"We are going to use powerful AI to teach kids to do jobs that AI will almost certainly do better in 10-20 years?"
I think understanding how to work well with AI and what its limitations are will be helpful regardless of what the outcome is.
Even if silicon brains achieve AGI or super-intelligence, I think it's highly unlikely that they will supersede biological brains along every dimension. Biological brains use physical processes that we have very little understanding of, and so they will likely not be possible to fully mimic in the foreseeable future even with AGI. We don't know exactly how we'll fit in and be able to continue being useful in the hypothetical AGI/super-intelligence scenario, but I think it's almost certain there will be gaps of various kinds that will require human brains to be in the loop to get the best results.
And even if we do assume that humans get superseded in every conceivable way, AGI does not imply infinite capacity, and work is not zero sum. Even if AI completely takes over for all the most important problems (for some definition of important), there will always be problems left over.
Right now, just because you aren't the best gardener in the world (or even if you're one of the worst), that doesn't mean you couldn't make the area around where you live greener and more beautiful if you spent a few months on it. There is always some contribution you can make to making life better.
I think understanding how to work well with AI and what its limitations are will be helpful regardless of what the outcome is.
Even if silicon brains achieve AGI or super-intelligence, I think it's highly unlikely that they will supersede biological brains along every dimension. Biological brains use physical processes that we have very little understanding of, and so they will likely not be possible to fully mimic in the foreseeable future even with AGI. We don't know exactly how we'll fit in and be able to continue being useful in the hypothetical AGI/super-intelligence scenario, but I think it's almost certain there will be gaps of various kinds that will require human brains to be in the loop to get the best results.
And even if we do assume that humans get superseded in every conceivable way, AGI does not imply infinite capacity, and work is not zero sum. Even if AI completely takes over for all the most important problems (for some definition of important), there will always be problems left over.
Right now, just because you aren't the best gardener in the world (or even if you're one of the worst), that doesn't mean you couldn't make the area around where you live greener and more beautiful if you spent a few months on it. There is always some contribution you can make to making life better.