Apple has very little ML talent. They're basically resting on their laurels in the phone market.

That's what I thought too; but all the upthread and peer commments keep asserting Apple is a (real, not just hypothetical/potential/future) player in the space.

Apple went too early to ML for Siri.

Siri was "decently good" when it was basically a voice-activated command line, and you could figure out what you needed to run to get it to do what you wanted.

When they added ML to it, it entirely went to shit, and it has never recovered.

Perhaps Apple will be able to use an LLM to make it finally not shit.

well yeah, people on this website opine a lot about things they know little about. they're starting to make moves, but they don't have a large talent base.

Apple will start using this tech now that it has been commoditized and since it is a big consumer tech company they will put a nice polish. But it's not where big research is being done.

Apple refuses to pay top of the market. They'll have to get really lucky with someone who grows from within.