I'd rather they pick up the pace, but I also don't see a "need" to be at the head of the pack either.

They should be doing better, focusing on easy wins, siri and etc. But if they get there they get there and nobody cares after the fact.

It's not a race they need to win. It's one they can take the long play on. The other companies at play? Their entire existence is the race.

Apple is very smart and very focused.

Apple is ruthlessly focused on extracting maximum current and long-term value from its high-end operating system which appeals to creative professionals, software developers, rich people, regular home users, and both the tech savvy and not very tech savvy.

In order to support its operating system experience, Apple has to make hardwarre.

Yes, Apple has a quite large ecosystem, but they do things slowly, intentionally, and by the book.

You know what's really not cool that the other companies all looked the other way on and are too big to fail on? Massive, massive copyright infringement. I suspect that Apple's legal team made the ethical choice to not do that.

Apple knows that models and inference are a race to the bottom, since the pirated content is broadly available, and since the model building is pretty openly published, and since there are several players already, and the hardware is basically commodity.

Apple can still be the glue that offers that platform on which people develop and make use of AI related technology.

The OP's assertion seems to miss that, for local AI, if one wants to spend under $10K or under $20K on an inference setup, or even under $1000 for that matter, (arguably) by far the the simplest and most efficient tokens per watt or tokens per dollar will come from buying Apple machines.

Apple appears to be simultaneously:

  - avoiding truly silly levels of hype train buy-in
  - letting market go where it wants to
  - supporting that with AI where it make sense for them (on-device for user privacy and easy to use where applicable; connecting users to interfaces to other companies where that makes sense)

To draw an analogy, Apple could make a Kagi/Google, or a GSheets/Excel, but they're not focused on those kind of things. Engineering stuff is expensive. Apple apparently runs a very tight ship.

Honestly, I think AI has little place in an OS where the user is well acquainted with it and in full control. Many may disagree, and that's fine. But maybe Apple is willing to let those people use another operating system?

* Disclaimer, the above are just my thoughts, I do not have any special knowledge.

In a "race" marked with significant gaffes, I'll wait for quality.

If it exists.

same as everything else.

sit back and enjoy the 30%