Online sensemaking often seems to require choosing between two kinds of information sources: smarter, more knowledgeable people with more incentives to lie, and dumber, more ignorant people with less incentive to lie.
The other 2 quadrants are usually either empty or easy to dismiss (dumb-ignorant with more incentives to lie = low-level easily spotted grift, smart-knowledgeable with less incentives to lie = likely to just not care enough to conscientiously supply information)
Caricature: Whom to believe: completely corrupt institutional authority vs completely moronic auto-didact conspiracy theorist?
Which signal is more useful assuming some post-processing?
Wrong-on-purpose or right-by-accident?
For wrong-on-purpose you have to model incentives, and deconvolute based gaps between what people say vs do. If you suspect a doctor is lying, observe treatments they choose for themselves or family. Works for medium-complex stuff.
When something is so complex, all apparent expertise can be considered superstition, the moron is probably the better bet. Might be right entirely by accident by entertaining more possibilities.
Gettier case of sorts. The stopped clock has a slight chance of telling the right time. A known-unknown slow clock has none.
Justified true lies vs unjustified true beliefs. Trap justified true beliefs somewhere in between.
“Do your own research” is the dumbest idea though because in most cases you can’t. At best you can learn how to test research for validity and decide whether or not to trust. N=1 broscience is not a real option for most situations.