Skip to content

Comment on Compressed filesystems à la language models

Comments

Interesting experiment but the author lists some caveats (Not exhaustive by any means):

"Of course, in the short term, there’s a whole host of caveats: you need an LLM, likely a GPU, all your data is in the context window (which we know scales poorly), and this only works on text data."

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.