Comment by istvanmeszaros

If it is going to be so cheap that you can run a good LLM model locally, it will certainly cheaper to run it at scale in the cloud.

Although, I hope I won't need be dependent on some cloud provider. So I am looking for a day that this comes.