[YmiYugy]

There is a place for local models. Typing prediction, speech to text, etc. But I have yet to see uses of local LLMs that are all that compelling.

[Aeroi]

Imagine in 10 years where local models will be, we're still in the 1st inning.

[bigyabai]

The transformer architecture will be 10 years old, next year. We're in the second or third inning of LLMs right now, depending on how much credit you give ELMo and BERT.