Forecasting AGI

For AGI to do most human work for <$25/hr by 2043, many things must happen.

Use this interactive web app to forecast individual events and see what they imply about the likelihood of transformative AGI by 2043.

We invent algorithms for transformative AGI 60%

We invent a way for AGIs to learn faster than humans 40%

Compute needed by AGI, relative to a human brain (1e20-1e21 FLOPS)

Compute cost efficiency, relative to 2023 (4e14 FLOPS/$/hr)

AGI inference costs drop below $25/hr (determined by above) 16%

We invent and scale cheap, quality robots 60%

We massively scale production of chips and power 46%

We avoid derailment by human regulation 70%

We avoid derailment by AI-caused delay 90%

We avoid derailment from wars (e.g., China invades Taiwan) 70%

We avoid derailment from pandemics 90%

We avoid derailment from severe depressions 95%

Probability of transformative AGI by 2043: 0.4%