3 ms·
I think compression algorithms are ripe for significant improvement with LLMs. It’s an ideal candidate problem you can have in a closed loop evaluation, and you
by whatever1 27d ago
I think compression algorithms are ripe for significant improvement with LLMs. It’s an ideal candidate problem you can have in a closed loop evaluation, and you can just let agent try things.
- Leftium 26d ago> Compression is prediction (According to recent article on HN[1]. So LLMs _are_ compression!) [1]: https://hn.leftium.com/i/49263497 https://hn.leftium.com/i/49263497