Hacker News Reader

> For a nonsensical but amusing demo that lets you watch what happens if you apply JPEG-style lossy compression to text, check out this post.

The link seems broken, since it points back to the same article at https://lcamtuf.substack.com/p/not-so-fast-mr-fourier

So I can feed a lossy text to an LLM and it can “delossy” it?

Does the lossy text take up less space so this could work as a weird compression?

Use an embedding, then quantize (i.e. rounding) maybe?

I wonder how that would look like if it used word2vec.