5 ms·
One thing worthy of study about Stack Overflow/Wikipedia/HN etc is the Reward System - or specifically their experiments at answering the question within a comm
by op03 6y ago
One thing worthy of study about Stack Overflow/Wikipedia/HN etc is the Reward System - or specifically their experiments at answering the question within a community or a society of who/what gets propped up and who/what gets kicked out. They haven't produced polarization, radicalization, scams, privacy issues like Facebook, Youtube or Twitter.
In the news/social media the reliance on Like/Follower/Click/View/Upvote counts as the basis for a reward/ranking system was not just that it raked in the cash or successfully captured attention, but that it was the easiest thing to implement. Back then it was just hard to do language understanding. NLP tech was pretty hopeless and so we got page rank and the Like/upvote count as half baked substitutes to work out who/what is interesting. With human moderation filling the gap.
But its obvious the human moderation on Wikipedia/Stackoverflow/HN etc have produced very different outcomes from what Twitter/Facebook/Youtube have.
So I just wish more of the debates on privacy, echo chambers, disinfo, polarization etc focused on Reward System. But most of the time debates on central/distributed, crypto, scaling, privacy, radicalization etc etc dont touch much on the Reward system and its role.
- throwawaybutwhy 6y ago> They haven't produced polarization, radicalization, scams, privacy issues like Facebook, Youtube or Twitter. You must be reading a different Stack Overflow, then. SO has had * a grand leak (or series of leaks), * revelations about spying on users with ads, * polarization between the company and power users/moderators, * a radical scandal over pronouns and thought policing, with a SO top manager spewing libel against an ex-mod in an interview leading to an out-of-court settlement, * and overall decline of quality across the board.
- knowhy 6y agoThe polarization between the company and power users/moderators and the scandal probably killed SO. The number of answered questions is now almost as low as it was in late 2011 [0]. I don't know how to plot the quality of questions and answers but I can confirm the observation of a general decline in quality. The problem with reward based sites is that people (or at least I am) are less stimulated to put in effort into a sinking ship. 0: https://data.stackexchange.com/serverfault/query/303570/answer-rate-over-time#graph https://data.stackexchange.com/serverfault/query/303570/answ...
- op03 6y agoDrama exists everywhere especially when systems aren't mature (ie still producing random effects). Even on Wikipedia. But my point was we have to go beyond the Counts and experiment on Reward/Punishment Systems more. Focus goes to the drama and after that getting all sides to focus back on a reward/punishment system improvement is really challenging these days.