No, at the level I think an AI needs to emulate humans I don’t think it’d have any great advantages for self-improvement. I think it might be possible to highly automate innovation and scientific discovery without intelligence (and all the risks that come with it) though—i.e. a system that applies the same “dumb” algorithms over and over again and spits out answers without the possibility of it ever having a goal or desire or making a decision (confusion would be the input, clarity the output). So I still think we might be able to solve all the world’s problems in my life time.
No, at the level I think an AI needs to emulate humans I don’t think it’d have any great advantages for self-improvement. I think it might be possible to highly automate innovation and scientific discovery without intelligence (and all the risks that come with it) though—i.e. a system that applies the same “dumb” algorithms over and over again and spits out answers without the possibility of it ever having a goal or desire or making a decision (confusion would be the input, clarity the output). So I still think we might be able to solve all the world’s problems in my life time.