People will, sometimes, very confidently state things they have no way of knowing. I'm not saying they make mistakes – we all make mistakes – but they'll say things essentially randomly. When they do this with confidence, it's time to just ignore everything they say for a while. Let them develop more trustworthiness, and in the meantime, chuck everything they say into the garbage.

The classic example of this is the 14-point spread in the 2010 survey about gays serving openly in the military vs homosexuals serving openly in the military. At least 14% of Americans were confused by the question, and should be ignored. But a similar effect is seen in a survey of expert opinions about the future of Artificial Intelligence.

In fact, only having 14% of people being confused (while an underestimate from the survey) is amazing compared to this survey of experts in their field. On average, they think (with huge error bars, but, ya know, of course) it will take 45 years (some of which has since elapsed) for machines to be able to do every task better and more cheaply than human workers, but 122 years for machines to do every task in every occupation... this is stone cold nonsense, obviously. And some of this is altered by the error bars, which are huge, and compounded by about 1/4 of the experts who don't think either event is likely any time soon at all, so medians and 50% cumulative probabilities skew upwards easier than downwards. But none of that explains the error – merely why, if people were answering randomly, the error would be that big.

But answering randomly is not appropriate! Not even for estimations and guesses! Imagine an actual state of the world, have an actual opinion, or don't offer it, and certainly don't offer it under the guise of a recognized and published expert in the topic.

It would be far more appropriate to say that you don't know something, if you truly don't. Or to take the time to develop a guess about the future. Or say, it'd take me X hours to complete a first estimate. And then not do it, if you don't have the time or budget.

Perhaps I am crazy, or so used to the constraints of engineering, that hearing people mouth off and say impossible things makes me think "this group of people is so totally unreliable they cannot get their minds around a coherent guess". Literally less sensible than the failed apocalypse cults – at least they had a guess! As improbable as those predictions were, anything is more likely than a logical impossibility.

This is all by means of contextualizing: there's a strange culture war battle, a tiny proxy battle in some weird corners of the thinking world, about whether this topic deserves more attention, or more generally, whether it deserves more respect. Let me emphatically say: every single human is measured, in some sense, by their impact on the world. To choose a path with a low probability of making a big difference means you adopt that relationship with strangers as well. With high probability, you will be disrespected and ignored, because what you are doing is a waste of time. And if you do your work poorly, there isn't a way of getting credit for trying to do something important and failing.

I'm not sure why a baker deserves less dignity than a fancy researcher at Stanford. So everyone follows the same rules. We're all in this together.