Have you ever heard the idiom “ripping off the band-aid? With thanks to Wiktionary for the definition, it means:
To perform a painful or unpleasant but necessary action quickly so as to minimize the pain or fear associated with it.
That’s not what Apple’s doing right now with the end stages of its Intel-to-Apple Silicon transition, which kicked off five years ago (fifteen years, ironically, after its previous announced transition to x86 CPUs, two decades ago). And although I’m (mostly) grateful for it on a personal level, I’m also annoyed by what’s seemingly the company’s latest (but definitely not the first, and probably also not the last) example of “obsolescence by (software, in this case) design”.
Nearing the end of the line for Intel
Upcoming MacOS “Tahoe” 26, publicly unveiled at this week’s Worldwide Developers Conference (WWDC) and scheduled for a “gold” release later this year (September or October, judging from recent history), still supports legacy Intel-based systems, but only four model/variant combos. That’s right; four (and in those few cases still absent any Apple Intelligence AI capabilities):
- MacBook Pro (16-inch, 2019)
- MacBook Pro (13-inch, 2020, Four Thunderbolt 3 ports)
- iMac (27-inch, 2020)
- Mac Pro (2019)
My wife owns the first one on the list, courtesy of a Christmas present from yours truly last year. I’m typing these words on the second one. The other two are the “end of the line” models of the Intel-based iMac and Mac Pro series, both of which subsequently also switched to Apple Silicon-based varieties. Not included, long-time readers may have already noticed, is my storage capacity-constrained 2018 Mac mini; its M2 Pro successor is already sitting on a shelf downstairs in storage, awaiting its turn in the spotlight (that said, I’ll probably cling to my 2018 model longer than I should in conjunction with OpenCore Legacy Patcher, even if only motivated by hacker curiosity and because I’m so fond of the no-longer-offered Space Gray color scheme…).
But let’s go back to the second (also my) system in the previous bullet list. Did it also seem strange to you that Apple specifically referenced the model with “Four Thunderbolt 3 ports”? That’s because Apple also sold a 2020 model year variant with two Thunderbolt 3 ports. If you compare the specs of the two options, you’ll see that there’s at least some tech rationalization for the supported-or-not differentiation; the two-port model is based on a 8th-generation “Coffee Lake” Intel Core i5 8257U SoC, while my four-port model totes a 10th-generation “Ice Lake” Core i5 1038NG7. That said, they both support the same foundation x86 instruction set, right? And the integrated graphics is Intel Iris Plus in generation for both, too. So…
A one-year delay in sentencing for (some of) the Dipert family system stable aside (my wife’s iPhone XS Max also gets the axe later this year), the endgame verdict for Intel on Apple is now coming into clear view. Apple confirmed that “Tahoe” is x86’s last hurrah; MacOS will be Arm-only beginning with next year’s (2026) spin. The subsequent 2028 edition will also drop Rosetta 2 dynamic software-translation support, so any x86-only coded apps will no longer run. And given these moves, along with Apple’s longstanding tradition of supporting not only the current but also prior two major O/S generations, it would make no sense for any developer to bother continuing to make and support “Universal” versions of apps (dual-coded for both x86 and Arm) once “Tahoe” drops off the support list in 2028…if they even wait that long, that is, considering that the predominant percentage of legacy Intel systems will be incapable of running a supported MacOS variant way before then. This forecast echoes what played out last time, when PowerPC was phased out in favor of x86.
The Liquid Glass interface
The other key announcement at the pre-recorded 1.5-hour keynote that kicked off this week:
which Apple itself condensed down to a 2 minute summary (draw your own “sizzle vs steak” conclusions per my recent comments on Microsoft and Google’s full and abridged equivalents):
involves the Liquid Glass UI revamp which, after conceptually originating with the two-years-back Vision Pro headset, now spans the broader product line. Translucency, rounded corners and expanded color vibrancy are its hallmarks; Apple even did a standalone video on it:
It looks…OK, I guess. On the Vision Pro, the translucency makes total sense, because UI elements need to be not only spatially arranged with respect to each other but also in front of the real-life scene behind them (and in front of the user), reproduced for the eyes by front-facing cameras and embedded micro-OLEDs. But for phones, tablets, watches, and the like…again, it’s OK, I guess. I’m not trying to minimize the value of periodic visual-experience refreshes, mind you; it’s the same underlying motivation behind re-painting houses and the rooms inside them. It just feels not only derivative, given the Vision Pro heritage, but also reactionary, considering that Google announced its own UI refresh less than a month ago (Android 16 is apropos downloading to my Pixels as I type these words, in fact), and Samsung had unveiled its own in January (six months later than originally promised, but I digress).
The iPad (finally?) grows up
There is, however, one aspect of Apple’s UI revamp that I’m very excited about, but which ironically has little (but not nothing) to do with Liquid Glass. For many years now, iPads (particularly the high-end Pro variants) have offered substantial hardware potential, largely untapped by the platform’s unrealized software capabilities. Specifically, I’m talking about Apple’s ham-fisted Split View and Slide Over schemes for (supposedly) unlocking multitasking. Frankly, the only times I ever used either of them were accidental, and my only reaction was to (struggle to figure out how to) undo whatever UI abominations I’d unintentionally activated.
Well, they’re both gone as of later this year. In their place is proper MacOS-like windowing, with menu bars, user customizable window locations and sizes (not to mention more simultaneously displayed windows), and the like. Hallelujah. Reiterating a point I’ve made before (although software imperfections blunted its at-the-time reality), Apple will need to be careful to not cannibalize its computer sales by tablet sales going forward. That said, as I’ve also previously noted, if you’re going to get cannibalized, it might as well be by yourself, not to mention that tablets offer Apple more competitive isolation than do computers.
Deep learning (local) model developer access
Apple also this week announced that it was opening up developer access to its devices’ locally housed deep learning inference models for use by third-party applications. Near term, I’d frankly be more enthusiastic about this move if the models themselves were better. That said, given that we’re talking about “walled garden” Apple here, they’re the only game in town, so I guess something’s better than nothing. And longer term, Apple now clearly realizes it’s behind its competitors in the AI race and is revamping its management and dev teams in response (not to mention dissing its competitors in the presumed hope of slowing down the overall market until it can catch up), so circumstances will likely improve tangibly here, in fairly short order.
Too much…too little…just right?
By the way, I’m sure many of you have already noticed the across-the-board naming revamp of the various operating systems to a consistent “dominant model year” approach…i.e. although the new versions will likely all roll out later this year, they’ll be majority-in-use in 2026. Whatever 
(in all seriousness, the numerical disparity between, for example, current-gen iOS 18, MacOS 15 and WatchOS 11 likely resulted in at least some amount of consumer confusion).
Broadly speaking, while I’m not trying to sound like Goldilocks with the header title of this concluding section of my 2025 WWDC coverage, I am feeling a bit of whiplash. Last year, Apple’s event was bloated with unrelenting AI hype (therefore my title “jab”), much of which still hasn’t achieved even a semblance of implementation reality even a year later (to developer and pundit dismay alike). This year, it felt like the pendulum swung (too far?) in the opposite direction, with excessive attention being drawn to minutia such as newly added gestures for AirPods earbuds and the Apple Watch (not that I can even remember the existing ones!), and no matter that the AI-powered real-time language translation facilities are welcome (albeit predictable).
Maybe next year (and, dare I hope, in future years as well) Apple will navigate to the “middle way” between these extremes. Until then, I welcome your thoughts on this year’s event and associated announcements in the comments!
—Brian Dipert is the Editor-in-Chief of the Edge AI and Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.
Related Content
- Microsoft Build 2025: Arm (and AI, of course) thrive
- The 2025 Google I/O conference: A deft AI pivot sustains the company’s relevance
- The 2024 WWDC: AI stands for Apple Intelligence, you see…
- The Apple 2023 WWDC: One more thing? Let’s wait and see
- Apple’s 2022 WWDC: Generational evolution and dramatic obsolescence
- Apple’s WWDC 2021: Software and services updates dominate
- Workarounds (and their tradeoffs) for integrated storage constraints
The post The 2025 WWDC: From Intel, Apple’s Nearly Free, and the New Interfaces Are…More Shiny? appeared first on EDN.
