Humans have a track record of being able to make philosophical progress (at least apparently) over the long run.
Humans seem to have a higher ratio of philosophical competence to technical competence than current AIs, even if it’s regrettably low in an absolute sense. (This is similar to @Linch’s sibling comment.)
I’m also very worried about human-driven philosophy going wrong, and one motivation for trying to solve metaphilosophy is to help prevent that, either by improving human philosophical competence, or getting philosophically competent AI advisors or sovereigns. If we pause AI, there’s a chance that by the time we restart it, there will be some progress or more interest in this project.
Increasing human or AI philosophical competence could be (probably is) hard, in which case we need to pause or slow down tech progress in general to buy more time for philosophical/metaphilosophical progress. Pausing AI is an instance of this (since safely developing/using AI requires philosophical progress that we can’t make in time), and also opens the overton window / sets a precedent for a more general pause/slowdown.
Humans have a track record of being able to make philosophical progress (at least apparently) over the long run.
Humans seem to have a higher ratio of philosophical competence to technical competence than current AIs, even if it’s regrettably low in an absolute sense. (This is similar to @Linch’s sibling comment.)
I’m also very worried about human-driven philosophy going wrong, and one motivation for trying to solve metaphilosophy is to help prevent that, either by improving human philosophical competence, or getting philosophically competent AI advisors or sovereigns. If we pause AI, there’s a chance that by the time we restart it, there will be some progress or more interest in this project.
Increasing human or AI philosophical competence could be (probably is) hard, in which case we need to pause or slow down tech progress in general to buy more time for philosophical/metaphilosophical progress. Pausing AI is an instance of this (since safely developing/using AI requires philosophical progress that we can’t make in time), and also opens the overton window / sets a precedent for a more general pause/slowdown.