Predictions

• August 4th, 2026: US real GDP growth from 2026-2031 will not average 4% annually.

  That is, there may be an AI construction boom, there may be some productivity growth from AI, but it will not be a radical change. AI will continue to be impressive, but actual economically impactful deployment will be gradual and not drive such shocking increases in the medium-term.

  Note: I originally made this prediction on Discord, where my prediction was a point estimate of 2.1% growth. I think that point estimate is reasonable, but it has huge error bars. I put at least a 1% chance on a major recession happening which could lead to 0% growth over that period. On the other side, 3% growth would be a mild surprise, but I wouldn't consider it shocking. So the one sided "no 4% growth prediction" is easier to make precise and testable.

  From the perspective of mainstream economics, this is a very boring prediction, but I made it in response to a online poll showing that SF tech folks are predicting 4.7% annual growth over the next five years. Manifold markets have a 10% probability world GDP doubles in the next 10 years (not equivalent, but it suggests much more rapid US growth than the status quo, which is the 2.0-2.5% range). So perhaps this prediction is shooting fish in a barrel, but they are real fish, not straw fish. (Note: I made this prediction to friends on Discord on August 4th, 2026. I posted it to the website sometime in early September).

• September 30th, 2026: We will not see major disemployment in the United States or other advanced economies because of AI in the next 3 years. This doesn't mean that there won't be specific jobs that take a big hit, but it does imply that there will be enough demand for people working that the affected people are able to find new jobs. I'd like to think more about how to make this precise and testable. The obvious thing to do is to make a prediction about the peak age labor force participation rate. But doing that requires thinking a little more, because those rates are affected by economic conditions and cohort effects. Certainly if that rate drops 5%, I'm wrong unless we're in a great depression, but I could probably make the prediction tighter.