Casey NewtonEmbarrassed by how much I miss Twitter's quote-tweet feature on here. It was mostly used for evil but I tried to use it for good!!
Craig Newmark@caseynewton Yeah, I think I'd like quote-tweet, always used it to promote others' work, or, comedy
Rob MillerI haven't yet seen any actual data verifying that QTs caused more harm than good, or that removing them causes significant reduction in harassment and abuse. I'm 100% behind removing them even though I miss them personally if that's actually true, but so far I see a lot of folks asserting that to be so w/o any evidence.
If anyone who happens to see this has info one way or the other I'd love to hear about it, please and thank you!
llimllib@nonsequitarian @craignewmark @caseynewton what form of data could actually answer your question? It's an aesthetic decision, isn't it?
@llimllib @craignewmark @caseynewton
Looking through replies to Casey's post I see phrases such as "Given the quote tweet was disproportionately used for harm..." and "I know the main use case was dunks..." The folks saying these things believe them, but are they true?
It's an honest question. I can def imagine that the negative effects of QTs (esp w engagement driven algorithmss) far outweigh the positive, I have no reason to doubt it. But I haven't yet seen any evidence to support it.
@nonsequitarian @craignewmark @caseynewton I believe you that it's an honest question! My point is just that it's not reasonable to ask for data to support an aesthetic decision.
Any metric you would use for "causes harm" is arbitrary and will swing your answer - you could convince the data to give you any answer you wanted.
My opinion is that you can only make the judgement based on aesthetics, appealing to data won't help you
@llimllib @craignewmark @caseynewton
Have to disagree with you there. All features can be potentially used to abuse or harass, and all abuse mitigation will be imperfect. If someone is seriously interested in dealing with these issues at scale, it's absolutely critical to have metrics to track whether the measures are working. Doing this well is an incredibly difficult problem, for reasons you mention, and (again) it will always be imperfect. But not doing it at all is an abdication of duty.