around a specific AGI date. Plan around the trend that's already certain: increasingly capable narrow AI is reshaping work now. The people and companies that adapt to today's tools will be best positioned whenever — or whether — AGI arrives.
Honest limits
This is a forecast about forecasts. The only intellectually honest position is wide uncertainty. Anyone selling you a precise AGI date is selling confidence they don't have.
