If LLMs have already reached the limits of available training data, what the hell is the point of all these city-sized data centers going up? What are they hoping to achieve with that much processing power?
@cinebox LLMs are also incredibly computationally expensive when producing output, not just during training
A small, community‐oriented Mastodon‐compatible Fediverse (GlitchSoc) instance managed as a joint venture between the cat and KIBI families.