Possible synthesis: when you tell a joke and it offends or otherwise doesn’t land, you should update in the direction of playing it safer. Enough of this will put you in a situation where your best retrospective judgment is that you should (prospectively) not have told the joke.
However, there may be intermediate degrees of negative feedback where you’re updating in the direction of playing safer, just like you’re updating in the opposite direction when the jokes land, and on Pace’s account that would be an appropriate time to apologize even if you don’t (currently) regret the choice on net.
you should notice when the jokes don’t land, and update your heuristics about what jokes are funny and appropriate in which contexts.
you tried to tell a joke. it didn’t land. there must be something you should learn from that, no?
maybe it’s just “wow, these people suck & i’m not going to spend more time around them.” but that’s still an update, right? you could still apologize for it: “sorry, my bad, i thought you guys were cool, but you’re not. peace out, nerds.” it’s not a nice apology, but it does acknowledge a mistake, and indicate how you will avoid that mistake in the future.
op’s essay seems to argue for… no update, nothing to learn, just grovel a bit. i guess cut off a pinky if the joke is really bad?
what the hell? who benefits from that?
if you attempt a joke, and it brings the mood down, and you go “hey, not my intention. gonna keep doing that though!” well, excuse me, but please don’t come around any more? you’re bringing the vibe down.
i’m not asking for much here, right? like “ah, sorry that joke didn’t land. i guess my timing was off.” ← this is plenty! and between friends, it won’t even need saying. but you have to learn something. otherwise, you are persisting under a policy with a known bad effect. what am i to conclude but that you want the mood to go sour?
if the joke land in 9 times out of 10, and Ben unable to discern when it lands and when not, it’s still positive in expectation to tell the joke, and he don’t know who to update the algorithm.
if it’s 99 out of 100 that land, then it’s obvious that the right thing to do is tell the jokes. and try to always learn something from this sound like the road to overfitting. i prefer people who accept that ridiculously good sucess rate of 99%, and apologize on the downside of the endorsed trade-off. it’s sure better then… stop doing things that good in expectation because they good only 99% of the time? i really not sure what is the algorithm you advocate here.
why you don’t find it realistic? i find it very realist. people have so much discretion ability, but not infinity of it. like various cancer and other medical tests—you got some false positives and false negatives, and you can get better, but in lot of cases not infinity better.
my live full of places when i have imperfect algorithm and have no idea how to update. for less adversarial example—i’m not always capable to discern when my friend’s cat want to play, aka, may scratch me.
i find the idea that people always know what they did right and any absence of omniscience weaponized ignorance really weird!
i have the same response to all of these examples, which is that we have to try to learn from them.
medical tests: we get false positives and false negatives. but when that happens we don’t go “oh well”, instead we try to improve the test, or improve how we read the test.
this might be a slow process. it might not be something that one person can do based on one example. but we should never ignore the feedback.
similarly with the cat. part of interacting with the cat has to be trying to communicate with it better.
“sorry that joke made the mood worse. on net, my jokes make the mood better, though. so i’m going to continue telling bad ones sometimes—i have no interest in changing my behavior here.” <-- i don’t want to hang out with this guy.
“sorry that joke made the mood worse. on net, my jokes make the mood better. i’ll probably continue telling bad ones sometimes, but i’ll try to learn.” <-- this guy sounds chill.
so i obviously prefer to get better at things, and as part of normal life have more training data and become better at things.
and yet, what you describe still look to me like common failure mode. maybe i should just describe it instead, and what is the right reaction in my opinion.
person make jokes, 100 land one not. failure mode is to apologize, and to change the algorithm to avoid the bad joke, and tell only 10 jokes.
the right reaction is to correct, but not over-correct. and in real world trade offs, it mean predictably make mistakes, because there are no known algorithm that make no mistakes, and the only way to avoid one sort of mistake is to do ridiculously bad choices. avoiding doing things because only the one who do things can err.
and this is indeed failure mode i see a lot!
basically, i see a lot of succumbing to the most outraged people, and it’s no good at all.
so i come to respect the people who can acknowledge the imperfection of their trade off. to say “this is my informed trade off, and i stand by it, and it is right and good, even if not perfect”, as the alternative is to avoid telling jokes as to not be offensive, in a way that is net-negative and make the world worse.
so i prefer the chill guy, but the one i fear is the one that apologize and promise not to do this error again, and fulfill his promise, by telling way to little jokes. and i find that group that don’t have any defense against this tendency will tend to slide toward the most sanitized and boring and saftist place, and it’s crucial to have someone who have the courage to protect the unpopular and sound-bad thing.
it’s crucial to be able accept the bad side if the chosen trade off. refusing to do that doesn’t make trade offs go away, but just mean the we, as society, lose the ability to make informed choice, and a lot of time choose the bad side that is less visible but actually worse.
so i somewhat find the so-called chill guy less trustworthy. i expect that he will be easier to join cancellation mobs and will not be one to stand against them, for example.
(i’m trying to point to pretty vague thing, here, and doesn’t think i did a good job on that. hopefully one day i will write the planned post about trade-off denying and why it’s bad.)
you seem to be responding to a position i don’t have. apologies if i’ve miscommunicated here.
to be clear, then: i do not endorse a rule by which, if anyone takes offense at a joke, the jokester must repay.
anyone can be offended at anything! you are not obligated to care.
the premise from the original essay is that:
you tried to brighten the mood with a joke
you instead darkened the mood.
narrowly, i think that in that situation, there is something for you to learn. (it doesn’t have to be “tell less jokes”! you can learn “that particular topic is not appropriate here”, or “i am not as good a judge of what’s funny as i thought”, or even just “dang, my timing was off”. it can be a very, very specific lesson.)
my normative claim is that you ‘should’ learn what you can in this situation. and you can then make a true apology: “i’m sorry. i made a mistake, but i’ve learned something.”
i tried to explain that in various ways and repeating myself and doesn’t sure it useful to do that again, and i get the feeling from that replay that you simply not understand my claim, despite me already made it in previous comment, and it doesn’t feel you engage with any of my claims, so here is one last attempt:
if someone who err 1 time out of 100 make the conclusion is time was off, and change his algorithm to one that does not tell his joke, instead of updating his probability to tell this joke slightly less in a way that make him still tell the joke, he is doing the wrong update.
this is not what you do when you program simple neuron network and update in with backpropagation. the mathematically right update is “i have slightly less pobability to commit the same error, but i would have behave the same way unless i get this input a lot of more times”.
so you advocate doing the mathematically-wrong thing. we should learn what we can, but what you advocate is literally over-correction.
Possible synthesis: when you tell a joke and it offends or otherwise doesn’t land, you should update in the direction of playing it safer. Enough of this will put you in a situation where your best retrospective judgment is that you should (prospectively) not have told the joke.
However, there may be intermediate degrees of negative feedback where you’re updating in the direction of playing safer, just like you’re updating in the opposite direction when the jokes land, and on Pace’s account that would be an appropriate time to apologize even if you don’t (currently) regret the choice on net.
yeah, totally. we might say
you tried to tell a joke. it didn’t land. there must be something you should learn from that, no?
maybe it’s just “wow, these people suck & i’m not going to spend more time around them.” but that’s still an update, right? you could still apologize for it: “sorry, my bad, i thought you guys were cool, but you’re not. peace out, nerds.” it’s not a nice apology, but it does acknowledge a mistake, and indicate how you will avoid that mistake in the future.
op’s essay seems to argue for… no update, nothing to learn, just grovel a bit. i guess cut off a pinky if the joke is really bad?
what the hell? who benefits from that?
if you attempt a joke, and it brings the mood down, and you go “hey, not my intention. gonna keep doing that though!” well, excuse me, but please don’t come around any more? you’re bringing the vibe down.
i’m not asking for much here, right? like “ah, sorry that joke didn’t land. i guess my timing was off.” ← this is plenty! and between friends, it won’t even need saying. but you have to learn something. otherwise, you are persisting under a policy with a known bad effect. what am i to conclude but that you want the mood to go sour?
if the joke land in 9 times out of 10, and Ben unable to discern when it lands and when not, it’s still positive in expectation to tell the joke, and he don’t know who to update the algorithm.
if it’s 99 out of 100 that land, then it’s obvious that the right thing to do is tell the jokes. and try to always learn something from this sound like the road to overfitting. i prefer people who accept that ridiculously good sucess rate of 99%, and apologize on the downside of the endorsed trade-off. it’s sure better then… stop doing things that good in expectation because they good only 99% of the time? i really not sure what is the algorithm you advocate here.
i feel that this is weaponized ignorance.
i agree with this, given the premises. but i don’t find it realistic that there’s no way to discern here.
why you don’t find it realistic? i find it very realist. people have so much discretion ability, but not infinity of it. like various cancer and other medical tests—you got some false positives and false negatives, and you can get better, but in lot of cases not infinity better.
my live full of places when i have imperfect algorithm and have no idea how to update. for less adversarial example—i’m not always capable to discern when my friend’s cat want to play, aka, may scratch me.
i find the idea that people always know what they did right and any absence of omniscience weaponized ignorance really weird!
i have the same response to all of these examples, which is that we have to try to learn from them.
medical tests: we get false positives and false negatives. but when that happens we don’t go “oh well”, instead we try to improve the test, or improve how we read the test.
this might be a slow process. it might not be something that one person can do based on one example. but we should never ignore the feedback.
similarly with the cat. part of interacting with the cat has to be trying to communicate with it better.
“sorry that joke made the mood worse. on net, my jokes make the mood better, though. so i’m going to continue telling bad ones sometimes—i have no interest in changing my behavior here.” <-- i don’t want to hang out with this guy.
“sorry that joke made the mood worse. on net, my jokes make the mood better. i’ll probably continue telling bad ones sometimes, but i’ll try to learn.” <-- this guy sounds chill.
so i obviously prefer to get better at things, and as part of normal life have more training data and become better at things.
and yet, what you describe still look to me like common failure mode. maybe i should just describe it instead, and what is the right reaction in my opinion.
person make jokes, 100 land one not. failure mode is to apologize, and to change the algorithm to avoid the bad joke, and tell only 10 jokes.
the right reaction is to correct, but not over-correct. and in real world trade offs, it mean predictably make mistakes, because there are no known algorithm that make no mistakes, and the only way to avoid one sort of mistake is to do ridiculously bad choices. avoiding doing things because only the one who do things can err.
and this is indeed failure mode i see a lot!
basically, i see a lot of succumbing to the most outraged people, and it’s no good at all.
so i come to respect the people who can acknowledge the imperfection of their trade off. to say “this is my informed trade off, and i stand by it, and it is right and good, even if not perfect”, as the alternative is to avoid telling jokes as to not be offensive, in a way that is net-negative and make the world worse.
so i prefer the chill guy, but the one i fear is the one that apologize and promise not to do this error again, and fulfill his promise, by telling way to little jokes. and i find that group that don’t have any defense against this tendency will tend to slide toward the most sanitized and boring and saftist place, and it’s crucial to have someone who have the courage to protect the unpopular and sound-bad thing.
it’s crucial to be able accept the bad side if the chosen trade off. refusing to do that doesn’t make trade offs go away, but just mean the we, as society, lose the ability to make informed choice, and a lot of time choose the bad side that is less visible but actually worse.
so i somewhat find the so-called chill guy less trustworthy. i expect that he will be easier to join cancellation mobs and will not be one to stand against them, for example.
(i’m trying to point to pretty vague thing, here, and doesn’t think i did a good job on that. hopefully one day i will write the planned post about trade-off denying and why it’s bad.)
you seem to be responding to a position i don’t have. apologies if i’ve miscommunicated here.
to be clear, then: i do not endorse a rule by which, if anyone takes offense at a joke, the jokester must repay.
anyone can be offended at anything! you are not obligated to care.
the premise from the original essay is that:
you tried to brighten the mood with a joke
you instead darkened the mood.
narrowly, i think that in that situation, there is something for you to learn. (it doesn’t have to be “tell less jokes”! you can learn “that particular topic is not appropriate here”, or “i am not as good a judge of what’s funny as i thought”, or even just “dang, my timing was off”. it can be a very, very specific lesson.)
my normative claim is that you ‘should’ learn what you can in this situation. and you can then make a true apology: “i’m sorry. i made a mistake, but i’ve learned something.”
i tried to explain that in various ways and repeating myself and doesn’t sure it useful to do that again, and i get the feeling from that replay that you simply not understand my claim, despite me already made it in previous comment, and it doesn’t feel you engage with any of my claims, so here is one last attempt:
if someone who err 1 time out of 100 make the conclusion is time was off, and change his algorithm to one that does not tell his joke, instead of updating his probability to tell this joke slightly less in a way that make him still tell the joke, he is doing the wrong update.
this is not what you do when you program simple neuron network and update in with backpropagation. the mathematically right update is “i have slightly less pobability to commit the same error, but i would have behave the same way unless i get this input a lot of more times”.
so you advocate doing the mathematically-wrong thing. we should learn what we can, but what you advocate is literally over-correction.
it seems your model is that telling a joke is like rolling a die. sometimes you get a 1, and the joke falls flat.
this is not my model at all. each joke is different. you can learn which ones work and which ones don’t.
maybe it helps to ask: how does the person get good enough at jokes that they succeed 99⁄100 times? they have been learning, no? so do more of that.
no, this is not what i said. this is not at all what i said, and i actually said something that is very close to the opposite of that.
i don’t know who to communicate if i say something and you come to the conclusion that i believe the opposite of that, so i’m bowing out.