LW is largely rationalist-porn for daydreamers, and that’s inevitable. However, even if we can avoid that a little bit, if we can be 95% porn and 5% actually effective, then that’s still a win worth having. Better than that is worth thinking about but hard to imagine in practice.
I think we will get more useful results if we have a thread in which we can pre-commit to writing such a thing up by the deadline, for the same reason that there should be a registrar of medical trials: a private commitment won’t be as effective in avoiding publication bias.
I am very wary of pre-commitment of this sort, it sounds like a huge temptation to play Bruce. (Woe is me, akrasia is my downfall, I told you all I’d invent Quantum gravity in a month and I never even cracked open a physics textbook...)
Also, it’s a crutch even if it works. Burning your boats should be reserved for big problems. You shouldn’t get used to coercing yourself into every fiddling little improvement!
Surely not pre-committing will offset that fear to exactly the extent that it gives people permission not to post if they failed, which is something else you’re keen to avoid?
I’m not thinking of it as a way to fight backsliding, but as a way to gather better data.
No, not really, at least that wasn’t my intention—the deadline should be absolute.
Hmm, how about: if your mind is unclouded with any attempt at self-coercion, you ought to announce in public for the sake of a more reliable strategy->win analysis.
That makes “actionable” much broader than it might sound at first blush. Could you give an example of an article which is not actionable, by this definition?
LW is largely rationalist-porn for daydreamers, and that’s inevitable. However, even if we can avoid that a little bit, if we can be 95% porn and 5% actually effective, then that’s still a win worth having. Better than that is worth thinking about but hard to imagine in practice.
I think we will get more useful results if we have a thread in which we can pre-commit to writing such a thing up by the deadline, for the same reason that there should be a registrar of medical trials: a private commitment won’t be as effective in avoiding publication bias.
I am very wary of pre-commitment of this sort, it sounds like a huge temptation to play Bruce. (Woe is me, akrasia is my downfall, I told you all I’d invent Quantum gravity in a month and I never even cracked open a physics textbook...)
Also, it’s a crutch even if it works. Burning your boats should be reserved for big problems. You shouldn’t get used to coercing yourself into every fiddling little improvement!
Surely not pre-committing will offset that fear to exactly the extent that it gives people permission not to post if they failed, which is something else you’re keen to avoid?
I’m not thinking of it as a way to fight backsliding, but as a way to gather better data.
No, not really, at least that wasn’t my intention—the deadline should be absolute.
Hmm, how about: if your mind is unclouded with any attempt at self-coercion, you ought to announce in public for the sake of a more reliable strategy->win analysis.
Require each article to be actionable and there will be no porn.
Authors should tag actionable posts as “actionable.” Then you could do a search on the tag.
The recommended action should be explicitly described.
I think this might rule out some valuable articles which are neither. Is, say, Argument screens off authority actionable? In what way? Is it porn?
What does actionable even mean in context?
An article that essentially says “please do try this at home”.
If being less irrationally credentialist counts as an action, yes.
That makes “actionable” much broader than it might sound at first blush. Could you give an example of an article which is not actionable, by this definition?