If the point of 3 is to find the flaws in a seemingly attractive theory, that’s very commendable, but it’s not necessary to do the work yourself … you can go to a forum populated by experts , and ask for comments.
And what forum is that? The point of Richard Ngo’s comment is that LW doesn’t do steps 3 and 5 either,[1] not frequently and reliably enough to be sufficient. And sadly, I’m not aware of any other forums that do this either. Particularly because the kinds of questions LW is interested in (rationality, AI, general sociological speculation) are not ones where established experts can directly point to existing literature that answers those questions clearly and unequivocally. Take something like Raemon’s attempts to do feedbackloop-first rationality as a representative example.
Doing steps 3 and 5 reliably seems to basically require a commitment similar to that of a full-time job.[2] And it requires a tremendous amount of already-existing expertise, and a fair bit of research taste, and a commitment to norms and principles of epistemic rationality, etc. All without pay and without the allure of increased reputation and publications on your CV that working in regular academia gives you.
Particularly because the kinds of questions LW is interested in (rationality, AI, general sociological speculation) are not ones where established experts can directly point to existing literature that answers those questions clearly and unequivocally
I was taking that for granted that you are unlikely to get definitive answers, and the exercise was more about avoiding known errors—“reinventing the wheel and making it square”.
Doing steps 3 and 5 reliably seems to basically require a commitment similar to that of a full-time job.[2]
What is the point of steps 3 and 5?
And it requires a tremendous amount of already-existing expertise, and a fair bit of research taste, and a commitment to norms and principles of epistemic rationality, etc.
Even if you are soliciting expert advice, not delving into primary sources?
Not even in comments
It’s possibly for experts to turn up and offer unsolicited critique as well...but the recipient needs to listen to benefit.
And what forum is that? The point of Richard Ngo’s comment is that LW doesn’t do steps 3 and 5 either,[1] not frequently and reliably enough to be sufficient. And sadly, I’m not aware of any other forums that do this either. Particularly because the kinds of questions LW is interested in (rationality, AI, general sociological speculation) are not ones where established experts can directly point to existing literature that answers those questions clearly and unequivocally. Take something like Raemon’s attempts to do feedbackloop-first rationality as a representative example.
Doing steps 3 and 5 reliably seems to basically require a commitment similar to that of a full-time job.[2] And it requires a tremendous amount of already-existing expertise, and a fair bit of research taste, and a commitment to norms and principles of epistemic rationality, etc. All without pay and without the allure of increased reputation and publications on your CV that working in regular academia gives you.
Not even in comments
As a necessary, but surely not sufficient, condition
If it’s philosophy, you can go to philosophyforums.com or philosophy.stackexchange.com or r/askphilosophy.
If it’s physics, you can go to physicsforums.com...etc.
I was taking that for granted that you are unlikely to get definitive answers, and the exercise was more about avoiding known errors—“reinventing the wheel and making it square”.
What is the point of steps 3 and 5?
Even if you are soliciting expert advice, not delving into primary sources?
It’s possibly for experts to turn up and offer unsolicited critique as well...but the recipient needs to listen to benefit.