Remix.run Logo
whatever1 2 hours 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.