Browser LLMs

[louis030195]

Hey, have you considered client side LLMs, seems like the 'ai' lib is very server oriented? Been waiting for a while for this to arrive and it seems some people run llama2 in the browser now

I assume llama2 will never run in edge due to the size?

[MaxLeiter]

I don't think this is something we're particularly interested in supporting with the SDK any time soon. That being said, the field is rapidly evolving so I don't want to shut anything down.