Steve Barnes' World of Happiness

Blog

macOS Tahoe: the Mac's most wayward update.

It's an odd feeling to approach a WWDC keynote, one of the most inspiring mornings of the year, when some of the greatest teams in software used every day around the world introduces its most cutting-edge decisions and creations, with even mild resentment.

WWDC is nominally for developers, and lately, when developers for Apple platforms are concerned, they voice it. I've developed for Apple platforms for 15 years now, I suppose I've just realized – it still bears the feeling of an aspiration. But the mild resentment isn't because I'm a developer; it's because I'm a long-time ordinary user and cherisher of the Mac.

There's been plenty of controversy around Apple lately, too. A few years ago, the headlines were about sphinxishly shifting requirements in the European Union as it legislated away at App Stores and rules around default browsers. Then came the "Apple Intelligence" hype – software features the iPhone 16 was "built for," which have yet to appear as the iPhone 18 approaches.

I didn't really mind any of this. I found the EU drama amusing and ridiculous – in some cases the result has been that Apple has withheld features there – but I'm not there, so I'm at least not affected. "Apple Intelligence" refers to features based on the installation of large system-managed on-device models, at least respectfully cordoned behind a single switch. Also easy enough to forget. What I minded was the questionably-guided revision to the Mac's beloved, refined interface. And it became clear I wasn't alone.

People have referred to the collective visual and presentational changes to the Mac's interface by the term "Liquid Glass." (Phrases like "I HATE LIQUID GLASS!!" have been online parlance for not loving Tahoe's appearance generally.) In marketing and with developers, Apple makes the distinction between the "new design" (with lower case letters), meaning the approach to interface layout and behaviour, and "Liquid Glass" (a proper noun), referring specifically to the virtual material.

I don't hate Liquid Glass. The "material" itself is very cool, precisely warping and refracting what it overlays, sometimes even splitting the whites subtly into thin rainbows around its curved edges, all in real time, all with no need for anyone – not even developers who adopt it – to think about how it happens. It's the kind of thing that would have audibly blown audiences' minds in Apple keynotes of a prior decade. It's a real example of Apple's traditional feat of making something complicated look simple, a kind of achievement so commonplace today that it spent the year being taken all but completely for granted. But macOS Tahoe was, least of all among its companion platforms this year, really about Liquid Glass.

So let's move to a term of my own:

"Rampant circularization."

Tahoe's "new design" is marked by an effort of leaning heavily into circles and circular edges, which Apple has characterized "harmony."

System-provided windows all over macOS now have circular corners which have struck many as comically or toyishly large. Where's the supposed "harmony"? Well, start with the traditional "close window" button at a window's upper-leading corner. That button is, as before, often housed within a sidebar. But the "new design," rather than dividing the window vertically into a sidebar and main area, now presents the sidebar as a rounded rectangle which is overlaid and inset within its alloted area. Due to that inset, its corners have their own radius: a value between the radius of the containing window's corners (the longest) and the radius of the close button (the shortest). Their corners are concentric arcs with increasing radii. If the window was housed in some sort of containing "window for windows," you'd guess it would have its own, even more bulbous, corners.

The most perceptibly immediate effect produced by this inset sidebar, though, is not a feeling of harmony. It's visual complication. The sidebar has gained three new edges that didn't need to be there, simply to hug the window's edges. The next effect is to prompt the inquiry: "Why? The Mac, nay, Apple, has always striven to identify and purge the extraneous, the redundant, to simplify. If there is sanity in this, where is it?"

I suppose this brings me to the main theme of my year with Tahoe. Every year to this point, since the dawn of OS X, while I've missed the odd touch or a trait as it faded into the past, and while the new has sometimes been briefly uncomfortable, I've understood the reasoning behind the design updates. I've perceived the taste. So as Tahoe began to seem more confusing and stressful than tasteful, I wanted to be charitable: if the reasoning didn't strike me, I should at least search for it.

Alan Dye's original introduction of Liquid Glass nodded to visionOS as its inspiration. Virtual glass controls and surfaces make a lot of sense in the Vision Pro, which has access to the appearance of real objects and backdrops. The much-criticized reduced readibility of text on glass buttons and overlays, if not defeated, is at least countered by depth perception. And the overlaidness of such controls would support their raison d'arriver on iPhone and iPad, where they sit inside the device's corners, atop the document or Web page, rather than claim an entire strip of screen space for an enclosing toolbar.

So I gathered that sidebars, which govern and pertain to the entirety of the containing window on both iPad and Mac, have been conceptually promoted to "controls." Controls are overlaid; sidebars are controls; ergo, sidebars are overlaid. Therefore they're inset and have rounded corners and more visual complexity. It's still visually chaotic, but at least there's a guess.

If it had stopped there, it might have been a dismissable compromise; one of those briefly uncomfortable but acceptable transitions. But that was just the start of a list. The definition of the middle-length sidebar corner radius seems shared by standard toolbar buttons all across the top of a window. More system buttons, such as in alerts, are now pill-shaped rather than rounded rectangles. Circles upon circles. I remember being routinely irked by the still-recent "modern" trend of patently circle-based fonts still flagshipped by Google and Twitch, and chuckling them away, thinking "at least Apple will never do that."

And then, there's the specular edge lighting.

Liquid Glass, when it's flowing, bouncing and morphing? Again, pretty cool in itself. When Alan Dye moves on to mention that app icons are now made of "multiple layers of Liquid Glass"? Well, app icons don't flow, bounce or morph. So, what does that mean?

Well, when updating from iOS 18 to iOS 26 – particularly for those like me, as someone who always leaves their phone in Dark Mode with a black background – it means that a tranquil Home Screen of lovely icons, their backdrops shaded a consistent dark grey by the system (a feature which had arrived only just the previous year), was suddenly a Home Screen of lovely icons, consistently shaded by the system, each also consistently outlined with a thin but glaring asymmetrical stroke of light around the edge. The previous year was about long-awaited home screen customization, and suddenly, here was an additional aspect that was deliberately never customizable. To add insult, the icons are drawn without the specular edge while smoothly animating from the foreground app to the Home Screen, then, after settling in, they regain the edge in a single frame.

Some iOS devices, like newer iPads, would cleverly shift the edge's asymmetry with the device's motion, suggesting physically real glass squares – but only some devices. And worse: only some icons! On my iPad mini, icons in the Home Screen's main grid shift while icons in the Dock remain gleaming at a fixed angle.

And, unpleasantly: this specular lighting imposed itself on Tahoe.

And, more unpleasantly: the same with widgets.

And, even more unpleasantly: widgets are also treated with additional styling. A "glassy" sheen instead of the pure shade or gradient the developer wanted, I guess.

I thought Apple – if not all of us – had surpassed this. Jony Ive's particular marks don't need to last forever; he left by choice. But he offered what once seemed an effortful revelation, a true stepping-back and rethinking about how computer software was displayed, that we needn't pretend everything on a screen is a real material. Windows don't all need thick "metal" title bars. A phone app doesn't always need a bulky border to contain text when the text's area is already surrounded by a phone. A computer's display area is made of pixels, of freely scuptable regions; not thread, wood, fur, plastic, stone. Software can be made of what computer displays are made of, can still even have the palpability and character of objects without pretending to be real objects. The industry, the visual language, the culture, felt like it grew up a little. So, what is Tahoe evincing? A straddling of both sides of that transition, sort of sloshed together without really considering it?

On one hand, I've gotten used to it on some level. On Mac, a row of circular toolbar buttons is almost stressful to look at, but pill-shaped groups of buttons look somewhat decent and organized, despite the presence of the requisite number of circularized ends.

On the other hand? Every time I glipmse an iPhone running iOS 18, or a screen shot of a Mac running Sequoia, I feel a pang of excitement and envy. It's exactly what you're supposed to feel when you see the upgrade, not the "downgrade."

To sum it up, I recall a Steve Jobs quote I haven't heard anyone else apply to this situation:

"So, it kinda just comes down to: what do you want to be looking at all day long?"

And this isn't even to mention the difficulty in resizing by the extra-rounded window corners, the problems they create when genuinely hiding the scroll thumb or the comparatively reasonable corners of elements in Web pages. The glyphs in the menus. If you're an enthusiast, you know at least a little about all of this.

In short: last year's WWDC week left me with a pang of concern. A year later, it proved merited.

I called my feeling mild resentment, going into tomorrow. But what it isn't? It's not anger, not doomsaying, not a lack of hope. Not what I've heard some online neighbours are experiencing.

That's an important point of perspective: compared to other computer users, Mac users are all but spoiled with excellence. Even in Tahoe, Spotlight is deeply reinvigorated. Live Activities are available. Control Center is hugely customizable. Ambient background sounds arrive at the system level. Despite their edges, app icons gain the uniform dark option. (And the creation of app icons for developers, traditionally an extreme chore, was elegantly and utterly simplified by a new utility app.) Folder icons are customizable with pleasant colours, emoji, and animations. A simple native Games app. The usual security advances and improved APIs. Even with Tahoe, the Mac is easily still the best personal computer. The main problem was that this on-par progress was so handily upstaged by the over-zealous design changes, and they raise concern only because the Mac has thus far been so beautifully and steadily refined by time and wisdom.

And there is at least some intention behind the "new design." visionOS has some "back to the Mac"-style stuff to legitimately offer the other platforms. There are now six platforms, and many people use many of them; consistency shouldn't cramp any device's style but should be considered for its value. Apple knows where they're headed with these platforms, and at least some of what we're enduring now will make at least make slightly more sense in retrospect.

Tahoe's appearance isn't a plane crash. It's a turbulent swerve, the type Apple has occasionally had, sometimes for a while, when they dare to try something. It takes time to regain stability, but it's in team at Apple to regain it.

And there are specific signs of this. No tweak between any two minor versions of Tahoe has been fully satisfying, but they've all been changes in the positive direction. The aforementioned Alan Dye, the public face of Liquid Glass and the "new design," left Apple in what seemed, even publicly, a surprise to the rest of Apple. His destination (Meta), and his public announcement (a conspicuous Steve Jobs quote on Instagram, badly styled and missing punctuation, in no way befitting a "Chief Design Officer") raise questions about where his heart ever really was, suggesting one wrong has already been righted. Steve Lemay, his successor in the role, is internally referenced with excitement by all the hearsay I've found. macOS 27 will the baggage of support for Intel, making it analogous to the beloved Snow Leopard, packed with features and refinements but often rememembered for a quiet, welcome advance in stability. And the shift comes with other leadership transitions, including the extremely implicative succession of Tim Cook as CEO by John Ternus.

No one expects a one-eighty on the Mac's design, but I expect that overt zeal to begin into something more reasonable, letting the Mac's cumulative greatness speak again for itself.

After the EU drama and the Apple Intelligence hype, I didn't really need a year of concern over something truly internal to Apple; something it fully controlled and could have done more carefully. A quiet year of staying the course would have been just right. But this coming year may indeed prove to be it.

My main wish for WWDC is to reach September, the day I have the option to install macOS 27, and observe no trace of trepidation; to be excited again to upgrade in the correct direction.