AWS has said that they believe the future of GenAI is an assortment of different models. I’m inclined to believe otherwise—one (or two, for images as well as text) great models that are highly versatile.
@Quinnypig at the moment, my team drifts toward gpt-4-turbo for everything unless it’s too expensive or slow. i imagine you’re right
RAG is all you need. if RAG doesn’t work it’s a hella lot cheaper to just go toward traditional ML algorithms rather than train an entire LLM (that’s insane)
@Quinnypig It’ll be one or two base models with an assortment of derived models tweaked for specific purposes, kind of like how Stable Diffusion has been working.
Add comment