You could say the same about all self-hosted software, teams with billions of dollars to produce and host SaaS will always have an advantage over smaller, local operations.
It only has to be good enough to do what we want. In the extreme, maybe inference becomes cheap enough that we ask “why do I have to wake up the laptop’s antenna?”
There might be also local/global bias strategies. A tiny local model trained on your specific code/document base may be better aligned to match your specific needs than a galaxy scale model. If it only knows about one "User" class, the one in your codebase, it might be less prone to borrowing irrelevant ideas from fifty other systems.
We're already very, very close to "smart enough for most stuff". We just need that to also be "tuned for our specific wants and needs".