Eh, it's not so clear cut. LLMs have definitely killed a lot of fun in reverse engineering old games by making it trivial to solve riddles that in the past would have been interesting problems to work on.
But the field still offers many mysteries that need human brains to be solved.
Whether LLMs are a net positive here depends if one is result-oriented (e.g. wants to rewrite an old game in a modern language) or process-oriented (is there for the investigation work and the intellectual challenges).
