Applying obscure techniques from one field of study to another does sound like kind of thing LLMs ought to be good at. I'm glad that at least some of these proofs are human-parseable, even if they may be replaced with more elegant and insightful proofs down the line