Even in people whose conscious world models are basically sane (by our exacting LW standards), when they're considering doing or planning for some weird, uncomfortable, and out-of-the-ordinary action seemingly justified by weighing costs and benefits, it seems to me that akrasia can sometimes be a rational stand-in for some considerations they don't keep conscious track of, including but not limited to: reputation costs, willpower loss / ego depletion, other limits to worry, the inference that if one weird thing seems especially important now others are going to seem especially important in the future, the possibility that one might waste resources by not following through on a weird plan, the desirability of keeping one's mind "cleaner" by keeping fewer chunks in one's planning space, benefits of long-term happiness and of not associating unhappiness with rationality to oneself or to other people, benefits of having one's actions and motivations make sense to other people, various self-image issues. There are going to be many unmodeled considerations in the other direction too, but I suspect they will normally be fewer.
It's better to forego a 1-expected-util hare-brained scheme than let it distract you into a 20% chance of foregoing a 10-expected-util hare-brained scheme, or so Morgensternsai tells me.
On the other hand, I don't want to hand people tools for mediocrity here; many times akrasia in these situations really is just irrational. I wish I had a better idea of when.
People have been encouraging me to share my anti-akrasia tricks, but it feels inappropriate to dedicate a top-level post solely to unproven techniques that work for some person and may not work for others, so:
Go ahead and share your anti-akrasia tricks!
Let's make it an open thread where we just share what works and what doesn't, without worrying (yet) about having to explain tricks with deep theories, or designing proper experiments to verify them. However, if you happen to have a theory or a proposed experiment in mind, please share.
Bragging is fine, but please share the failures of your techniques as well – they are just as valuable, if not more.
Note to readers – before you read the comments and try the tricks, keep in mind that the techniques below are not yet
provensupported or explained by proper experiments, and are not yet backed by theory. They may work for their authors, but are not guaranteed to work for you, so try them at your own risk. It would be even better to read the following posts before rushing to try the tricks: