ParentFull threadLeftium·> Compression is prediction(According to recent article on HN[1]. So LLMs _are_ compression!)[1]: https://hn.leftium.com/i/49263497View on HN