hmm that’s true - I think if you could get a measure of interest for the paper like views on arxiv or number of mentions online, that could be a good metric to use.
it’s also possible that the LLM will be able to determine well which results are important vs which are minor improvements / changes that are not important for me to listen to.
the economist does a „world in brief” news summary daily, and having that for the papers that are relevant to me would be great!