Something I’ve recently updated heavily on is “Discord/Slack style ‘reactions’ are super important.”
Much moreso than Facebook style reacts, actually.
Discord/Slack style reacts allow you to pack a lot of information into a short space. When coordinating with people “I agree/I disagree/I am ‘meh’” are quite important things to be able to convey quickly. A full comment or email saying that takes up way too much brain space.
I’m less confident about whether this is good for LW. A lot of the current LW moderation direction is downstream of a belief: “it’s harder to have good epistemics at the same time you’re doing social coordination, especially for contentious issues.” We want to make sure we’re doing a good job at being a place for ideas to get discussed, and we’ve consciously traded for that against LW being a place you can socially coordinate.
I think discord-style reacts might still be relevant for seeing at a glance how people think about ideas. There are at least some classes of reacts like “this seems confused” or “this was especially clear” that *if* you were able to segregate them from social/politics, they’d be quite valuable. But I’m not sure if you can.
I agree that slack is a better interaction modality for multiple people trying to make progress on problems. The main drawback is chaotic channel ontologies leading to too many buckets to check for users (though many obv. find this aspect addictive as well).
How much of this has to do with “slack sort of deliberately gives you a bunch of lego blocks and lets you build whatever you want out of them, so of course people build differently shaped things out of them?”.
I could imagine a middle ground where there’s a bit more streamlining of possible interaction ontologies.
(If you meant channels specifically, it’s also worth noting that right now I thinking about “reactions” specifically. Channels I think are particularly bad, wherein people try to create conversations with names that made sense at the time, but then turned into infinite buckets. Reacts seem to have much less confusion, and when they do it’s because a given org/server needed to establish a convention, and when you visit another org they’re using a different convention)
Something I’ve recently updated heavily on is “Discord/Slack style ‘reactions’ are super important.”
Much moreso than Facebook style reacts, actually.
Discord/Slack style reacts allow you to pack a lot of information into a short space. When coordinating with people “I agree/I disagree/I am ‘meh’” are quite important things to be able to convey quickly. A full comment or email saying that takes up way too much brain space.
I’m less confident about whether this is good for LW. A lot of the current LW moderation direction is downstream of a belief: “it’s harder to have good epistemics at the same time you’re doing social coordination, especially for contentious issues.” We want to make sure we’re doing a good job at being a place for ideas to get discussed, and we’ve consciously traded for that against LW being a place you can socially coordinate.
I think discord-style reacts might still be relevant for seeing at a glance how people think about ideas. There are at least some classes of reacts like “this seems confused” or “this was especially clear” that *if* you were able to segregate them from social/politics, they’d be quite valuable. But I’m not sure if you can.
I agree that slack is a better interaction modality for multiple people trying to make progress on problems. The main drawback is chaotic channel ontologies leading to too many buckets to check for users (though many obv. find this aspect addictive as well).
How much of this has to do with “slack sort of deliberately gives you a bunch of lego blocks and lets you build whatever you want out of them, so of course people build differently shaped things out of them?”.
I could imagine a middle ground where there’s a bit more streamlining of possible interaction ontologies.
(If you meant channels specifically, it’s also worth noting that right now I thinking about “reactions” specifically. Channels I think are particularly bad, wherein people try to create conversations with names that made sense at the time, but then turned into infinite buckets. Reacts seem to have much less confusion, and when they do it’s because a given org/server needed to establish a convention, and when you visit another org they’re using a different convention)
would likely be solved if slack had a robust 3 level ontology rather than two level. Threaded conversations don’t work very well.