> Unless you know how to assess the "proof" then you'd just be stuck in a state of permanent but useless skepticism.

Agreed, but I'd suggest that that state would (at least often) be better than the current state of affairs, where most people outsource the formation of opinions to various third parties. This is how we get things like the Iraq war, or anti-vaxxers.

> Also, many times a "proof" is inappropriate or just fundamentally unproduceable, but we are still required to engage with the content and that content could still be valuable and 'true'.

100% agree. And yet, look how many genuinely intelligent people on HN hold very strong conclusions on topics that are unprovable, with no sense of uncertainty.

> I've studied epistemology in some undergraduate philosophy classes and I've generally found it much less useful than say the reasonably complete theory of media provided by Manufacturing Consent (though my study of epistemology likely helped me understand that book).

Expressing ideas in narrative form is often an easier way to communicate as it's easier to conceptualize when it's put in relatable examples. One of the most useful ideas I think could improve the world is that there aren't only two two states of knowledge, True or False, but also (at least) one additional: Unknown. Of course, everyone will gladly admit this in a thread about philosophy, but move to a thread on politics, and watch that knowledge vanish.

> If I wasn't knowledgeable about AI at all, what would I do with a "proof" if it was given to me?

Two (there are others) options are: outsource your opinion to a third party, or remain undecided. Which one is preferable depends on the situation.