WWDC2026 - All the emotions and who do *you* trust?
Apple’s WWDC 2026 keynote just wrapped. I come away having gone from spending most of the keynote thinking “wow, this looks really good” to it leaving me feeling (probably temporarily) disappointed.
To be clear: there are no other OSes I would rather use than MacOS on PC and iOS on phone. I am wholly invested in Apple’s success here.
(Just to note: as I wrote this post, my thinking in part two - the lack of Siri AI in the EU - kind of became the point - I’m sorry for ‘burying the lead’).
The good is the Siri AI stuff around surfacing context and supporting the smart creation of appointments and reminders; the use of AI to build shortcuts; and the improvements overall to the various systems and interface design. Great stuff!
I still think the generative AI stuff is off-brand for Apple (cf. non-consensual use of photos of other people and the general hoovering up of the work of artists, designers, and writers). I’d have rather seen them take a stand against that kind of use of AI, but I guess the market wants what it wants.
No local model love for iPhone 16 Pro models I bought my iPhone 16 Pro Max in 2024 (less than 2 years ago) on the promise that it was the phone for Apple Intelligence. The privacy-first pitch was that the models would run on-device for a lot of things, only seeking Private Cloud Compute (off-device, in Apple’s cloud) when more complex tasks were asked for. Apple never really delivered that for me: the local models that have run for the past 18 months are weak and don’t deliver the promised context-aware abilities (mind you, neither do the cloud models).
Now, the new local models for ‘Siri AI’ reportedly need 12 GB of RAM to run - more than my phone has, so the door is closed on local models providing context awareness on the iPhone 16 Pro Max.
Whether by design or by a hardware ceiling they probably should have seen coming, the result is the same: the 16 Pro phones will never run the version that makes good on their 2024 sales premise. Honestly, this feels really really bad. I’ve even written my first email to Tim Cook!
*EU Digital Markets Act shenanigans * Once again, Apple is doing some (in my view) ‘malicious compliance’ in the EU because of the Digital Markets Act.
Apple’s rough stance is that the DMA forces it to hand any third-party assistant the same deep access to my device that Siri AI has - reading messages, acting across apps - and that this is a security risk.
At the platform level I agree - it would be a new attack surface with deep access to the system. At the user level, I also agree - using another AI that is less privacy-focused than Apple would also be a risk. Let me be fair and quote Apple here (from www.apple.com/newsroom/… accessed 09.06.2026 at 21.22 CET):
“According to EU regulators, the DMA requires Apple to give any AI system nearly unlimited access to a user’s device, as well as the ability to act on that access autonomously without a user’s ongoing visibility and control. That includes the ability to read and send messages, make purchases, access files, and execute actions across any app. Security researchers have already shown that AI systems can be hijacked to steal personal data — like passwords and photos — and to permanently alter files and account settings without a user’s consent. As AI systems gain more capabilities, these risks are quickly increasing in frequency and scope. > Given the serious risks to users, Apple designed a solution called Trusted System Agent — an intermediary that would allow virtual assistants to safely access the same features and capabilities as Siri AI for devices in the EU. Apple also shared a plan to launch Siri AI in the EU while gradually rolling out this new solution over an 18-month period. The European Commission said no. In fact, the European Commission did not agree to any of Apple’s proposals. > Apple will continue working to bring these features to the European Union as safely as possible. However, given the clear dangers to EU users and the regulators’ failure to acknowledge these risks, there is currently no timeline for Siri AI’s availability in the EU on iOS and iPadOS.”
The first question is: do we take Apple at their word here? I would say we shouldn’t entirely. I believe them that they offered options that the EU rejected. But the fix they propose - a ‘Trusted System Agent’ that every other assistant has to route through - still rests on the premise that Apple can be trusted in ways that makers of other assistants can’t be. I’m sure that holds against some rando AI, but it’s almost certainly not true for the likes of Claude or Mistral.
The second question is: what’s really going on here? To which the answer is broadly ‘Apple doesn’t really like competition.’
At the user level, the most liberal position is that I should be allowed to make my own choice to take the risk on a rando AI or not (I wouldn’t - I want Siri AI). Apple’s duty of care to me is real, but it doesn’t overrule my right to make bad choices for myself. Apple can help me not make bad choices, but it shouldn’t be permitted to forbid me from doing so if doing so would unfairly disadvantage competitors who are as trustworthy as Apple. Indeed, my right to choose is precisely what the DMA is seeking (imperfectly) to protect.
Actually, in not wanting to give me a choice, Apple is not giving me the thing I actually want: to use their model! Oh the irony!
At the platform level, if Apple genuinely embraced the liberal position at the user level, they could put in the work to manage the attack surface with proper APIs (not a Trusted System Agent intermediary). Indeed the DMA does allow Apple to take reasonable steps to keep the system secure, so the real argument is over what ‘reasonable’ means, and the EU seems to think Trusted Agent is not reasonable as an approach (and I tend to believe them, because I agree with the principle of the law and assume I would make the same assessment).
My broad take is that Apple is both genuine in its desire for user security and quality of experience and understandably resistant to giving their competitors any advantage (hence the need for regulation here!). So we find ourselves in a position where they try to comply in creative ways that do protect user privacy and security, but do so by co-opting both of those to also optimise for Apple’s own interest - i.e. that avoid embracing a truly competitive level playing field.
The Trusted Agent is actually a great example of a genuine security mechanism that also happens to be an Apple-controlled chokepoint every competitor has to route through, which is why Apple can pitch it in good faith and the EU can reject it in good faith. The prize is both short-term GenAI subscription money from iCloud+ and - perhaps more strategically - the long-term ownership of the agent that sits between us and our phones and learns everything about us.
*All-in-all * Despite these wrinkles, Apple’s OSes - for me - remain far ahead of Microsoft’s and Google’s, and Shortcuts remains a powerful and delightful way to ‘program’ my computers to do what I want them to do. So I’m frustrated that I come away feeling disappointed from an otherwise rather promising keynote. Eventually, the feeling will pass. I will, after all, be able to use Siri AI in the EU eventually (oh please let that be true!), and my iPhone 16 Pro Max will run Siri AI through Private Cloud Compute… it’s just not what I felt I was promised, or what I was hoping for.
*One more thing… * Reflecting on this also makes me realise that much of the prominent coverage of Apple over the next days and weeks is going to be from US-based or US-focused people who really like the company and - even when they seek to be critical - often give it a pass. This often gets dismissed as ‘fanboy-ism’, but I actually think the root of the problem is their extensive insight into Apple and how it works. In essence the best commentators half-think like Apple execs; they deeply understand the company and see the constraints and the rationale behind the decisions it makes, so it becomes a double-edged sword: they’re both more insightful and less critical.
baratheonpilled
prguitarman
jacobvanloon
ellievsbear
turnnoffyourmind