[kenjackson]

It may be that offline LLMs will be common in a few years.

[tonypace]

That is definitely a potential issue, but so far any text models that run on laptops are tremendously slow. Still, something to look out for.