Meta Secretly Trained Its AI on a Notorious Piracy Database, Newly Unredacted Court Docs Reveal

The llama-1 paper acknowledged the use of the books dataset, libgen isn't mentioned in any of the papers so this is new info.

It's called Llama and it's one of the more prominent open source LLMs. So it's got that going for it at least.

Meta ai is pretry much the basis of all open weight models. I fact i use a meta model as my main model, a dolphin fine tuning but still meta based.

IIRC Meta was one of the first companies to publicly release a language model that you could run on your computer, called llama. Hence the naming of projects like llama.cpp and ollama.

I use it to release stress, I just hurl abuse at it randomly.

Article is paywalled but had no idea either.

I’ve just not used fb or insta for years but I asked my wife who uses both regularly and she hadn’t heard of it either.

Oh they definitely are. AI aside if people want to see change in the way these companies operate, they need to abandon anything they make.

SPICY CHICKEN SANDWICH REPRESENTATION