A wonderful post. The binding constraint has always been questions. When I was a graduate student, the very eminent (and curmudgeonly) Bryce DeWitt used to always say that the most important thing in research is having taste: asking the right question. I would not, however, be so sure that with continued development, AI would not develop taste. It does not quite have it now--but it could. It "runs" on the same laws of physics as we do, so there is no law stopping it from developing taste.
Thanks! I have the podcast you shared the other day and your book lined up for this weekend. I guarantee finishing the podcast - the book looks like it'll take a bit more than a couple of days to process :)
A wonderful post. The binding constraint has always been questions. When I was a graduate student, the very eminent (and curmudgeonly) Bryce DeWitt used to always say that the most important thing in research is having taste: asking the right question. I would not, however, be so sure that with continued development, AI would not develop taste. It does not quite have it now--but it could. It "runs" on the same laws of physics as we do, so there is no law stopping it from developing taste.
Thanks! I have the podcast you shared the other day and your book lined up for this weekend. I guarantee finishing the podcast - the book looks like it'll take a bit more than a couple of days to process :)
Thanks! I hope you enjoy them. Also, check this out, for instance (and this is already far behind the state of the art): https://www.anthropic.com/research/vibe-physics