← Research & Proposals
Active

AI Futures Simulator

Updated 2026-06-13

Status
Active — maintained, real outputs
Method
Monte Carlo, 3 seeds × 400 runs
Horizon
2026–2050
Code
Public

A joint Monte Carlo simulator for AI-related futures from 2026 to 2050: a capability growth spine, discrete plot events, and a society-feedback loop that compound into emergent outcome regions — doom, utopia, friction, and severe-but-recoverable — rather than a single hand-picked forecast.

The headline numbers are emergent, not hand-tuned: across 3 seeds × 400 runs, the simulator lands at roughly 7% doom, 18% utopia, 69% friction, and 6% severe outcomes.

Every event's cumulative probability in the model's configuration traces back to a specific claim, a stated reason, and cited evidence — the simulator is a companion to a public forecast essay and an accompanying evidence index, not a black box producing a number.