On the one hand, I like the gesture of commenting nice but relatively empty stuff like “this was great”. On the other hand I dislike spam, and this feels kind of redundant with the karma system. Not sure what I think about this.
I’m curious what you think are the suboptimal ways (and the optimal one) to implement this and what detrimental behavior they can incentivize?
As one random example: I could pay 10$ to any comment which agrees with me, and to any comment which criticizes one of my own critics. I don’t even need to say this explicitly, and yet over time it would still absolutely warp discussions.
And what if a critic notices that and offers 20$ each? Then we’re suddenly in an arms race.
I’m not sure how to prevent failure modes like that.
Oh, well the way I imagined it the reward isn’t visible or influences how comments are shown. So I guess it could hypothetically create an incentive to make comments that agree with you, but if it’s not visible and has no direct influence that seems very unlikely. If it is then it’s more likely, but It still doesn’t seem like it would be a big problem. Maybe moderators can have automatic monitoring for that kind of thing like they have for mass up/down voting?
On the one hand, I like the gesture of commenting nice but relatively empty stuff like “this was great”. On the other hand I dislike spam, and this feels kind of redundant with the karma system. Not sure what I think about this.
As one random example: I could pay 10$ to any comment which agrees with me, and to any comment which criticizes one of my own critics. I don’t even need to say this explicitly, and yet over time it would still absolutely warp discussions.
And what if a critic notices that and offers 20$ each? Then we’re suddenly in an arms race.
I’m not sure how to prevent failure modes like that.
Oh, well the way I imagined it the reward isn’t visible or influences how comments are shown. So I guess it could hypothetically create an incentive to make comments that agree with you, but if it’s not visible and has no direct influence that seems very unlikely. If it is then it’s more likely, but It still doesn’t seem like it would be a big problem. Maybe moderators can have automatic monitoring for that kind of thing like they have for mass up/down voting?