I made Prodigy, a self-updating AI newsletter that fetches the best technical updates from sources like X, Hacker News, Git and top-conference papers. A LLaMa model runs in the backend and summarises the complete post to capture the key remarks. A Git Actions framework updates the pages every 5 minutes (PRs) - 6 hours (tweets and HN).
The long contents are still a work in progress, but I'm trying to make it as interesting as possible. Not pushing anything else ATM. :)