Why did these two questions have very different reception?

These two questions are, at best, superficially similar, for the reasons you write down. However, if one looks just a tiny bit deeper, the vote patterns are completely justified.

Since the age of the internet (and before, but it was a lot harder for them to reach an audience before that), cranks have been trying to claim proofs of the most famous number theory/number theory adjacent conjectures like RH, Goldbach, Twin-Primes, Collatz, etc. By sheer volume, virtually every tactic possible has already been tried. This was one of the tactics. I have seen it on MO at least half a dozen times in my own memory. Inevitably, if you allow the person to continue, there will be a later post in the form of a comment or an answer that says "actually, I just proved RH, here's my 5 page 'paper'", and often they feel the need to insult someone (Terry Tao unfortunately has borne a far greater share of this than is reasonable or deserved) along the way, because they are the 'true genius'.

The second question is extremely timely. I am sure that a lot of researchers are starting to experiment with the popular LLM's for research purposes, and if their experience is anything like mine (and Carlo's, apparently), they would find themselves astonished. Most likely the LLM's have produced some useable result, so how to properly give credit or discuss the output is a highly relevant topic. This is reflected in the high up-vote count.

If we look at the broader context beyond these two specific questions, MO ultimately functions as a total democracy as a core concept. People vote up, down, or not at all at their own discretion, and moderators only intervene in extreme cases of manipulation or foul play. If something just happens to be more popular than another, there's nothing anyone can or should do.