May have more resources but I suspect googles servers also have some fantastic caches of common misspelling based on what your locale is and what most people are searching for.
A lot of that derived from aggregate data of searches that were more likely to be engaged with or lead to click through.
That kind of data could be used on a phone if they wanted to download a large cache of it every so often but since it could be time and location dependent.
Olzempik
Right now… Olympics.
A few months from now… you’re in Washington state. Maybe you want Olympus?
Recent searches for diabetes… Ozempic?
Some of that is just too hard to reasonable calculate and cache on a mobile device.
hoshikarakitaridia@lemmy.world 3 months ago
I would add that it’s way more efficient because Google doesn’t have everyone using it mistyping all at once. So they can save up CPU cycles until they need them.