See my comment here (on titotal’s post) for some concrete evidence that the slow progress scenario is too high as a baseline, minimum progress scenario.
For example, the slow progress scenario predicts the development of household robots that can do various chores by December 2030. Metaculus, which tends to be highly aggressive and optimistic in its forecasts of AI capabilities progress, only predicts the development of such robots in mid-2032. To me, this indicates the slow progress scenario stipulates too much progress.
Metaculus is already aggressive, and the slow progress scenario seems to be more aggressive — at least on household robots, depending how exactly you interpret it — than Metaculus.
See my comment here (on titotal’s post) for some concrete evidence that the slow progress scenario is too high as a baseline, minimum progress scenario.
For example, the slow progress scenario predicts the development of household robots that can do various chores by December 2030. Metaculus, which tends to be highly aggressive and optimistic in its forecasts of AI capabilities progress, only predicts the development of such robots in mid-2032. To me, this indicates the slow progress scenario stipulates too much progress.
Metaculus is already aggressive, and the slow progress scenario seems to be more aggressive — at least on household robots, depending how exactly you interpret it — than Metaculus.