A Postscript on the Singular Nature of Mark Gurman’s Reporting

Sunday, 16 March 2025

My post Friday commenting (read: wise-cracking) on Mark Gurman’s explosive report on an all-hands Siri team meeting at Apple was begging for a bit of meta commentary on the reporting itself. But I’ve been doing so much of that regarding Gurman lately that I thought it best to hold it for a postscript. Here’s that postscript.

Both of these things are true:

Mark Gurman is a singular reporter in the Apple media sphere. He publishes an extraordinary number of exclusives, both regarding leaks of upcoming products, and leaks like this Siri team meeting.
Gurman often gets things wrong, and when he does, he never acknowledges those mistakes, let alone corrects them. He also tries to take credit for having called things he completely missed. He’s not an oracle but presents himself as one. And he writes for a publication, Bloomberg, that shares his insistence on never acknowledging let alone correcting mistakes, even massive ones. What gives me such joy pointing out his boners isn’t that he made them in the first place but that he refuses to acknowledge they happened, presenting an air of infallibility with a provably fallible track record.

In short, I do actually suspect — but can claim zero sources familiar with the matter to confirm — that Gurman hangs his toilet paper in an improper underhand fashion.

So let’s just examine how extraordinary and singular Gurman’s Friday report was. Nobody else reported on this meeting. Every other article about it — including mine — was commenting on Gurman’s exclusive report about the meeting. I’ve not seen one other report even confirming the meeting took place, let alone describing it in detail, replete with copious quotes from Siri senior director Robby Walker, who, according to Gurman, led the meeting. Not one. I’m not pointing that out to cast suspicion that the meeting did not take place or that Gurman’s report cast it inaccurately or that his direct quotations were not, in fact, direct quotations. I’m pointing out just how singular and extraordinary Mark Gurman is in this sphere. If it wasn’t for Gurman’s report we, outside Apple (and probably outside the Siri team inside Apple) wouldn’t even know the meeting occurred.

How did Gurman not only get the scoop on this meeting, but copious direct quotes from Walker’s remarks to the team? Well, it was “according to people with knowledge of the matter, who asked not to be identified because the gathering was private”. In other words, more than one member of the Siri team, and at least one of which either recorded the meeting surreptitiously and slipped the recording to Gurman, or at least one of whom takes notes at the pace and accuracy of a court stenographer. Either way, these sources — plural — surely knew how the meeting would make Apple look if it were to leak.

I’ve long made my opinions about Bloomberg’s institutional journalistic credibility well known. But I don’t think they’re bereft of credibility — it’s the fact that they are deservedly well-regarded that makes their refusal to ever admit their own glaring mistakes so notable. When a Gurman reports says “people” that means “more than one” and, I believe, he must be able to confirm to his editors that he got this information from more than one source. If he’s reporting direct quotes, I think that means he’s heard a recording. That’s extraordinary.

But I’d feel a lot better about our collective conventional wisdom regarding the nature of this particular all-hands Siri meeting if it had leaked to, and been reported on by, more than one reporter at more than one publication. ★

Something Is Rotten in the State of Cupertino

Wednesday, 12 March 2025

In the two decades I’ve been in this racket, I’ve never been angrier at myself for missing a story than I am about Apple’s announcement on Friday that the “more personalized Siri” features of Apple Intelligence, scheduled to appear between now and WWDC, would be delayed until “the coming year”.

I should have my head examined.

This announcement dropped as a surprise, and certainly took me by surprise to some extent, but it was all there from the start. I should have been pointing out red flags starting back at WWDC last year, and I am embarrassed and sorry that I didn’t see what should have been very clear to me from the start.

How I missed this is twofold. First, I’d been lulled into complacency by Apple’s track record of consistently shipping pre-announced products and features. Their record in that regard wasn’t perfect, but the exceptions tended to be around the edges. (Nobody was particularly clamoring for Apple to make a multi-device inductive charging mat, so it never generated too much controversy when AirPower turned out to be a complete bust.) Second, I was foolishly distracted by the “Apple Intelligence” brand umbrella. It’s a fine idea for Apple to brand its AI features under an umbrella term like that, similar to how a bunch of disparate features that allow different Apple devices to interoperate are under the “Continuity” umbrella. But there’s no such thing, technically speaking, as “Continuity”. It’s not like there’s an Xcode project inside Apple named Continuity.xcodeproj, and all the code that supports everything from AirDrop to Sidecar to iPhone Mirroring to clipboard sharing is all implemented in the same framework of code. It’s a marketing term, but a useful one — it helps Apple explain the features, and helps users understand them.

The same goes for “Apple Intelligence”. It doesn’t exist as a single thing or project. It’s a marketing term for a collection of features, apps, and services. Putting it all under a single obvious, easily remembered — and easily promoted — name makes it easier for users to understand that Apple is launching a new initiative. It also makes it easier for Apple to just say “These are the devices that qualify for all of these features, and other devices — older ones, less expensive ones — get none of them.”

Let’s say Apple were to quietly abandon the dumb Image Playground app next year. It just disappears from iOS 19 and MacOS 16. That would just be Apple eliminating a silly app that almost no one uses or should use. That wouldn’t be a setback or rollback of “Apple Intelligence”. I would actually argue that axing Image Playground would improve Apple Intelligence; its mere existence greatly lowers the expectations for how good the whole thing is.¹

What I mean by that is that it was clear to me from the WWDC keynote onward that some of the features and aspects of Apple Intelligence were more ambitious than others. Some were downright trivial; others were proposing to redefine how we will do our jobs and interact with our most-used devices. That was clear. But yet somehow I didn’t focus on it. Apple itself strongly hinted that the various features in Apple Intelligence wouldn’t all ship at the same time. What they didn’t spell out, but anyone could intuit, was that the more trivial features would ship first, and the more ambitious features later. That’s where the red flags should have been obvious to me.

In broad strokes, there are four stages of “doneness” or “realness” to features announced by any company:

Features that the company’s own product representatives will demo, themselves, in front of the media. Smaller, more personal demonstrations are more credible than on-stage demos. But the stakes for demo fail are higher in an auditorium full of observers.
Features that the company will allow members of the media (or other invited outside observers and experts) to try themselves, for a limited time, under the company’s supervision and guidance. Vision Pro demos were like this at WWDC 2023. A bunch of us got to use pre-release hardware and in-progress software for 30 minutes. It wasn’t like free range “Do whatever you want” — it was a guided tour. But we were the ones actually using the product. Apple allowed hands-on demos for a handful of media (not me) at Macworld Expo back in 2007 with prototype original iPhones — some of the “apps” were just screenshots, but most of the iPhone actually worked.
Features that are released as beta software for developers, enthusiasts, and the media to use on their own devices, without limitation or supervision.
Features that actually ship to regular users, and hardware that regular users can just go out and buy.

As of today — March 2025 — every feature in Apple Intelligence that has actually shipped was at level 1 back at WWDC. After the keynote, dozens of us in the press were invited to a series of small-group briefings where we got to watch Apple reps demo features like Writing Tools, Photos Clean Up, Genmoji, and more. We got to see predictive code completion in Xcode. What has shipped, as of today, they were able to show, in some functional state, in June.

For example, there was a demo involving a draft email message on an iPad, and the Apple rep used Writing Tools to make it “more friendly”. I was in a group of just four or five other members of the media, watching this. As usual, we were encouraged to interrupt with questions. Knowing that LLMs are non-deterministic, I asked whether, as the Apple rep was performing this same demo for each successive group of media members, the “more friendly” result was exactly the same each time. He laughed and said no — that while the results are very similar each time, and he hopes they continue to be (hence the laughing), there were subtle differences sometimes between different runs of the same demo. As I recall, he even used Undo to go back to the original message text, invoked Writing Tools to make it “more friendly” again, and we could see that a few of the word choices were slightly different. That answered both my explicit question and my implicit one: Writing Tools generates non-deterministic results, and, more importantly, what we were watching really was a live demo.

We didn’t get to try any of the Apple Intelligence features ourselves. There was no Apple Intelligence “hands on”. But we did see a bunch of features demoed, live, by Apple folks. In my above hierarchy of realness, they were all at level 1.

But we didn’t see all aspects of Apple Intelligence demoed. None of the “more personalized Siri” features, the ones that Apple, in its own statement announcing their postponement, described as having “more awareness of your personal context, as well as the ability to take action for you within and across your apps”. Those features encompass three main things:

“Personal context” — Knowing details and information about you from a “semantic index”, built from the contents of your email, messages, files, contacts, and more. In theory, eventually, all the information on your device that you wish to share with Siri will be in this semantic index. If you can look it up on your device, Siri will be able to look it up on your device.
“Onscreen awareness” — Giving Siri awareness of whatever is displayed on your screen. Apple’s own example usage: “If a friend texts you their new address, you can say ‘Add this address to their contact card,’ and Siri will take care of it.”
“In-app actions” — Giving Siri the ability, through the App Intents framework, to do things in and across apps that you can do, the old fashioned way (yourself) in and across apps. Again, here’s Apple’s own example usage:

You can make a request like “Send the email I drafted to April and Lilly” and Siri knows which email you’re referencing and which app it’s in. And Siri can take actions across apps, so after you ask Siri to enhance a photo for you by saying “Make this photo pop,” you can ask Siri to drop it in a specific note in the Notes app — without lifting a finger.

There were no demonstrations of any of that. Those features were all at level 0 on my hierarchy. That level is called vaporware. They were features Apple said existed, which they claimed would be shipping in the next year, and which they portrayed, to great effect, in the signature “Siri, when is my mom’s flight landing?” segment of the WWDC keynote itself, starting around the 1h:22m mark. Apple was either unwilling or unable to demonstrate those features in action back in June, even with Apple product marketing reps performing the demos from a prepared script using prepared devices.

This shouldn’t have just raised a concern in my head. It should have set off blinding red flashing lights and deafening klaxon alarms.

Even the very engineers working on a project never know exactly how long something is going to take to complete. An outsider observing a scripted demo of incomplete software knows far less (than the engineers) just how much more work it needs. But you can make a rough judgment. And that’s where my aforementioned hierarchy of realness comes into play. Even outsiders can judge how close a public beta (stage 3) feels to readiness. A feature or product that Apple will allow the press to play with, hands-on (stage 2) is further along than a feature or product that Apple is only willing to demonstrate themselves (stage 1).

But a feature or product that Apple is unwilling to demonstrate, at all, is unknowable. Is it mostly working, and close to, but not quite, demonstratable? Is it only kinda sorta working — partially functional, but far from being complete? Fully functional but prone to crashing — or in the case of AI, prone to hallucinations and falsehoods? Or is it complete fiction, just an idea at this point?

What Apple showed regarding the upcoming “personalized Siri” at WWDC was not a demo. It was a concept video. Concept videos are bullshit, and a sign of a company in disarray, if not crisis. The Apple that commissioned the futuristic “Knowledge Navigator” concept video in 1987 was the Apple that was on a course to near-bankruptcy a decade later. Modern Apple — the post-NeXT-reunification Apple of the last quarter century — does not publish concept videos. They only demonstrate actual working products and features.

Until WWDC last year, that is.

My deeply misguided mental framework for “Apple Intelligence” last year at WWDC was something like this: Some of these features are further along than others, and Apple is showing us those features in action first, and they will surely be the features that ship first over the course of the next year. The other features must be coming to demonstratable status soon. But the mental framework I should have used was more like this: Some of these features are merely table stakes for generative AI in 2024, but others are ambitious, groundbreaking, and, given their access to personal data, potentially dangerous. Apple is only showing us the table-stakes features, and isn’t demonstrating any of the ambitious, groundbreaking, risky features.

It gets worse. Come September, Apple held its annual big event at Apple Park to unveil the iPhone 16 lineup. Apple Intelligence features were highlighted in the announcement. Members of the media from around the world were gathered. That was a new opportunity, three months after WWDC, for Apple to demonstrate — or even better, offer hands-on access to the press to try themselves — the new personalized Siri features. They did not. No demos, at all. But they did promote them, once again, in the event keynote.²

But yet while Apple still wouldn’t demonstrate these features in person, they did commission and broadcast a TV commercial showing these purported features in action, presenting them as a reason to purchase a new iPhone — a commercial they pulled, without comment, from YouTube this week.

Last week’s announcement — “It’s going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year” — was, if you think about it, another opportunity to demonstrate the current state of these features. Rather than simply issue a statement to the media, they could have invited select members of the press to Apple Park, or Apple’s offices in New York, or even just remotely over a WebEx conference call, and demonstrated the current state of these features live, on an actual device. That didn’t happen. If these features exist in any sort of working state at all, no one outside Apple has vouched for their existence, let alone for their quality.

Duke Nukem Intelligence

Why did Apple show these personalized Siri features at WWDC last year, and promise their arrival during the first year of Apple Intelligence? Why, for that matter, do they now claim to “anticipate rolling them out in the coming year” if they still currently do not exist in demonstratable form? (If they do exist today in demonstratable form, they should, you know, demonstrate them.)

I’m not trying to be obtuse here. It’s obvious why some executives at Apple might have hoped they could promote features like these at WWDC last year. Generative AI is the biggest thing to happen in the computer industry since previous breakthroughs this century like mobile (starting with the iPhone, followed by Android), social media (Meta), and cloud computing (Microsoft, Google, and Amazon). Nobody knows where it’s going but wherever it’s heading, it’s going to be big, important, and perhaps profitable. Wall Street certainly noticed. And prior to WWDC last year, Apple wasn’t in the game. They needed to pitch their AI story. And a story that involved nothing but table-stakes AI features isn’t nearly as compelling a story as one that involves innovative, breakthrough, ambitious personal features.

But while there’s an obvious appeal to Apple pitching the most compelling, most ambitious AI story possible, the only thing that was essential was telling a story that was true. If the truth was that Apple only had features ready to ship in the coming year that were table stakes compared to the rest of the industry, that’s the story they needed to tell. Put as good a spin on it as possible, but them’s the breaks when you’re late to the game.

The fiasco here is not that Apple is late on AI. It’s also not that they had to announce an embarrassing delay on promised features last week. Those are problems, not fiascos, and problems happen. They’re inevitable. Leaders prove their mettle and create their legacies not by how they deal with successes but by how they deal with — how they acknowledge, understand, adapt, and solve — problems. The fiasco is that Apple pitched a story that wasn’t true, one that some people within the company surely understood wasn’t true, and they set a course based on that.

The Apple of the Jobs exile years — the Sculley / Spindler / Amelio Apple of 1987–1997 — promoted all sorts of amazing concepts that were no more real than the dinosaurs of Jurassic Park, and promised all sorts of hardware and (especially) software that never saw the light of day. Promoting what you hope to be able to someday ship is way easier and more exciting than promoting what you know is actually ready to ship. However close to financial bankruptcy Apple was when Steve Jobs returned as CEO after the NeXT reunification, the company was already completely bankrupt of credibility. Apple today is the most profitable and financially successful company in the history of the world. Everyone notices such success, and the corresponding accumulation of great wealth. Less noticed, but to my mind the more impressive achievement, is that over the last three decades, the company also accumulated an abundant reserve of credibility. When Apple showed a feature, you could bank on that feature being real. When they said something was set to ship in the coming year, it would ship in the coming year. In the worst case, maybe that “year” would have to be stretched to 13 or 14 months. You can stretch the truth and maintain credibility, but you can’t maintain credibility with bullshit. And the “more personalized Siri” features, it turns out, were bullshit.

Keynote by keynote, product by product, feature by feature, year after year after year, Apple went from a company that you couldn’t believe would even remain solvent, to, by far, the most credible company in tech. Apple remains at no risk of financial bankruptcy (and in fact remains the most profitable company in the world). But their credibility is now damaged. Careers will end before Apple might ever return to the level of “if they say it, you can believe it” credibility the company had earned at the start of June 2024.

Damaged is arguably too passive. It was squandered. This didn’t happen to Apple. Decision makers within the company did it.

Who decided these features should go in the WWDC keynote, with a promise they’d arrive in the coming year, when, at the time, they were in such an unfinished state they could not be demoed to the media even in a controlled environment? Three months later, who decided Apple should double down and advertise these features in a TV commercial, and promote them as a selling point of the iPhone 16 lineup — not just any products, but the very crown jewels of the company and the envy of the entire industry — when those features still remained in such an unfinished or perhaps even downright non-functional state that they still could not be demoed to the press? Not just couldn’t be shipped as beta software. Not just couldn’t be used by members of the press in a hands-on experience, but could not even be shown to work by Apple employees on Apple-controlled devices in an Apple-controlled environment? But yet they advertised them in a commercial for the iPhone 16, when it turns out they won’t ship, in the best case scenario, until months after the iPhone 17 lineup is unveiled?

When that whole campaign of commercials appeared, I — along with many other observers — was distracted by the fact that none of the features in Apple Intelligence had yet shipped. It’s highly unusual, and arguably ill-considered, for Apple to advertise any features that haven’t yet shipped. But one of those commercials was not at all like the others. The other commercials featured Apple Intelligence features that were close to shipping. We know today they were close to shipping because they were either in the iOS 18.1 betas already, in September, or would soon appear in developer betas for iOS 18.2 and 18.3. Right now, today, they’ve all actually shipped and are in the hands of iPhone 16 users. But the “Siri, what’s the name of the guy I had a meeting with a couple of months ago at Cafe Grenel?” commercial was entirely based on a feature Apple still has never even demonstrated.

Who said “Sure, let’s promise this” and then “Sure, let’s advertise it”? And who said “Are you crazy, this isn’t ready, this doesn’t work, we can’t promote this now?” And most important, who made the call which side to listen to? Presumably, that person was Tim Cook.

Even with everything Apple overpromised (if not outright lied about) at the WWDC keynote, the initial takeaway from WWDC from the news media was wrongly focused on their partnership with OpenAI. The conventional wisdom coming out of the keynote was that Apple had just announced something called “Apple Intelligence” but it was powered by ChatGPT, when in fact, the story Apple told was that they — Apple — had built an entire system called Apple Intelligence, entirely powered by Apple’s own AI technology, and that it spanned from on-device execution all the way to a new Private Cloud Compute infrastructure they not only owned but are powering with their own custom-designed server hardware based on Apple Silicon chips. And that on top of all that, as a proverbial cherry on top, Apple also was adding an optional integration layer with ChatGPT.

So, yes, given that the news media gave credit for Apple’s own actual announced achievements to OpenAI, Apple surely would have been given even less credit had they not announced the “more personalized Siri” features. It’s easy to imagine someone in the executive ranks arguing “We need to show something that only Apple can do.” But it turns out they announced something Apple couldn’t do. And now they look so out of their depth, so in over their heads, that not only are they years behind the state-of-the-art in AI, but they don’t even know what they can ship or when. Their headline features from nine months ago not only haven’t shipped but still haven’t even been demonstrated, which I, for one, now presume means they can’t be demonstrated because they don’t work.

‘So Why the Fuck Doesn’t It Do That?’

In May 2011, Fortune published an extraordinary look inside Apple by Adam Lashinsky, at what we now know to be the peak, and (alas) end, of the Steve Jobs era. The piece opens thus:

Apple doesn’t often fail, and when it does, it isn’t a pretty sight at 1 Infinite Loop. In the summer of 2008, when Apple launched the first version of its iPhone that worked on third-generation mobile networks, it also debuted MobileMe, an e-mail system that was supposed to provide the seamless synchronization features that corporate users love about their BlackBerry smartphones. MobileMe was a dud. Users complained about lost e-mails, and syncing was spotty at best. Though reviewers gushed over the new iPhone, they panned the MobileMe service.

Steve Jobs doesn’t tolerate duds. Shortly after the launch event, he summoned the MobileMe team, gathering them in the Town Hall auditorium in Building 4 of Apple’s campus, the venue the company uses for intimate product unveilings for journalists. According to a participant in the meeting, Jobs walked in, clad in his trademark black mock turtleneck and blue jeans, clasped his hands together, and asked a simple question:

“Can anyone tell me what MobileMe is supposed to do?” Having received a satisfactory answer, he continued, “So why the fuck doesn’t it do that?”

For the next half-hour Jobs berated the group. “You’ve tarnished Apple’s reputation,” he told them. “You should hate each other for having let each other down.” The public humiliation particularly infuriated Jobs. Walt Mossberg, the influential Wall Street Journal gadget columnist, had panned MobileMe. “Mossberg, our friend, is no longer writing good things about us,” Jobs said. On the spot, Jobs named a new executive to run the group.

Tim Cook should have already held a meeting like that to address and rectify this Siri and Apple Intelligence debacle. If such a meeting hasn’t yet occurred or doesn’t happen soon, then, I fear, that’s all she wrote. The ride is over. When mediocrity, excuses, and bullshit take root, they take over. A culture of excellence, accountability, and integrity cannot abide the acceptance of any of those things, and will quickly collapse upon itself with the acceptance of all three. ★

Image Playground would make a ton of sense not as a consumer-facing app, but as an example project for developers. Long ago, Apple used to share the source code for TextEdit as an example project for Mac developers. (TextEdit is actually a low-key great application, though. It’s genuinely useful, reliable, and understandable.) Apple shares tons of sample code at WWDC each year. Image Playground would be a great sample project. The silly app icon even looks like something from a WWDC sample project. What Image Playground is not is a credible useful generative AI tool. Yet Apple keeps talking about it — and showing it off in new hardware demonstrations — like it’s something they should be proud of and that anyone might credibly use for real-world work or even personal purposes. Image Playground does exemplify just how state-of-the-art the generative AI features are in Apple Intelligence, but not in the way Apple seems to think. ↩︎
Skip to the 53-minute mark of Apple’s September “It’s Glowtime” event introducing the iPhones 16, and it’s Craig Federighi who says the following:

“Siri will be able to tap into your personal context to help you in ways that are unique to you. Like pulling up the recommendation for the TV show that your brother sent you last month. And Siri will gain onscreen awareness. So when your friend texts you about a new album, you’ll be able to simply say, ‘Play that.’ And then you’ll be able to take hundreds of new actions in your apps, like updating a friend’s contact card with his new address, or adding a set of photos to a specific album. With Siri’s personal context understanding and action capabilities, you’ll be able to simply say, ‘Send Erica the photos from Saturday’s barbecue’, and Siri will dig up the photos and send them right off.”

That’s about 40 seconds of keynote time I bet Federighi regrets — and that I suspect he was skeptical about including. It’s telling though, that unlike WWDC, Apple didn’t show those features or spend even a full minute talking about at the iPhone 16 event — despite the fact that, ostensibly, those features should have been three months closer to shipping than they were in June. Federighi’s title is SVP of software, and Apple Intelligence and Siri are “software”, but John Giannandrea (SVP of machine learning and AI strategy) is Federighi’s peer, not subordinate, on the org chart — both report directly to Tim Cook — and is responsible for Siri and Apple Intelligence. Why it was Federighi, not Giannandrea, pitching those features in the iPhone 16 event keynote almost certainly comes down to Federighi’s presentation skills and stage presence, not responsibility for the features themselves. But who’s going on camera to pitch these features and promise their future availability the next time? ↩︎︎

A New System-Wide UI Look for iOS — Let Alone MacOS, Too — Would Be a Huge Deal

Tuesday, 11 March 2025

Ben Lovejoy, writing at 9to5Mac:

Our editor-in-chief Chance Miller wryly commented that a radical new look would serve as a great way to distract from the ever-slowing progress on the new Siri. But in truth, I think many more Apple users will be wowed by a new look than would ever care about Siri.

If that’s the way it works out — with a new visual look drawing attention from lackluster progress on the AI front — surely the timing will be coincidental, but some accidents are happy accidents, as Bob Ross used to say.

Sure, the new look introduced with iOS 7 was a massively controversial one, and many thought that then Apple design chief Jony Ive should never have been allowed anywhere near the software side of the business. But love it or hate it, you certainly couldn’t ignore it.

The same will be true of a new 3D look, which might even include some (much more modern) skeuomorphic elements. Probably as many will hate it as love it when it’s first introduced, as that seems to be true of any significant change made by the company, but it will likely make more of a bang than any improvement to Siri the company may introduce, then or later.

There should be no question that all of what Lovejoy is saying here is true. If Apple launches an all-new systemwide UI theme for iOS 19, something even half as radical a change as iOS 7’s theme was, it will be the only thing most users notice or opine about. Humans are fundamentally visual creatures and we notice visual changes. And, most humans are resistant to change. A lot of people simply dislike change, even changes for the better. (Honestly, I think that’s what the iOS 18 Photos complaints are about. Complaints about change itself.)

Part of what makes Apple Apple is that the company is (or at least should be) led by people who both have great taste and trust their own instincts. No one’s taste is perfect. Even Steve Jobs pushed through a few clunkers. But if you have great taste and confidence, you’ll do what your gut says is right.

When a company only pushes out changes that avoid controversy, it leads to paralysis. Then stagnation. Earlier today, Cabel Sasser wrote:

One of my strongest early developer memories was being in the “UI Feedback Forum” at WWDC after they introduced Aqua. Think of a live Q&A, but developers giving notes to a team of Apple engineers.

To these veteran Mac coders, the reaction to Aqua was universally negative. People were actively very angry. It’s a waste! It’s ugly! It’s confusing! How could you. It went on and on, and I was surprised because Aqua looked cool and fun to me.

After that WWDC, they never did another Feedback Forum.

It’s true. What a lot of Mac users wanted was a Mac OS X that looked like Mac OS 9.¹ The Aqua look and feel was definitely polarizing. And Apple dialed back its most exuberant details with each subsequent Mac OS X update — less transparency, subtler pinstripes (pinstripes!), etc. But iOS 7 was equally polarizing, and its excesses also got dialed back (or perhaps better, said, dialed back up) with each successive iOS release — a little more depth, some subtle hints of texture.

Either Apple is never going to ship an altogether new UI theme, or they’ll ship one and a large number of people will declare it utter garbage and proof that Apple has completely lost its way. Maybe it will be garbage and proof that Apple has lost its way! Or, maybe it will actually prove to be a great new look that starts a decade-long industry-wide trend that all other companies will soon follow — which is what happened with Aqua’s 3D “lickability”, and happened again with iOS 7’s austere flattening. But either way, you won’t be able to judge it by asking for a show of hands from the general population when it’s unveiled. You either have taste or you don’t, and most people don’t, at least when judging something new and unfamiliar. ★

Which is actually what 1999’s Mac OS X Server 1.0 looked like. It really was the best-looking version of the classic Platinum UI theme Apple ever released. ↩︎

Apple Is Delaying the ‘More Personalized Siri’ Apple Intelligence Features

Friday, 7 March 2025

Here’s a statement I got this morning from Apple spokeswoman Jacqueline Roy, verbatim:

“Siri helps our users find what they need and get things done quickly, and in just the past six months, we’ve made Siri more conversational, introduced new features like type to Siri and product knowledge, and added an integration with ChatGPT. We’ve also been working on a more personalized Siri, giving it more awareness of your personal context, as well as the ability to take action for you within and across your apps. It’s going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year.”

This is a Friday sort of wah-waaah sad-trombone news drop, but it’s good that Apple is getting on top of it, and setting expectations accurately.

Reading between the lines, and based on my PhD-level fluency in Cupertino-ese, what Apple is saying here is that these “more personalized Siri” features are being punted from this year’s OS cycle to next year’s: to iOS 19 and MacOS 16. Apple’s years in this context aren’t calendar years, but Apple’s OS product years. Those years effectively start at WWDC. The problem, from Apple’s perspective, messaging-wise, is that they don’t talk about future products. They haven’t officially acknowledged there is going to be an iOS 19 or MacOS 16. They haven’t even yet announced that there’s going to be a WWDC 2025. But they do want to set expectations accurately, especially for a feature as high profile — from Apple’s own marketing push — as Apple Intelligence. So here we are.

It was already pretty obvious these features weren’t coming in iOS 18.4/MacOS 15.4, because they’re not in the developer betas that are already out. And if these features were coming in iOS 18.5/MacOS 15.5, or even iOS 18.6/MacOS 15.6, Apple wouldn’t have felt the need to issue this “It’s going to take us longer than we thought...” statement today. Delivering the “more personalized Siri” in iOS 18.5 or even 18.6 would have been delivering them on the originally announced schedule — the “coming year” that began with WWDC 2024. That would have been the tail end of that “year”, but within it. But Apple, being limited to the self-imposed opaque verbiage of Cupertino-ese, isn’t going to spell that out. But trust me, you can bank on it: what they’re announcing today is that these features are coming in iOS 19/MacOS 16.

Siri has improved in several useful, but iterative, ways under the Apple Intelligence umbrella in this year’s OS cycle (iOS/iPadOS 18 + MacOS 15 Sequoia), but the big “more personalized Siri” that was shown at WWDC last June is the Siri-related Apple Intelligence stuff that people are, justifiably, most excited about. This is the “App Intents” stuff — the features requiring Siri to have access to and, effectively, understanding of your personal information, in your apps, stored privately on your devices. These are the most helpful-sounding, most practical, most futuristic features of Apple Intelligence. And they’re the sort of features Apple is almost uniquely positioned to offer, alongside only Google, as the provider of a mobile platform and device ecosystem.

From Apple’s own announcement, after the WWDC keynote:

Siri will be able to deliver intelligence that’s tailored to the user and their on-device information. For example, a user can say, “Play that podcast that Jamie recommended,” and Siri will locate and play the episode, without the user having to remember whether it was mentioned in a text or an email. Or they could ask, “When is Mom’s flight landing?” and Siri will find the flight details and cross-reference them with real-time flight tracking to give an arrival time.

“When is Mom’s flight landing?” is easily understood, by we humans, and a really good example of the sort of Siri interaction that would be obviously useful. But you can also pretty easily see how complex it is on Apple’s side. Siri has long been able to do the “Mom” thing. Your spouse, your kids — you’ve long able to tell Siri to call or text those people by nickname like that. But getting your mom’s flight information out of your email (which I think was the premise of this demo in the WWDC keynote — that the flight info was stored in a message in Mail) or from text messages, and then knowing how to check for current flight status information for that flight, is something far beyond the ken of current Siri. Getting the information requires access to your most personal information — your own email and messages. Getting the correct information from your messages requires the common sense of understanding that you’re obviously asking about some sort of current trip your mom is taking, not, say, an email she sent you about a flight back in 2019, or one from 2022, or even one from December last year. The Siri we currently know lacks that sort of common sense. And then lastly, after getting the correct flight from your email, performing this task requires Siri to have the agency to go look up the current status of that particular flight — and get it right every time it provides an answer.

If Siri answers, “Your mom’s flight is arriving at 4:30 this afternoon, and it’s currently showing an on-time arrival,” you really want that information to be correct if you’ve promised to pick her up at the airport. You’ll be annoyed if you drive to the airport and it turns out she’s not actually landing until 7:30. And you’ll be in real trouble if she’s landing at 1:30 and you show up three hours later, having banked on Siri’s answer being accurate.

It’s very unusual, to say the least, for Apple to announce a product delay.¹ The white iPhone 4 was supposed to debut alongside the black one in June 2010, but, bizarrely, was delayed several times (which I had fun with more than once) and wound up not shipping until the end of April the next year, 10 months late. The original AirPods were announced alongside the iPhones 7 in September 2016, and were supposed to ship in October. They wound up being delayed until December. Apple’s statement in October 2016, regarding the AirPods being delayed, as given to Matthew Panzarino at TechCrunch:

“The early response to AirPods has been incredible. We don’t believe in shipping a product before it’s ready, and we need a little more time before AirPods are ready for our customers.”

This is exactly why Apple hews to an almost religious policy of not announcing products before they’re ready. When they do make exceptions — when they lapse, as it were, as cited above (and below) — they often regret it. You can’t break a promise you never made, and you can’t miss a ship date you never announced. Even with talented teams and realistic managers, there’s often a long span between how long you expect a project to take and how long it actually winds up taking. Engineering schedules often slip even with predictable technologies, and the non-deterministic nature of LLMs makes their future inherently unpredictable.

Most of Apple Intelligence has felt like Apple has pushed it out a year ahead of the company’s usual level of baked-ness. I’ve held all year long that if the entire industry — along with Wall Street — weren’t in the midst of a generative-AI/LLM mania, that “Apple Intelligence” wouldn’t have been announced until this year’s WWDC, not last year’s. Today’s announcement is disappointing, but to me utterly unsurprising. The Siri integration with personal data through App Intents has never felt like something that was going to be ready for real customer usage in this year’s OS cycle. And if it’s Apple’s general policy not to ship a product before it’s ready, that applies tenfold for a product involving LLM access to deeply private on-device information. It’s better to ship late, but ready, than to ship something unfinished or unreliable to hit a promised deadline. And with something like next-generation Siri, I’d say that’s downright essential. Apple can take the bad publicity today’s announcement is engendering; they cannot afford to squander the trust of their users. ★

The granddaddy of modern-era Apple product delays, of course, is AirPower — the multi-device (iPhones, Apple Watch, AirPods) charging device that Apple pre-announced in September 2017, alongside the iPhone X and iPhone 8. When announced, Apple said AirPower would ship in “2018”. It wound up never shipping, but the plug wasn’t officially pulled (so to speak) until 29 March 2019, an absurd 18 months after it was announced. ↩︎

ABC Shuts Down FiveThirtyEight, and Pulls the Plug on Its Website

Thursday, 6 March 2025

Oliver Darcy, reporting at Status (paywalled, alas), on the sudden demise of FiveThirtyEight:

On Wednesday morning, shortly after sending an all-staff memo announcing layoffs at ABC News, network president Almin Karamehmedovic joined a virtual meeting with the FiveThirtyEight team. The 15 staffers at the political analysis site owned by ABC News had learned the previous evening — via a Wall Street Journal report — that they were being let go. Now, they were hearing it from Karamehmedovic himself.

But Karamehmedovic, like other managers carrying out mass layoffs at ABC News that day, had little to say about it. I’m told he appeared to be reading from a script as he delivered the devastating news. He offered no explanation for why the Disney-owned network had decided to shutter FiveThirtyEight, which it acquired in 2018 along with data whiz and then-editor Nate Silver. He took no questions. He simply thanked the employees for their work, told them a human resources staffer would follow up, and ended the meeting. The entire affair, I’m told, lasted about 15 minutes. Soon after, employees had their access to ABC News’ systems shut off and the FiveThirtyEight website was pulled offline. Just like that, it was over.

Business is business, and layoffs and closures are never anything but traumatic. But there’s a right way to do things and a wrong way. And one of the wrong ways is having the staff find out they’re all fired by reading it from The Wall Street Journal. This guy Karamehmedovic sounds like a real Grade-A dickhead.

And what is the deal with these companies that just unceremoniously pull the plug on websites when they close them? Why not keep the FiveThirtyEight site up and running — at least for a while, if not in perpetuity? It costs practically nothing to run a server for a static/archived website. I don’t get it. It betrays a profound level of disrespect for the work that the site hosted. It’s ABC’s business if they want to close FiveThirtyEight as an ongoing concern, but the years of work deserves to remain online, both out of respect for the people who made it and for the audience that might still want to refer to it. Disney didn’t burn the movies from now-closed subsidiary studios like Touchstone Pictures.

There’s a vast gulf between the old world of print and the new world of online publishing. The difference is profound, but largely overlooked, because it’s human nature to just adapt to the nature of different forms of media and quickly accept current media forms as “normal”. With print, issues were largely ephemeral to the audience. You’d get a newspaper every day, and magazines every week or month, and you’d throw them out after you’d finished reading them. If you wanted to look back at what the paper had published about something a year, a month, or even just a few days ago, you were out of luck, in terms of being able to just look it up on the spot. Librarians did the yeoman’s work of archiving newspapers and magazines to microfilm, so nothing was truly gone or forgotten, but it was an enormous hassle to access microfilm archives. You pretty much had to know the dates of the issues you wanted to see, and you had to do the “searching” with your eyeballs, page by page. The ability of anyone, today, to just instantly access week-, month-, year-, and even decades-old stories is a breathtaking advantage of web publications. And for the last 15 years or so, we’ve had the ability to do this from little machines we all carry in our pockets everywhere we go. It’s truly remarkable.

But the standard behavior when closing a web publication is to just pull the plug. When the whole company goes under, that’s one thing. But when there’s a parent company, especially a thriving one, there’s no justification for pulling the plug other than spiteful disregard for the work. From the perspective of a company the size of Disney, it would cost veritable pennies to keep FiveThirtyEight’s website around forever. What a disgrace. ★

Monday, 17 March 2025

Sunday, 16 March 2025

A Postscript on the Singular Nature of Mark Gurman’s Reporting

Sunday, 16 March 2025

Saturday, 15 March 2025

Friday, 14 March 2025

Thursday, 13 March 2025

Something Is Rotten in the State of Cupertino

Wednesday, 12 March 2025

Duke Nukem Intelligence

‘So Why the Fuck Doesn’t It Do That?’

Wednesday, 12 March 2025

A New System-Wide UI Look for iOS — Let Alone MacOS, Too — Would Be a Huge Deal

Tuesday, 11 March 2025

Tuesday, 11 March 2025

Monday, 10 March 2025

Saturday, 8 March 2025

Apple Is Delaying the ‘More Personalized Siri’ Apple Intelligence Features

Friday, 7 March 2025

ABC Shuts Down FiveThirtyEight, and Pulls the Plug on Its Website

Thursday, 6 March 2025