This is sort of a non-answer, but I think some people who use these methods have a different epistemology (subjectivist, non-positivist), so wouldn’t necessarily have generalization as their goal.

But are your epigraphs epigrammatic?

He says, bolting the stable door as the horse disappears over the horizon

Is it large enough that people are writing books with audio rather than reading in mind? I would think the biggest change would be in the complexity of writing.

As @tootlefish.bsky.social alludes to in another reply, AI chatbots probably make academic referencing _more_ important to avoid hallucinations.

Footnotes are optimal for any academic, but I will grudgingly settle for online endnotes if the alternative is nothing at all.

People speaking to a popular audience are making factual claims, too, so I don’t see why those should be more difficult to verify.

This, incidentally, is one of the big problems with Acemoglu & Johnson’s Power and Progress. LOADS of claims, no citations!

But they are essential for scientific reproducibility and verification.

For balance, some techno-optimism: The Adobe Scan app is at least a 10% productivity boost in photographing archive documents, mostly through quality improvement and ease of labeling/filing.

Yes—it had no idea of what I was hoping to communicate. At least I understood what that was, so I knew not to use it!