Disagree with the premise and conclusion. Most of the consumer-facing "AI" products out there suck. Copilot sucks, badly. Gemini is middling. We'll see what the new Alexa/Claud AI brings, but really, playing music and setting timers, hardly impressive. Alexa can whisper at you; wow, so impressive (not).
Perhaps Apple is taking its sweet-ass time rolling out their assistant, because they know most of these "AI assistants" are useless slop. And that's the truth; if you're even remotely efficient with a keyboard and shortcut toggles, the AI interface slows down users and gets in the way. How many of you have turned off Apple Mail "categories"? That was the first thing I did when it came out: off.
And this is just the kindergarten-level AI assistants for noobs. Get to some real coding tasks, and the LLMs leave a lot to be desired. These models can't even write good code in Z80, let alone Linux, C++, Java, etc.
Apple executives, including the C-suite, were really, really smart to not dump billions into developing an Apple LLM and AI apps no one wants. DeepSeek R1 (and soon R2) has already proved the folly of burning money on this. Llama 3 is open-source. How long until "OpenAI," a misnomer if there ever was, has to flush their model onto the open market too?
I think the slow roll is the way to play this. Didn't MSFT's Nardella say, just last week, the GDP impact of all this AI spending is, thus far, negligible? So how the F is Apple purportedly behind?
Finally, look at the hardware/firmware side. Silicon, the bish that counts. Apple is killing it in speed, efficiency, a new C1 chip/radio, M4 and M5 hardware coming out at a rapid clip. I don't understand this hair-on-fire narrative. At all.