[[https://www.lesswrong.com/s/ynMFrq9K5iNMfSZNg/p/nK5jraMp7E4xPvuNv][On Enjoying Disagreeable Company - LessWrong]]

url
https://www.lesswrong.com/s/ynMFrq9K5iNMfSZNg/p/nK5jraMp7E4xPvuNv
  1. Reduce salience of disliked traits.

Identify the traits you don't like about the person - this might be a handful of irksome habits or a list as long as your arm of deep character flaws, but make sure you know what they are.  Notice that however immense a set of characteristics you generate, it's not the entire person.  ("Everything!!!!" is not an acceptable entry in this step.)  No person can be fully described by a list of things you have noticed about them.  Note, accordingly, that you dislike these things about the person; but that this does not logically entail disliking the person.  Put the list in a "box" - separate from how you will eventually evaluate the person.

When the person exhibits a characteristic, habit, or tendency you have on your list (or, probably just to aggravate you, turns out to have a new one), be on your guard immediately for the fundamental attribution error.  It is especially insidious when you already dislike the person, and so it's important to compensate consciously and directly for its influence.  Elevate to conscious thought an "attribution story", in which you consider a circumstance - not a character trait - which would explain this most recent example of bad behavior.1  This should be the most likely story you can come up with that doesn't resort to grumbling about how dreadful the person is - that is, don't resort to "Well, maybe he was brainwashed by Martians, but sheesh, how likely is that?"  Better would be "I know she was up late last night, and she does look a bit tired," or "Maybe that three-hour phone call he ended just now was about something terribly stressful."

Reach a little farther if you don't have this kind of information - "I'd probably act that way if I were coming down with a cold; I wonder if she's sick?" is an acceptable speculation even absent the least sniffle.  If you can, it's also a good idea to ask (earnestly, curiously, respectfully, kindly!  not accusatively, rudely, intrusively, belligerently!) why the person did whatever they did.  Rest assured that if their psyche is fairly normal, an explanation exists in their minds that doesn't boil down to "I'm a lousy excuse for a person who intrinsically does evil things just because it is my nature."  (Note, however, that not everyone can produce verbal self-justifications on demand.)  Whether you believe them or not, make sure you are aware of at least one circumstance-based explanation for what they did.

Notice which situations elicit more of the disliked behaviors than others.  Everybody has situations that bring out the worst in them, and when the worst is already getting on your nerves, you should avoid as much as possible letting any extra bubble to the surface.  If you have influence of any kind over which roles this person plays in your life (or in general), confine them to those in which their worst habits are irrelevant, mitigated, or local advantages of some kind.  Do not ask for a ride to the airport from someone who terrifies you with their speeding; don't propose splitting dessert with someone whose selfishness drives you up the wall; don't assign the procrastinator an urgent task.  Do ask the speeder to make a quick run to the bank before it closes while you're (ever so inconveniently) stuck at home; do give the selfish person tasks where they work on commission; do give the procrastinator things to do that they'll interpret as ways to put off their other work.

Is action under a description subject to correspondence bias? The description must be well enough constructed to avoid this, one may conclude, but does that just reduce it to empty / unverifiable in a majority of cases? there is a subject of outcomes that seem "clearly intended" by an actor, but this subset of outcomes is very clearly unverifiable in a majority of cases. however, abducting your way towards good-enough models of these seems to be a good-faith (or maybe more accurately a fair-play) social move. And certain kinds of bad actors or maladaptive actors can be characterised as employing bad strategies (harmful strategies) to accomplish their ultimate intentions (as chained up from their actions), and the identifying of these cannot be done without this kind of social immune response.

but for every immune response is a new hack to get around it, and the explicit design of permissive norms cannot happen from heuristic conclusions. how to balance the paucity of rigourously obtained information with the unavoidability of acting at all?

Permitting everything, and building causal models, gives us the option of constructing tailored responses for specific situations. present to the procrastinator as procrastinating. rotate and twist people into the configurations that make more sense for what you're trying to achieve. and disinvest in them before doing it, so you don't go trying to live fiteen thousand people's lives.

give no shits, and play five chess games.

to consider for the elements of my desired life list - what games do I want to be playing with people?