I think this post makes an excellent core point re the evolutionary origin of human values and I think it’s other points, while obvious in some sense, are valuable to articulate because they have gone underappreciated in recent discussions. I hope to see a greater continued focus on the evolutionary origin of values in future alignment discussions.
Thanks! I was raised on The Selfish Gene and have a Biology minor, and yes, all of this seems fairly obvious to me — but I agree, it does seem to often get overlooked in discussions of alignment. Human values, including our moral intuitions, are complex and fragile, but they have a well-understood source and one description of them fits in a ~4GB genome — they’re not some mysterious philosophical problem, they’re Evolutionary Psychology: you can make theoretical predictions about them, you can study them with standard psychological methods, you could investigate their genetic basis, and with a sufficiently large and accurate simulation you could even do experiments on their evolution (modulo possible moral issues in the experimental design).
I think this post makes an excellent core point re the evolutionary origin of human values and I think it’s other points, while obvious in some sense, are valuable to articulate because they have gone underappreciated in recent discussions. I hope to see a greater continued focus on the evolutionary origin of values in future alignment discussions.
Thanks! I was raised on The Selfish Gene and have a Biology minor, and yes, all of this seems fairly obvious to me — but I agree, it does seem to often get overlooked in discussions of alignment. Human values, including our moral intuitions, are complex and fragile, but they have a well-understood source and one description of them fits in a ~4GB genome — they’re not some mysterious philosophical problem, they’re Evolutionary Psychology: you can make theoretical predictions about them, you can study them with standard psychological methods, you could investigate their genetic basis, and with a sufficiently large and accurate simulation you could even do experiments on their evolution (modulo possible moral issues in the experimental design).