I appreciate you writing the list of examples, but also don’t find them currently particularly compelling, mostly because they are hard to verify from my current perspective (which is fine and the real value might just be hard to communicate).
If I say that a benefit is a feeling of enormous peace and relief, you’ll have a hard time verifying that too. Seems to me like you’re Goodharting on what you’re allowing to be compelling.
I’m really quite disheartened that you find the value of e.g. mending a long-hurt relationship to be “hard to communicate”. That’s… I hurt for you.
I have very high priors on people making up post-hoc stories about the things that happened to them, and the reasons for why they happened. While I don’t disbelieve that a long-hurt relationship of yours has gotten better, I am very skeptical of any specific story you tell of how it has come to that, and so don’t really take that account to be super strong evidence.
You can frame it as a “lemons problem”, or you can frame it as an inevitable result of information assymetries, but in either case this makes benefits like the ones you describe hard to communicate.
This would make sense to me applied to a generic person about whom you know very little; in context it feels to me like you’re outside viewing too hard here. It may be socially much less comfortable to give an inside view description of why you don’t find Val’s comments persuasive, so you don’t need to interpret this as a request for such a description, but this strikes me as reasoning that’s been optimized for defensibility instead of truth-seeking and that seems worth pointing out.
What I know about Val does not make me update downwards particularly strongly about the tendency to make up post-hoc stories. Very few people strike me as being particularly more trustworthy in that direction, which in itself might be a failure of outside-viewing too hard, but that’s definitely my considered epistemic state, taking into account what I know about Val.
(May it be said that my general distrust of people having developed real rationality skills has been a source of disagreement with many, and some source of personal tension between me and others, including me and Val in the past.)
[Meta note to onlookers who seem to have maybe downvoted Qiaochu’s comment]
I think a comment like the one Qiaochu wrote would indeed be out of place and weird for people who didn’t have a personal relationship with me, but given the kind of relationship me and Qiaochu do have, I found the comment to be well-placed and helpful, and expect that the general algorithm of highlighting these potential faults in my thinking patterns will be very useful for me in the long-run, if executed by Qiaochu.
Ah. I didn’t realize you were referring to information asymmetry. That makes sense to me.
I still think you’re doing the updating math wrong. Qiaochu’s thing about you outside-view-ing too hard rings true from where I stand. But I think this is an epistemic disagreement you and I have had for way longer than four months. So… I guess my examples don’t get to be compelling to you. Oh well.
I appreciate you writing the list of examples, but also don’t find them currently particularly compelling, mostly because they are hard to verify from my current perspective (which is fine and the real value might just be hard to communicate).
If I say that a benefit is a feeling of enormous peace and relief, you’ll have a hard time verifying that too. Seems to me like you’re Goodharting on what you’re allowing to be compelling.
I’m really quite disheartened that you find the value of e.g. mending a long-hurt relationship to be “hard to communicate”. That’s… I hurt for you.
I have very high priors on people making up post-hoc stories about the things that happened to them, and the reasons for why they happened. While I don’t disbelieve that a long-hurt relationship of yours has gotten better, I am very skeptical of any specific story you tell of how it has come to that, and so don’t really take that account to be super strong evidence.
You can frame it as a “lemons problem”, or you can frame it as an inevitable result of information assymetries, but in either case this makes benefits like the ones you describe hard to communicate.
This would make sense to me applied to a generic person about whom you know very little; in context it feels to me like you’re outside viewing too hard here. It may be socially much less comfortable to give an inside view description of why you don’t find Val’s comments persuasive, so you don’t need to interpret this as a request for such a description, but this strikes me as reasoning that’s been optimized for defensibility instead of truth-seeking and that seems worth pointing out.
What I know about Val does not make me update downwards particularly strongly about the tendency to make up post-hoc stories. Very few people strike me as being particularly more trustworthy in that direction, which in itself might be a failure of outside-viewing too hard, but that’s definitely my considered epistemic state, taking into account what I know about Val.
(May it be said that my general distrust of people having developed real rationality skills has been a source of disagreement with many, and some source of personal tension between me and others, including me and Val in the past.)
[Meta note to onlookers who seem to have maybe downvoted Qiaochu’s comment]
I think a comment like the one Qiaochu wrote would indeed be out of place and weird for people who didn’t have a personal relationship with me, but given the kind of relationship me and Qiaochu do have, I found the comment to be well-placed and helpful, and expect that the general algorithm of highlighting these potential faults in my thinking patterns will be very useful for me in the long-run, if executed by Qiaochu.
Thanks Oli, I appreciate your candor and your support.
Ah. I didn’t realize you were referring to information asymmetry. That makes sense to me.
I still think you’re doing the updating math wrong. Qiaochu’s thing about you outside-view-ing too hard rings true from where I stand. But I think this is an epistemic disagreement you and I have had for way longer than four months. So… I guess my examples don’t get to be compelling to you. Oh well.