That’s well and good, but how are we supposed to evaluate the accuracy of random HN comments without anything resembling somewhat objective metrics? People say

That’s well and good, but how are we supposed to evaluate the accuracy of random HN comments without anything resembling somewhat objective metrics? People say all manner of things, and usually it’s contradictory. What heuristic do you propose?