I've been tinkering with a new algorithm that compresses data by exploiting self‑similar patterns—might be handy for your next prototype. what do you think?