The problem is that special knowledge is also prone to bias.
Some correct, but unpopular expert opinion can be also voted to be wrong, when the level of required knowledge is substantially high.
I think the point at which "popular but wrong" starts winning out is way lower than you think are implying it is when you're talking about internet cesspools that have a cross section of the general public (Twitter, Reddit, 4Chan, etc).
If you're talking about a forum or somewhere that the casually interested internet riff raff won't dominate the bar moves higher.
That said, I'd rather have internet idiots do group fact checking than have established institutions be in charge of it.
