I think the upvote/downvote model for evaluating content is fundamentally not a good one, and this deficiency is the root problem for a lot of sites (Reddit, Quora, etc.) that were founded around the same time, circa 2005-2012. It lends itself too much to groupthink, popularity contests, and self-promotion.
X/Twitter has a million other problems, but I do like their new Community Notes of fact-checking claims. It's essentially anonymous and somewhat immune to self-promotion.
Going forward, a better path for FAQ-type sites is probably something similar to that model. The hard part is how you get users to answer questions while removing the self-promotion and gamification/scoreboard incentives.
It's possible that AI might just destroy this space entirely, but I still think you need some form of human fact-checking in order to avoid hallucinations.
X/Twitter has a million other problems, but I do like their new Community Notes of fact-checking claims. It's essentially anonymous and somewhat immune to self-promotion.
Going forward, a better path for FAQ-type sites is probably something similar to that model. The hard part is how you get users to answer questions while removing the self-promotion and gamification/scoreboard incentives.
It's possible that AI might just destroy this space entirely, but I still think you need some form of human fact-checking in order to avoid hallucinations.