Apple never claimed to be an open ecosystem that allowed users to install anything they liked without interference.

Actually, from 2007 (first iPhone) to 2008 (the App Store), they kind of did.

Apple made it perfectly clear that iDevices would be a walled garden when they released the SDK for third party applications.

That’s retconning at its best - there was no support for 3rd party apps then, just websites.

That's the point, in the beginning there were not supposed to be native apps, just web apps, some of them advertised in Apple site but not otherwise privileged.

The point is that customers who bought an iDevice knew in advance that they were buying into a walled garden ecosystem with a web browser.

It was Google that lied to the public about Android being an "open" ecosystem that allowed you to do anything you liked without interference.

Offline speech recognition (using an open weights model like Whisper) is a good example of a feature that you can run on your NPU.

Sure, but it's something you can run on the NPU, not something you must run on the NPU. Whisper will run just fine on your CPU or GPU as well - as will the decades of offline speech recognition tools before.

So what's the actual feature? Slightly higher performance? Slightly lower latency? Slightly longer battery life? And that for something most people are rarely - if ever - going to use?

The battery life, if you use enough AI, is the main feature.

Local speech recognition is something useful that runs well on a (power efficient) NPU.

Dragon Naturally Speaking used to cost real money and wasn't as accurate as an open model like Whisper.

> On one hand it is kinda radioactive for a former Apple engineer to work on this

If this were the case, then Wine and its forks like Proton would have the same issue.

Former Microsoft employees have contributed to Wine in the past.

The speed with which we went from pearl clutching that Apple hasn't added as many AI features as Google and Microsoft have to pearl clutching that they have finally done so is amusing.

Meanwhile, Google just announced "laptops" that are locked down to the Play Store, and the sound of crickets is palpable.

> Google has confirmed to Ars that Googlebooks will enforce developer verification requirements for apps similar to Android phones. Under this system, only apps from developers who have verified their identities with Google will be eligible for installation

https://arstechnica.com/gadgets/2026/09/googlebook-laptops-l...

I have met zero people and read zero accounts of people wishing there was "more AI" in Apple devices.

Your first sentence of this reply doesn't make any sense to me.

And I don't see how your second sentence relates to the first.

It is a few paragraphs down on the linked page:

> The following features are known not to work: suspend, display out on the USB-C ports (HDMI works), and Thunderbolt/USB4. Additionally, power consumption may be higher than expected (not very relevant for a desktop); we have set conservative limits for max speaker volume for safety reasons; and shutdown/reboot are flaky (you can just use the power button anyways). Every feature not listed here works and has been tested.

The issue is that the moment you want to run the more capable models that will no longer fit in a single 5090's memory, performance falls off a cliff.

The M5 generation added tensor instructions to the GPU cores.

> What exactly the M7 improvements are who knows.

> Apple's planned M7 Ultra chip is being designed to support up to 1.5 TB of unified memory and to push AI performance toward the class of Nvidia's Blackwell accelerators

https://www.tomshardware.com/tech-industry/semiconductors/ap...

The next Ultra, supposedly on deck in 2028:

> Apple's planned M7 Ultra chip is being designed to support up to 1.5 TB of unified memory and to push AI performance toward the class of Nvidia's Blackwell accelerators, according to a new Bloomberg report published by Mark Gurman...

Apple plans to release a base M6 chip this fall for entry-level Macs... a base M7 in the first half of 2027, M7 Pro and M7 Max at the end of 2027, and the M7 Ultra in 2028.

https://www.tomshardware.com/tech-industry/semiconductors/ap...