I think I just stumbled into a legitimate use case for small local #llm #ai models.
I just threw together a quick python app which pulls the #rss feeds from #arxiv and then throws the titles of each article at a tiny model running on my local machine with instructions to rank them in order to the relevance of my research. The program then shows the feed in order of relevance.
The result is an amazing benefit of both worlds. I still get a personal "algorithm" curating my feeds, but I am the one who controls the algorithm and explicitly told it what to prioritize.
I plan to #OpenSource it soon once I get the code in a readable and portable state.
Edited 246d ago