I’m surprised to see finding cruxes contrasted with value of information considerations.
To me, much of the value of looking for cruxes is that it can guide the conversation to the most update-rich areas.
I try to optimize my posts and comments for value of information (e.g. bringing up new ideas).
Correct me if I’m wrong, but I would guess that part of your sense of taste about what makes something an interesting new idea is whether it’s relevant to anything else (in addition to maybe how beautiful or whatever it is on its own). And whether it would make anybody change their mind about anything seems like a pretty big part of relevance. So a significant part of what makes an idea interesting is whether it’s related to your or anybody else’s cruxes, no?
The game to me is about win-win trade of interesting ideas.
Setting aside whether debates between people who disagree are themselves win-win, cruxes are interesting (to me) not just in the context of a debate between opposing sides located in two different people, but also when I’m just thinking about my own take on an issue.
Given these considerations, it seems like the best argument for not being explicit about cruxes, is if they’re already implicit in your sense of taste about what’s interesting, which is correctly guiding you to ask the right questions and look for the right new pieces of information.
That seems plausible, but I’m skeptical that it’s not often helpful to explicitly check what would make you change your mind about something.
I think caring about agreement first vs VoI first leads to different behavior. Here’s two test cases:
1) Someone strongly disagrees with you but doesn’t say anything interesting. Do you ask for their reasons (agreement first) or ignore them and talk to someone else who’s saying interesting but less disagreeable things (VoI first)?
2) You’re one of many people disagreeing with a post. Do you spell out your reasons that are similar to everyone else’s (agreement first) or try to say something new (VoI first)?
The VoI option works better for me. Given the choice whether to bring up something abstractly interesting or something I feel strongly about, I’ll choose the interesting idea every time. It’s more fun and more fruitful.
I’m surprised to see finding cruxes contrasted with value of information considerations.
To me, much of the value of looking for cruxes is that it can guide the conversation to the most update-rich areas.
Correct me if I’m wrong, but I would guess that part of your sense of taste about what makes something an interesting new idea is whether it’s relevant to anything else (in addition to maybe how beautiful or whatever it is on its own). And whether it would make anybody change their mind about anything seems like a pretty big part of relevance. So a significant part of what makes an idea interesting is whether it’s related to your or anybody else’s cruxes, no?
Setting aside whether debates between people who disagree are themselves win-win, cruxes are interesting (to me) not just in the context of a debate between opposing sides located in two different people, but also when I’m just thinking about my own take on an issue.
Given these considerations, it seems like the best argument for not being explicit about cruxes, is if they’re already implicit in your sense of taste about what’s interesting, which is correctly guiding you to ask the right questions and look for the right new pieces of information.
That seems plausible, but I’m skeptical that it’s not often helpful to explicitly check what would make you change your mind about something.
I think caring about agreement first vs VoI first leads to different behavior. Here’s two test cases:
1) Someone strongly disagrees with you but doesn’t say anything interesting. Do you ask for their reasons (agreement first) or ignore them and talk to someone else who’s saying interesting but less disagreeable things (VoI first)?
2) You’re one of many people disagreeing with a post. Do you spell out your reasons that are similar to everyone else’s (agreement first) or try to say something new (VoI first)?
The VoI option works better for me. Given the choice whether to bring up something abstractly interesting or something I feel strongly about, I’ll choose the interesting idea every time. It’s more fun and more fruitful.
Gotcha, this makes sense to me. I would want to follow the VoI strategy in each of your two test cases.