Tech News

MG Siegler returns to the show to talk about the drama surrounding Siri and Apple Intelligence.

Sponsored by:

• WorkOS: The modern identity platform for B2B SaaS — free up to 1 million monthly active users. Check out their latest features from Launch Week.
• BetterHelp: Give online therapy a try at BetterHelp and get on your way to being your best self.
• OpenCase: MagSafe perfected that’s thinner, lighter, and more secure. Save 10% with code TALKSHOW.

How has your week been? My week was ... busy. That includes a new episode of The Talk Show recorded yesterday, dropping in your favorite podcast app soon. Amidst all the writing (and talking) I’ve been doing, I’m also working on filling up open weeks on the sponsorship schedule for Q2.

After a very full February and March, I’ve got a bunch of openings in the next few months — and openings for the next two weeks, starting with this Monday. Update: The coming week just sold, but the next week, starting March 31, remains open.

Weekly sponsorships have been the top source of revenue for Daring Fireball ever since I started selling them back in 2007. They’ve succeeded, I think, because they make everyone happy. They generate good money. There’s only one sponsor per week and the sponsors are always relevant to at least some sizable portion of the DF audience, so you, the reader, are never annoyed and hopefully often intrigued by them. And, from the sponsors’ perspective, they work. My favorite thing about them is how many sponsors return for subsequent weeks after seeing the results.

If you’ve got a product or service you think would be of interest to DF’s audience of people obsessed with high quality and good design, get in touch.

My thanks to WorkOS for sponsoring DF, once again, this last week. This has been WorkOS’s Launch Week, and they’ve got a slew of new features to show. Honestly, though, you should check out their Launch Week page just to look at it — it’s beautiful, fun retro-modern pixel-art goodness. Great typography too. I wish every website looked even half this cool.

New features launched just this week include:

• WorkOS Connect — “Sign in with [Your App]”
• WorkOS Vault — Encryption Key Management (EKM) and Bring-Your-Own-Key (BYOK)
• AuthKit Integrations — Native support for several new identity providers including LinkedIn, Slack, GitLab, BitBucket, Intuit, and more.

Ookla, the company behind the Speedtest download/upload bandwidth testing app:

Although it’s early in the adoption curve for the iPhone 16e, we analyzed the performance of the new device from March 1st through March 12th, and compared it to the performance of iPhone 16, which has a similar design and the same 6.1” screen. Both devices run on the same Apple-designed A18 SoC.

When we compare Speedtest Intelligence data from the top 90th percentile (those with the highest performance experience) of iPhone 16e and iPhone 16 users from all three of the top U.S. operators, we see the iPhone 16 performing better in download speeds. However, at the opposite end, with the 10th percentile of users (those who experience the lowest performance) we see the iPhone 16e performing better than the iPhone 16.

There are some differences, but overall the 16e’s cellular performance seems great for the frequencies it supports. And given the efficiency claims from Apple, it might be the better overall modem. (I also think the frequencies it doesn’t support don’t really matter all that much in real-world practice. If you know that you really make use of the crazy-high speeds of mmWave from Verizon, then you know the C1 modem is not for you.)

Emma Roth, The Verge:

TechCrunch has a new owner, again. Yahoo has sold the tech news site to the private equity firm Regent for an undisclosed sum, according to an announcement on Friday.

Regent is the same company that snapped up Foundry, the firm behind outlets like PCWorld, Macworld, and TechAdvisor on Thursday. Founded in 2005, TechCrunch has experienced many shakeups in ownership after AOL acquired the site in 2010.

A lot of shakeups in a lot of media companies’ ownership lately. Steady as she goes here at The Daring Fireball Company, a subsidiary of Fedora World Media Industries.

Matthew Belloni has a very good take on Apple TV+ at Puck (that’s a gift link that should get you through their paywall — but which requires you creating a free account, sorry):

All of which fed into the self-centered fears of my lunch date. What, if anything, does the current state of Apple mean for its entertainment business? After all, more than five years into the Apple TV+ experiment, it’s never been entirely clear what C.E.O. Tim Cook and services chief Eddy Cue are up to in Hollywood. Certainly not making money, at least not in the traditional sense. The Information reported today that Apple lost $1 billion on Apple TV+ last year, following a Bloomberg report that more than $20 billion has been shoveled into making original shows and movies since 2019. That’s not nothing, even for a company worth $3 trillion.

The “loss” number is a bit misleading, of course, considering Apple has always said that a key goal is to leverage Leo DiCaprio and Reese Witherspoon to thicken its brand halo and the device “ecosystem,” ultimately boosting its other businesses. But still… for all its billions, Apple TV+ has accumulated only about 45 million subscribers worldwide, according to today’s Information report and other estimates.

That’s far less than Disney+, Max, and Paramount+, all of which launched around the same time. Those rival services are attached to legacy studios with rich libraries, but they’re not attached to a company with $65 billion in cash on hand and a device in the pockets of 1 billion people that also delivers bundle-friendly music, news, and games. Apple declined to confirm or comment on any numbers, but a source there suggested the subscriber number is higher than 45 million and that the global nature of the sub base is being undercounted by U.S.-oriented research firms. Maybe. The company reveals zero performance data beyond B.S. “biggest weekend ever!” press releases that the trades accept without skepticism and producers like Ben Stiller and David Ellison post with “blessed” emojis on their social media. No one outside the company really knows how the Apple TV+ business is performing.

One interesting nugget is this chart, which suggests that subscriptions to TV+ have boomed since Apple and Amazon worked out a deal to sell TV+ subscriptions through Amazon Channels in Prime Video at the end of last year. That deal has, seemingly, moved the needle. Another interesting nugget is that TV+ seems to suffer from a higher churn rate than other streaming services. Said Belloni’s Puck colleague Julia Alexander, “Fewer than 35 percent of all subscribers keep the service for longer than six months.”

That’s kind of crazy. I’d think TV+ would have less churn, not more, than the industry average — that the Apple TV+ audience is small but loyal. Perhaps this is the unsurprising side effect of Apple giving away 3-month trials when you purchase new devices. But I also truly wonder if TV+ subscriptions are the hardest for industry groups to measure, because so many people who do subscribe watch through tvOS (or, on their phones, on iOS) where everything is private. Belloni hints at this, and says little birdies at Apple told him the TV+ subscriber base is larger than they’re getting credit for.

And how do you count Apple One subscribers toward TV+’s subscriber base? My vague theory about Cue and Cook’s thinking about getting into this business has been about making it one leg among several on the stool of reasons to subscribe to Apple One. That Apple will take subscribers who are only subscribed to TV+, or only subscribed to TV+ and Apple Music, but what they really want is to get people to subscribe to Apple One, which, because it includes iCloud storage, almost certainly has very little churn.

Belloni closes thus:

Apple wouldn’t be the first tech powerhouse to dabble in professionally produced content only to retreat. [...] Neither Cook nor Cue has suggested anything like that, and Apple, in just over five years, has become a reliable partner and a high-quality buyer for Hollywood shows and movies. In some ways, it’s remarkable how fast Apple TV+ became part of the entertainment community. Whether that lasts is the question.

Here’s where I will point out that Apple isn’t like other tech companies. Apple isn’t a move fast and break things company. They’re a measure twice, cut once company. When they commit to something, they tend to stay committed. And they’re very, very good at playing long games that require patience, especially when entering new markets. Look at Apple Pay. 10 years ago, it was widely panned as a flop after a slow first year. Now it’s everywhere.

Jill Goldsmith, Deadline:

Apple is losing more than $1 billion a year on streamer Apple TV+, according to a report in the Information that cited two people familiar with the matter. The tech giant has spent over $5 billion a year on content since launching Apple TV+ in 2019 but trimmed that by about $500 million last year, the report said.

The headline on Wayne Ma’s report at The Information set the framework: “Apple Streaming Losses Top $1 Billion a Year” — the story got picked up widely, and almost everyone who did framed it in terms of losing or a loss. But is it a loss when Apple expected the business to be unprofitable for a decade or more? From Scharon Harding’s paraphrasing at Ars Technica of Ma’s paywalled report:

Apple TV+ being Apple’s only service not turning a profit isn’t good, but it’s also expected. Like other streaming services, Apple TV+ wasn’t expected to be profitable until years after its launch. An Apple TV+ employee that The Information said reviewed the streaming service’s business plan said Apple TV+ is expected to lose $15 billion to $20 billion during its first 10 years.

For comparison, Disney’s direct-to-consumer streaming business had operating losses of $11.4 billion between the launch of Disney+ in fall 2020 and April 2024. Disney’s streaming business became profitable for the first time in its fiscal quarter ending on June 29, 2024.

The above two paragraphs of essential context are buried 13 paragraphs down. If Apple expected TV+ to operate in the red, to the tune of $15–20 billion over its first decade, and halfway through that decade (TV+ debuted in November 2019) it operated in the red to the tune of $1 billion for the year — doesn’t that mean costs are exactly in line with their expectations?

The insinuation here is that Apple’s pissing this money away and doesn’t know what they’re doing. Maybe they are! But if so it was exactly Eddy Cue and Tim Cook’s strategy to piss this money away. If Apple had expected TV+ to be profitable or break-even in 2024, then a $1 billion operating loss would be a story. But as it stands it’s just a cost. How much did Apple “lose” on electricity bills last year?

Juli Clover, writing for MacRumors last week:

With new iOS software updates, Apple has been automatically turning Apple Intelligence on again even for users who have disabled it, a decision that has become increasingly frustrating for those that don’t want to use Apple Intelligence .

After installing iOS 18.3.2, iPhone users have noticed that Apple Intelligence is automatically turned on, regardless of whether it was turned off prior to the update being installed. There is an Apple Intelligence splash screen that comes up after updating, and there is no option other than tapping “Continue,” which turns on Apple Intelligence .

If you’ve updated to iOS 18.3.2 and do not want Apple Intelligence enabled, you will need to go the Settings app, tap on Apple Intelligence, and then toggle it off. When Apple Intelligence is enabled, it consumes up to 7GB of storage space for local AI models, which is an inconvenience when storage space is limited.

I’d been seeing complaints about this, including from some friends who are developers and/or had previously worked on iOS as engineers at Apple. A bunch of regular DF readers have written to complain about it too. I wouldn’t call it a deluge, but I’ve gotten an unusual number of complaints about this. (And at CNet, Jeff Carlson reports the same thing happening with MacOS 15.3.2.)

I hadn’t experienced it personally because I have Apple Intelligence enabled on my iPhone. But my year-old iPhone 15 Pro was still running iOS 18.2. So I disabled Apple Intelligence on that phone, then updated it to 18.3.2. When it finished, Apple Intelligence was re-enabled. I also tried this on my iPhone 16e review unit, which was still running iOS 18.3.1 (albeit a version of 18.3.1 with a unique build number for the 16e). I turned Apple Intelligence off, upgraded to 18.3.2, and on that iPhone, Apple Intelligence remained off after the software upgrade completed.

So I don’t know if this is a bug that only affects some iPhones, or a deliberate growth hacking decision from Apple to keep turning this back on for people who have explicitly turned it off. But it’s definitely happening.

And while the 7 GB of storage space required for the model is a legitimate technical reason to turn it off, I think (judging from my email from DF readers) the main reason people disable Apple Intelligence is that they don’t like it, don’t trust it, and to some degree object to it. It could take up no additional storage space at all and they’d still want it disabled on their devices, and they are fucking angry that Apple’s own software updates keep turning it back on. Put aside the quality or utility of Apple Intelligence as it stands today, and there are people who object to the whole thing on principle or, I don’t know, just vibes alone. Feelings are strong about this. Turning it back on automatically, after a user had turned it off manually, leads those users to correctly distrust Apple Intelligence specifically and Apple in general.

If it’s a bug, it’s a bug that makes Apple look like a bunch of gross shysters. If it’s not a bug, it means Apple is a bunch of gross shysters. I’d wager on bug — especially after seeing it not happen on my 16e review unit. I’m thinking it’s something where it’s supposed to be enabled by default, once, for people who’ve never explicitly turned Apple Intelligence on or off previously, but that for some devices where it has been turned off explicitly, somehow the software update is mistaking it for the setting never having been touched. Apple needs to get it together on this one.

Ina Fried, reporting for Axios:

The suit, filed Wednesday in U.S. District Court in San Jose, seeks class action status and unspecified financial damages on behalf of those who purchased Apple Intelligence-capable iPhones and other devices.

“Apple’s advertisements saturated the internet, television, and other airwaves to cultivate a clear and reasonable consumer expectation that these transformative features would be available upon the iPhone’s release,” the suit reads. “This drove unprecedented excitement in the market, even for Apple, as the company knew it would, and as part of Apple’s ongoing effort to convince consumers to upgrade at a premium price and to distinguish itself from competitors deemed to be winning the AI-arms race.”

Most of these class action lawsuits are bullshit, but it’s hard to argue with the basic premise of this one.

This is beautiful and crazy, and no, I’m not going to buy one, but damn I’m tempted and I’d sure like to try one. I’m glad it exists.

Mark Gurman, with a blockbuster scoop for Bloomberg:

Apple Inc. is undergoing a rare shake-up of its executive ranks, aiming to get its artificial intelligence efforts back on track after months of delays and stumbles, according to people familiar with the situation.

Chief Executive Officer Tim Cook has lost confidence in the ability of AI head John Giannandrea to execute on product development, so he’s moving over another top executive to help: Vision Pro creator Mike Rockwell. In a new role, Rockwell will be in charge of the Siri virtual assistant, according to the people, who asked not to be identified because the moves haven’t been announced.

Rockwell will report to software chief Craig Federighi, removing Siri completely from Giannandrea’s command. Apple is poised to announce the changes to employees this week. The iPhone maker’s senior leaders — a group known as the Top 100 — just met at a secretive, annual offsite gathering to discuss the future of the company. Its AI efforts were a key talking point at the summit, Bloomberg News has reported. [...]

My quick take on this is that it’s a turf battle that Craig Federighi just won. It’s not just putting a new executive in charge of Siri, it’s moving Siri under Federighi’s group.

How Gurman got this scoop before Apple had announced the changes — even internally — is rather unbelievable. It’s not “Bloomberg” that got this scoop. It’s Mark Gurman. And trust me, Apple PR did not leak this to him deliberately. I’m sure they’re now accelerating an announcement, at least internally, framing it on their own terms. I can only guess that Gurman hinted at his sourcing in the passage above: Tim Cook must have announced these changes at the Top 100 retreat this week, and at least two of those attendees leaked the news to Gurman. Unprecedented.

Also:

Rockwell is currently the vice president in charge of the Vision Products Group, or VPG, the division that developed Apple’s headset. As part of the changes, he’ll be leaving that team and handing the reins to Paul Meade, an executive who has run hardware engineering for the Vision Pro under Rockwell.

I don’t find it surprising at all that Rockwell was given this task.

Giannandrea will remain at the company, even with Rockwell taking over Siri. An abrupt departure would signal publicly that the AI efforts have been tumultuous — something Apple is reluctant to acknowledge. Giannandrea’s other responsibilities include oversight of research, testing and technologies related to AI. The company also has a team reporting to Giannandrea investigating robotics.

This I find a little surprising. But maybe I shouldn’t. I don’t buy Gurman’s argument that dismissing Giannandrea would “signal publicly that the AI efforts have been tumultuous”. Apple already signaled that publicly when they announced that all of the ambitious features for Siri and Apple Intelligence that were promised for this year’s OS cycle would be postponed until next year’s OS cycle. That’s public tumult. But I mean, you can see for yourself that Apple’s AI efforts have been “tumultuous” by asking Siri on your iPhone, right now, what month it is.

What Apple needs to signal is that they don’t expect to deliver a significantly better Siri without making significant changes to the team behind Siri.

But maybe the answer is as simple as that Giannandrea is good at leading and managing teams doing advanced research that is abstracted from product. So move the products out of his division and into Federighi’s, and put someone who knows how to ship directly in charge of Siri. Leave Giannandrea in charge of a division focused on research and technology. Attention has moved on from “machine learning” to LLMs, but Apple’s machine learning game has gotten very good.

Here’s an update I just appended to my post yesterday, after linking to Gus Mueller’s suggestion that Apple open up a semantic index to third-party AI apps:

HealthKit already works a lot like what Mueller is suggesting here (for, say, “SemanticKit”). With explicit user permission — that can be revoked at any time — third party apps can both read from and write to your Health data. Apple does a lot of that itself, both through Apple Watch and from the various activity-related things an iPhone can track, but third-party apps and devices are welcome participants, in a private, easily-understood way.

Nobody is suggesting Apple should give up on AI. Quite the opposite. They really need to go from being a joke to being good at it, fast. But there’s no reason at all they should build out a strategy that relies on Apple doing all of it themselves, and Apple users relying solely on Apple’s own AI. Do it like Health — a model that has proven to be:

• profitable (for Apple itself, selling devices like Watches);
• popular (with users, who actually use it, understand it, and like it);
• private;
• and open to third-party developers, device makers, and medical service providers.

(Thanks to Bill Welense for the suggestion.)

Last March, when Apple introduced the then new M3 MacBook Airs, they moved the base model 13-inch M2 MacBook Air into the magic $999 spot in their own lineup, replacing the M1 MacBook Air. But mid-March it was announced that Walmart would begin selling the M1 MacBook Air — in one tech-spec configuration (8 GB RAM, 256 SSD), but three colors (gold, silver, space gray) for just $700.

This year Apple replaced the entire lineup of MacBook Airs that it sells itself with M4-based models, including the $999 starting-price model. Online, Walmart sells a handful of MacBook models now, at, per Walmart’s brand, slightly lower prices than Apple itself. But the one and only MacBook they seem to stock in their retail stores is the classic wedge-shaped M1 MacBook Air — now down to $650.

It’s over four years old now, and yes, 8 GB RAM and 256 GB of storage are meager, but it’s almost certainly the best new laptop you can buy for that price. Assuming Apple thinks this partnership is a success, eventually they’ll have to replace this with a more recent MacBook Air. But I suspect the main reason it’s still the M1 Air (and hasn’t been replaced by, say, the M2 Air) is not about the specs or performance, per se, but rather simply how it looks. It looks like an older MacBook. Walmart might not get an updated MacBook with a more-recent-than-M1 chip until Apple refreshes the industrial design on its current MacBook Airs.

Whole Reddit thread examining this simple question: “What month is it?” and Siri’s “I’m sorry, I don’t understand” response (which I just reproduced on my iPhone 16 Pro running iOS 18.4b4). One guy changed the question to “What month is it currently?” and got the answer “It is 2025.”

Update: Ask Siri (with Apple Intelligence™) “ChatGPT, what month is it?” and, though you’ll have to wait a few extra seconds, you’ll get the right answer each time. Perhaps the current month is “broad world knowledge” and Siri shouldn’t even attempt to answer such a complex question on its own?

News from Apple that I let slip by a few weeks ago, but that seems apt again today:

Apple Intelligence, the personal intelligence system that delivers helpful and relevant intelligence, will soon be available in more languages, including French, German, Italian, Portuguese (Brazil), Spanish, Japanese, Korean, and Chinese (simplified) — as well as localized English for Singapore and India.

These new languages will be accessible in nearly all regions around the world with the release of iOS 18.4, iPadOS 18.4, and macOS Sequoia 15.4 in April, and developers can start to test these releases today.

With the upcoming software updates, iPhone and iPad users in the EU will have access to Apple Intelligence features for the first time, and Apple Intelligence will expand to a new platform in U.S. English with Apple Vision Pro — helping users communicate, collaborate, and express themselves in entirely new ways.

Given that Apple Intelligence isn’t exactly setting the world on fire, I think in the grand scheme of things, it’ll wind up being filed away under “Oh yeah, remember that?” that the EU got it 4-5 months after it debuted. (Clean Up in Photos is often great, and I genuinely enjoy notification summaries and miss them now that they’re disabled for news apps; the rest I don’t use, and the most ambitious aspects of Apple Intelligence are (you may have heard) delayed for everyone, not just the EU.)

Apple was concerned that the EU’s hardline interpretation of the DMA was such that the European Commission considered it a violation of the DMA that Apple Intelligence wasn’t an interchangeable component. Like the way the EC forced Apple to open up iOS to alternative app marketplaces — there was uncertainty whether they’d demand the same for system-integrated AI. And if that’s what the EC had demanded, they simply wouldn’t have gotten system-integrated AI for years. But I’m not sure how to square up today’s decisions — requiring Apple to enable third-party alternatives to system-level features like AirPlay and AirDrop — with an interpretation that the EU will be fine with Apple Intelligence only offering Apple’s own AI (along with Apple’s approved partners, like OpenAI).

I think the regime change at the European Commission has changed things to some degree, but quietly. Former competition chief Margrethe Vestager was a firebrand. Back in June last year, after Apple had announced that Apple Intelligence would be delayed indefinitely in the EU for iOS, she made clear that she thought it was anti-competitive:

“I find that very interesting that they say we will now deploy AI where we’re not obliged to enable competition. I think that is that is the most sort of stunning open declaration that they know 100% that this is another way of disabling competition where they have a stronghold already.”

But Vestager is gone, and until today we hadn’t heard a whit about DMA compliance from her successor, Teresa Ribera. In September, when the proceedings that resulted in today’s decisions opened, I wrote:

Also worth noting: Margrethe Vestager is on her way out, about to be replaced by Spanish socialist Teresa Ribera, a career climate expert (which, possibly, might give her an affinity for Apple, far and away the most climate-friendly large tech company) with no experience in competition law. To me that makes Ribera an odd choice for the competition chief job, but apparently that makes sense in the EU. It remains unclear to me whether Ribera supports Vestager’s crusade against the DMA’s designated “gatekeepers”. If she doesn’t, is this all for naught?

Until today, that remained an open question. Now it appears the Commission’s crusading course is unchanged — it’s just no longer accompanied by inflammatory commentary from the commissioners in charge.

The European Commission, today:

Today, the European Commission adopted two decisions under the Digital Markets Act (DMA) specifying the measures that Apple has to take to comply with certain aspects of its interoperability obligation. [...]

The first set of measures concerns nine iOS connectivity features, predominantly used for connected devices such as smartwatches, headphones or TVs. The measures will grant device manufacturers and app developers improved access to iPhone features that interact with such devices (e.g. displaying notifications on smartwatches), faster data transfers (e.g. peer-to-peer Wi-Fi connections, and near-field communication) and easier device set-up (e.g. pairing).

Benjamin Mayo, reporting for 9to5Mac:

In a statement to 9to5Mac, Apple firmly rebuked the EU decision announced today about specific interoperability requirements the company must implement over the coming months.

Apple said “Today’s decisions wrap us in red tape, slowing down Apple’s ability to innovate for users in Europe and forcing us to give away our new features for free to companies who don’t have to play by the same rules. It’s bad for our products and for our European users. We will continue to work with the European Commission to help them understand our concerns on behalf of our users”.

In regards to customer privacy, Apple is especially concerned with the requirements surrounding opening up access to the iOS notification system. The company indicated these measures would allow companies to suck up all user notifications in an unencrypted form to their servers, sidestepping all privacy protections Apple typically enforces.

My interpretation of the adopted decision is that the EU is requiring Apple to treat iOS like a PC operating system, like MacOS or Windows, where users can install third-party software that runs, unfettered, in the background.

Apple’s statement makes clear their staunch opposition to these decisions. But at least at a superficial level, the European Commission’s tenor has changed. The quotes from the Commission executives (Teresa Ribera, who replaced firebrand Margrethe Vestager as competition chief, and Henna Virkkunen) are anodyne. Nothing of the vituperativeness of the quotes from Vestager and Thierry Breton in years past. But the decisions themselves make clear that the EU isn’t backing down from its general position of seeing itself as the rightful decision-maker for how iOS should function and be engineered, and that Apple’s core competitive asset — making devices that work better together than those from other companies — isn’t legal under the DMA.

Sebastiaan de With:

You can speculate what the ‘e’ in ‘16e’ stands for, but in my head it stands for ‘essential’. Some things that I consider particularly essential to the iPhone are all there: fantastic build quality, an OLED screen, iOS and all its apps, and Face ID. It even has satellite connectivity. Some other things I also consider essential are not here: MagSafe is very missed, for instance, but also multiple cameras. It would be reasonable to look at Apple’s Camera app, then, and see what comprises the ‘essential’ iPhone camera experience according to Apple.

Alex Cheema is the founder of EXO Labs, an AI company focused on “AI you can trust with your data” by making systems that run locally, on computers you own and control. Apple provided him with two M3 Ultra Mac Studios, each maxed out with 512 GB of unified memory. Within a day, he had them linked together by Thunderbolt 5 and had the full DeepSeek R1 model running on his desk.

Sure, that’s over $20,000 of computing hardware. But to my knowledge there is no other way in the world to run the full DeepSeek R1 model for even close to $20,000, let alone doing it on your desk rather than a data center. It’s an exclusive advantage, made possible by Apple Silicon’s general performance and the breakthrough of Apple’s unified memory architecture, which lets the GPU cores access the same RAM as the CPU cores.

Apple has tremendous technical advantages to offer in AI. But they’re marketing Genmojis of hot dogs carrying briefcases.

Gus Mueller:

A week or so ago I was grousing to some friends that Apple needs to open up things on the Mac so other LLMs can step in where Siri is failing. In theory we (developers) could do this today, but I would love to see a blessed system where Apple provided APIs to other LLM providers.

Are there security concerns? Yes, of course there are, there always will be. But I would like the choice.

The crux of the issue in my mind is this: Apple has a lot of good ideas, but they don’t have a monopoly on them. I would like some other folks to come in and try their ideas out. I would like things to advance at the pace of the industry, and not Apple’s. Maybe with a blessed system in place, Apple could watch and see how people use LLMs and other generative models (instead of giving us Genmoji that look like something Fisher-Price would make). And maybe open up the existing Apple-only models to developers. There are locally installed image processing models that I would love to take advantage of in my apps.

The analogy I used, talking with Jason Snell during my guest stint on Upgrade last week, was to the heyday of desktop publishing. The Mac was the platform for graphic design because it was the best platform for using design apps. Fonts worked better and looked better on the Mac. Printing worked better from Macs. Peripherals worked better. The apps themselves looked better on the Mac than they did on Windows. The Mac had taste and designers (hopefully) have taste. Graphic designers could understand how their machines worked, and maintain them themselves, in a way they couldn’t with PCs.

But Apple didn’t make any of the actual apps. Companies like Adobe and Macromedia and Aldus did. Independent small developers made niche extensions for use inside apps like Photoshop, FreeHand, and QuarkXPress. When a new app came along like InDesign — which quickly ate Quark’s lunch — the Mac remained the dominant platform to use.

Making a great platform where other developers can innovate is one of Apple’s core strengths. Apple got even better at it once Mac OS X hit its stride in the 2000s — the Cocoa APIs really did empower outside developers to make world-class apps providing experiences that couldn’t be matched on other platforms like Windows or Linux. Then it happened again, with a much bigger audience, with iOS. What desktop publishing was to the Mac in the 1990s, social media was to the iPhone in the 2010s. Apple didn’t make the apps — they made the best platform to use those apps.

Apple should be laser focused on doing this for AI now. Where I quibble with Mueller is that I don’t want Apple to get out of the way. I want Apple to pave the roads to create the way. Apple doesn’t have to make the cars (literally) — just pave the best roads. Make the Mac the best platform for outside developers to create innovative AI systems and experiences. Make iOS the best consumer device to use AI apps from any outside developer. Work on APIs and frameworks for the AI age. No company has ever been better than Apple at designing and delivering those sort of APIs. Lean into that. It’s as useful, relevant, and profitable an institutional strength (and set of values) today as ever.

In a follow-up post, Mueller shows he’s thinking like I’m thinking:

But off the top of my head, here’s one idea that I think could really help and reap benefits for both Apple and developers.

Build a semantic index (SI), and allow apps to access it via permissions given similar to what we do for Address Book or Photos.

Maybe even make the permissions to the SI a bit more fine-grained than you normally would for other personal databases. Historical GPS locations? Scraping contents of the screen over time? Indexed contents of document folder(s)? Make these options for what goes into the SI.

And of course, the same would be true for building the SI. As a user, I’d love to be able to say “sure, capture what’s on the screen and scrape the text out of that, but nope - you better not track where I’ve been over time”.

HealthKit already works a lot like what Mueller is suggesting here (for, say, “SemanticKit”). With explicit user permission — that can be revoked at any time — third party apps can both read from and write to your Health data. Apple does a lot of that itself, both through Apple Watch and from the various activity-related things an iPhone can track, but third-party apps and devices are welcome participants, in a private, easily-understood way.

Scharon Harding, writing for Ars Technica:

Reports of Roku customers seeing video ads automatically play before they could view the OS’ home screen started appearing online this week. A Reddit user, for example, posted yesterday: “I just turned on my Roku and got an ... ad for a movie, before I got to the regular Roku home screen.” Multiple apparent users reported seeing an ad for the movie Moana 2. The ads have a close option, but some users appear to have not seen it.

When reached for comment, a Roku spokesperson shared a company statement that confirms that the autoplaying ads are expected behavior but not a permanent part of Roku OS currently. Instead, Roku claimed, it was just trying the ad capability out. [...]

“Our recent test is just the latest example, as we explore new ways to showcase brands and programming while still providing a delightful and simple user experience.”

What I’d find delightful and simple is disconnecting my Roku box and throwing it out the window.