Posts

Sorted by New

Wiki Contributions

Comments

LessWrong Has Agree/Disagree Voting On All New Comment Threads

Claim 2: Agree/disagree buttons are confusing or even harmful for comments that are making multiple claims. This is significant enough that there should not be an agree/disagree button for comments where agree/disagree buttons are not suitable.

  • Agree: The negative consequences are significant enough that there should not be agree/disagree buttons for certain types of comments. For example, authors may be able to decide if they will allow agree/disagree votes on their comment.
  • Disagree: It is acceptable to have agree/disagree votes even for posts/comments where this does not make sense, e.g. because people will adjust accordingly. We can add in a feature to disable agree/disagree votes for certain comments, but it is also okay if we don't.
LessWrong Has Agree/Disagree Voting On All New Comment Threads

Claim 1C: See claim 1A.

  • Agree: I may or may not think that I/other users have this experience, but I think the effects are negative and significant enough, or have the potential to be significant enough that we should see if there are ways to address this when designing a new voting system.
  • Disagree: I may or may not think that I/other users have this experience, but I think that the effects are not negative or are negligible enough that we do not need to factor this into the design of a new voting system.
     
LessWrong Has Agree/Disagree Voting On All New Comment Threads

Claim 1B: See claim 1A.

  • Agree: This may or may not match my experience, but I believe that for majority (>50%) of users on LW, they are less likely to write replies expressing agreement/disagreement because they can now vote agree/disagree.
  • Disagree: This may or may not match my experience, but I believe that majority (>50%) of users on LW, would still write a reply even if they can just vote agree/disagree.
LessWrong Has Agree/Disagree Voting On All New Comment Threads

Claim 1A:  Agree/disagree buttons disincentivizes productive conversations because clicking the disagree button satisfies the need for expressing disagreement (or agreement) with lower cost (less effort & no reputational cost since votes are anonymous) than writing out a reply. This is a significant enough concern that we should consider its effects when deciding whether or not to go with the new voting system.

  • Agree: This matches my experience: I am less likely to write replies expressing agreement/disagreement because I am now able to vote agree/disagree.
  • Disagree: This does not match my experience: If I was already going to write a reply, I would still write one even if I can just vote agree/disagree.
LessWrong Has Agree/Disagree Voting On All New Comment Threads

This comment is an experiment. I'm trying out a variant of the proposed idea of voting by headings/block quotes: this comment contains my comment, and the replies below contain claims extracted from my comment for agree/disagree voting. 

Agree/disagree buttons incentivizes knee-jerk, low-effort reactions rather than deliberate, high-effort responses 

Something I like about LW's system of upvotes meaning "things you want to see more of" and having no agree/disagree button is that there's no simple way of expressing agreement or disagreement. This means that when there's something I disagree with, I'm more incentivized to write a comment to express it. That forces me to think more deeply because I need to be able to state clearly what it is I'm agreeing or disagreeing with, especially since it can be quite nuanced. It also feels fairer because if someone went to the effort of writing a comment, then surely it's only fair that I do likewise when disagreeing. (Unless of course it was a low effort comment, in which case I could always just downvote.)

I suspect that if there's an agree/disagree button, the emotional part of me would be satisfied with clicking the disagree button, whereas currently, it pushes me to express my disagreement as a (thought-through, reasoned) reply. I aspire to be someone who responds thoughtfully, but that is not an instinctive behavior. With the disagree button available, I would be fighting instinct rather than working with it. It encourages emotional, knee-jerk reactions rather than deliberate responses.

(It's nice to be able to get a rough gauge of the community's opinion of a statement though. It's not of much practical use in terms of evaluating the truth of a statement, because I prefer to weight different people's opinions differently based on the topic, but it does give a general sense of the community's opinion.)

Agree/disagree buttons are confusing or even harmful for comments that are making multiple claims

There are Telegram channels by news agencies where they post messages for articles. Each message contains a headline and a link to the article, and people would then react to the chat message using emojis. It's quite amusing when there are headlines like "Person X convicted of Y and sentenced to Z" and you see many thumbs up and thumbs down. It makes me wonder, are you showing approval/disapproval for the crime, the conviction, or the punishment? It also seems to contribute to typical mind fallacy/confirmation bias problems.

Similarly, having agree/disagree buttons on comments that have multiple claims doesn't really make sense, because we can't tell which part is being agreed/disagreed with and people might end up interpreting the votes according to their own beliefs. 

Suggested alternative: agree/disagree buttons for claims created specifically for voting

Others have suggested allowing voting per heading or by block quote but I think that the way I phrase my comments is different from how I would craft a claim. Also, some statements aren't meant for people to evaluate  (e.g. sharing of personal stories, giving encouragement, sharing related posts). 

Thus, one possibility I can think of is to let users create claims specifically for agree/disagree voting. Other users (besides the author) can also add in separate claims extracted from the comment. Hopefully, when claims are designed to be agreed/disagreed with, it makes the agree/disagree votes easier to interpret. (Ideally, there should probably be a third option that says "this is not a statement that can be meaningfully agreed/disagreed with" or "this is not a well-crafted statement".) 


Thoughts after trying it out

  • This is really, really hard. I'm not sure if my claims as well-formulated, and I'm not sure which claims are meaningful to extract from my comment. When there are many ways to disagree with a statement, I can't tell which ways are more meaningful (i.e. where to draw the line for agree/disagreeing or which ones are worth creating separate claims for). It's also very high effort to write compared with a typical comment.
  • I notice that a part of me seems to prefer uncontroversial claims to get that validation, while another part of me wants more controversial claims so that it'll be more fun (i.e. won't know what answer to expect).
  • Would it make sense to have a separate section where we can view all the claim comments for a post? (probably needs to support some form of sorting by relevance) Would that be a way to help the community reason something out collectively?
  • I wonder if it has a tendency to focus attention on a narrower set of ideas, simply because those are the options offered. 
Benign Boundary Violations

MSRayne is saying "no, not in my experience," but afaict MSRayne has also self-identified as being in the set of [people whose personal boundaries already lie outside of the social boundary, such that even things which do not violate the social boundary are already violating their personal boundary].

Yes I'd read about this in the other comment but I think it didn't really register until I saw MSRayne's reply above.

The reply was enough for something to click in my head, possibly because it was a more concrete explanation, but your explanation made the misunderstanding more explicit to me, so thanks!

Benign Boundary Violations

Oh I think I see what you mean. If there's always a cost to saying no, then all boundary violations are basically threats and hence aggressive.

And I think you always lose something if you say no to someone - always. It is always coercive. It just may not be visible on the surface - but they will resent you a little bit for it, and the more you do it the more resentment will build up.

I recognize this, or at least something like it - it's like when people ask for your opinions. People say that there is no wrong answer and that you should say what you really think, but I always felt that that wasn't true. There are wrong answers, and you will know that they are wrong because people will respond negatively to them (e.g. they like you less afterwards because your opinion differed from theirs). People don't really want to hear what you have to say; they just want validation.

To avoid saying the wrong thing, I ended up trying to figure out what people were hoping to hear (e.g. based on how they phrase their questions), so that I could tell them what they wanted me to say. I didn't even notice that habit until one day when someone asked me a question and I couldn't tell what they wanted - they were completely blank to me. I ended up giving an answer truer to myself, and was expecting a negative response. Yet they didn't show disapproval, and more surprisingly, neither did they show approval. They really just did want my answer!

The experience showed me that something I thought was a trait of all humans was actually more like an attribute that varies based on the individual. Some people just want validation, but others genuinely want to hear what you have to say. That changes the game, because it means it's not actually my job to say what people want to hear, it's just how some people prefer to be dealt with. I can always keep my true thoughts aside for people who want to hear them.

Some time after, I shared my opinion with someone who responded dismissively. Yet days later, they asked me a question that showed that they were thinking about what I'd said. I learned that just because someone responds negatively, it doesn't necessarily mean they are upset with me and want me to be different; sometimes it's just a natural response to hearing something you don't like or even just something new. What's interesting is that had I continued saying what I thought others would want to hear, I wouldn't have realised that people are ok with listening to what I have to say.

There are things I tend to avoid because they weren't good experiences in the past and when I think of doing them now, it just feels like a bad idea. Sometimes when I'm with the right people or in the right context though, my mind realizes that there is a very low likelihood of something terrible happening, it's just my heart that's convinced that something awful will happen. But when my heart wants something badly enough, the risk becomes worth it and so I try it even though it feels scary. So far, it's paid off every time. Sure, sometimes it doesn't go the way I hope for, but then again nothing terrible happened either.

I think the difference is that where I used to pay attention to just my negative experiences, I now also pay attention to when there isn't a negative response, both for my myself and when watching others interact. I notice that the ratio is different from what I'd always thought it was (1:0), because the people I'm with are different, because people change, and because I pay attention to a broader slice of reality. That's why to feels safer to try (with the right people). (There's also that I'm more capable now, and can therefore cope better with anything that might happen.)

I think it's quite interesting how sometimes you can't tell if your beliefs are wrong unless you are willing to do things that past experiences say you shouldn't, and create opportunities to prove your beliefs false. It's like confirmation bias, except I'd never thought to apply it to personal/emotional experiences.

I don't know, can't know what your experiences are like - I couldn't even understand Caperu_Wesperizzon's and your comments. I want to say though, that I think people who are nice and good with boundaries do exist, and I hope that you get to meet them someday.

Benign Boundary Violations

If you fail to respond adequately you decrease the respect of your comrades (because you can't take it like a man or whatever) and thus by proxy decrease intimacy.

Hmm if you lose respect for responding wrongly then it doesn't really seem like a benign boundary violation anymore? The way I see it, a boundary violation can be considered benign only if you are capable of saying no, and the other person is genuinely capable of accepting and respecting a no. Otherwise, it's more like coercion. (And the violation shouldn't have very negative consequences for the person, based on what can be anticipated. )

If your friend takes your things without asking and you tell them to stop doing it because you don't like it, and they apologise and stop doing it, then that was a benign boundary violation. If they stop but then go around telling others that you are selfish, or they stop and then complain about how they always have to give in to your demands, or they ignore you and tell you that best friends share everything, then that's not benign at all. You can't really tell from the boundary violating action though, only from their response when you say no.

People who are more powerful (e.g. physically stronger, higher social status) are more capable of saying no because the consequences of saying no are less severe for them. In that sense, things that seem like benign boundary violations are more likely to be benign for them, so they tend to see it as benign (and may not realise that this is not the case for others). I don't think it's benign just for the masculine though, because it works the other way around as well. If the person who is violating the boundaries is responsible about it (e.g. sensitive to potential power imbalances), it can also work. Also, boundary violations don't have to be aggressive (?). Here are some examples that are milder/more feminine that I think also count as benign boundary violations (if done properly):

  • affectionate nicknames (For a female version of the faggot example, I had a schoolmate who called people "bitch" only if she considered them a friend, e.g. greeting them with "Hey bitch!")
  • playing with/braiding someone's hair without asking
  • adjusting someone's collar when you see the tag sticking out
  • giving someone very sour candy without telling them that beforehand
  • untying someone's shoelaces (making sure they notice it before standing up so they don't accidentally trip)
  • asking "Can I borrow your pen pretty pretty please? Just 5 seconds! Thanks!" and taking it before you hear them say yes
  • asking sensitive questions like salary or asking a woman for her age
  • playful emotional manipulation like making puppy eyes at someone to persuade them to share their snack with you (only works if the other person is capable of saying no if they genuinely don't want to do it, and you are capable of truly accepting the rejection, and both parties understand that it's play)
AGI Safety FAQ / all-dumb-questions-allowed thread

We have dangerous knowledge like nuclear weapons or bioweapons, yet we are still surviving. It seems like people with the right knowledge and resources are disinclined to be destructive. Or maybe there are mechanisms that ensure such people don't succeed. What makes AI different? Won't the people with the knowledge and resources to build GAI also be more cautious when doing the work, because they are more aware of the dangers of powerful technology?

AGI Safety FAQ / all-dumb-questions-allowed thread

In AI software, we have to define an output type, e.g. a chatbot can generate text but not videos. Doesn't this limit the danger of AIs? For example, if we build a classifier that estimates the probability of a given X-ray being abnormal, we know it can only provide numbers for doctors to take into consideration; it still doesn't have the authority to decide the patient's treatment. This means we can continue working on such software safely?

Load More