Siri AI
apple.com
apple.com
I really feel like there's a fascinating valley of death between simple things that actually work and things of real value that are actually still beyond the horizon. They either aren't reliable enough, aren't accessible to the tech, or exceed the sophistication of our existing trust models. For example, I'm planning a trip. Booking a multiday holiday - there's a real beast that is time consuming, complex and painful. I test out the AI tools. They fail. Hard. Hallucinations all over the place, false confidence, inability to act, inability of me to trust their actions.
It's just nowhere near practical utility yet. Not "nearly there" but "not nearly half way there". I got the top tier of Gemini AI. Can it rent me a car? "As an AI I can absolutely guide you through the process of renting the car, but I can't physically access the web site or type in the details for you".
When we talk about “the market”, the customer base, remember it’s a market that typically doesn’t know how to or care to even install an adblocker.
Because the problem now took a whole afternoon to solved and sapped your creative energy instead.
eg: if my booking is wrong, they will cover the cost and compensate me. It would sort of just come down to buying premium travel insurance for everyone that uses it. And insurance for anything else they do. It has to be one of two things - they either believe the risks are worth it (so then there should be a financial model that can absorb the cost of insurance to do it), or in fact, the risks are too great. At some point, if they keep offering the tech on a "use at your own risk" basis, they are implicitly communicating that they themselves think the risks are too great - so YOU shouldn't trust it either.
That would be nice, but it's the wrong angle. The reason people like real secretaries is not because somebody is compensated when things go wrong. It's because things don't go wrong. I don't use this thing if I need to fear things go wrong, even if I'd be compensated.
Maybe it would provide the right incentives for the companies though.
Though a compensation that high sounds like it would invite fraud, where the customer would be glad to have something go "wrong" and get a fat check. Not sure if that's a solvable problem.
This whole venue of technology is an exercise in ivory tower construction completely disconnected from ordinary people.
They never asked for one becase they never imagined being able to afford one.
The amount of administration organizing a normal household takes I suspect most would be glad to leave to someon/something they trust and that can be held accountable.
Today that someone needs to be a person (imo). But who knows, a startup may be plotting accountable digital assistans as we speak.
That quote has an unexpressed precondition to the effect of "In order for an organisation to be objectively well run..." or "In order for an organisation to equitably benefit all stakeholders at all levels..." etc
But that's absolutely not what he's describing. He wants not to think about it, that's exactly the opposite of a disorder.
Anyway, an LLM assistant is also exactly the worst technology to use there, on every dimension.
I have zero use for AI. What's it going to do? Read the 3 emails about bills I get?
Certainly someone out there has a use for this functionality, but when you say "household admin tasks" the last thing I think about are /digital/ admin tasks.
They don't even seem to get the basics right, why would I want another layer on top?
- don't travel frequently,
- don't have so many complex inquiries that require someone to research,
- don't have super complicated taxes to file,
- don't go eating out in fancy restaurants that require special skills to get reservations in,
- don't have so many meetings to attend,
- don't receive hundreds of emails per day,
- don't work on multiple projects at the same time,
- don't organise festivities and social gatherings all the time.
The one thing that would be useful - a kind of "chief of staff" that monitors your entire digital life and prioritises your every next step - is the antithesis to Siri and the like, which are merely reactive to your requests, not proactive in figuring out what needs your attention next. Let alone that that would be a total privacy nightmare, and a prime candidate for mass manipulation at scale.
Like you said, the non-digital things are where people need assistance most. Fold my laundry, clean my house, clean my car, make me dinner. At an affordable cost. I don't need you to book my trip to an all inclusive resort that I go on once a year at most.
We're at this place where AI/LLMs is truly incredible technology, but the futuristic vision of robot assistants doing things for you at an attainable cost isn't there yet - so a lot of companies/startups are trying to force feed purely digital consumer AI products (assistants/agents) that no one wants.
https://www.normaltech.ai/p/new-paper-towards-a-science-of-a...
They failed most of the time. Simple things like finding the right password for Gmail sometimes was out of reach. Anti bot techniques sometimes stopped it.
Impressively, sometimes they'd successfully write hugely complex bash or python scripts to do tasks on web pages they hadn't managed to do with the browser automation.
Huh? Maybe my uni was an outlier / it is a UK thing, but there were ~99 female and 1 male psych students in my year. This was not considered unusual.
So I end up pulling out the trusty old Garmin gpsmap with cycle/hiking maps, that survived drops from 1.5 meters at 30 km/h as I was gliding of a mountain with my bike.
But I generally prefer to use a Garmin GPS or watch. They work for days without charging (the older models even work with two AA batteries), very robust (e.g. their gpsr survives drops), work well offline, and transflective displays work better in direct sunlight.
For planned routes, I make then in NodeMapp or some other focused application and send the GPX overs to a gpsmap unit or Fenix watch. Many national parks, etc. also have great GPX files for recommended hiking/cycling routes.
It used good models and did a lot of searching, including searches in other languages. It got nothing right, riddled with fake places and times. It also found some weird and unique places I never would have considered.
I had a blast, brought me back to traveling pre-internet, requiring a level of spontaneity I had forgotten we used to depend on. 100% recommend it.
I am now reminded of a short trip with less tech savy folks, where I also on the trip noticed that the plan was a bit .. not working. And the person organizing it complaining to the bus driver, why they were not going what the internet told him, they were going. The internet being ChatGPT.
I told it my preferences and of the group members, where we arrived and departed, at what times. I gave it my itinerary and then asked it to plan two new itineraries and also suggest a location to book a hotel that was convenient for the early flight on the last day.
I went away for 20 mins and gave me a 20 page document with a good summary and decent options. I did choose some of the activities it suggested.
I did this 10 months ago. It’s probably better now.
But Gemini has access to google maps, so it can estimate travel times, and know which lunch places are near which sites and which hotels have good reviews. So if you want AI to work for travel panning you need to ground it in good data.
As you guessed, there's a ton of info in the training data on this topic, but there's some value in being able to see it on one place with different options.
There are a lot of extremely legitimate concerns, like the environmental impact and so on.
But I just laugh when they point out that LLMs are merely clever regurgitators of their previous inputs… as if this isn’t how we as humans operate nearly all of the time. People realllllllllly want to think they’re special snowflakes.
Ask a human to plan a trip:
They do research, Pick destinations led by their own experience/likes/dislikes Compare to other guides Plan itineraries so they can get there Check and share
Ask an LLM to plan a trip:
It takes the prompt and continues it based on weights in the training data. If there is no data it picks the most likely thing (maybe made up). If there is it’ll mostly add things from that data. Maybe it’ll make tool calls and pull in data that way too but you can’t actually trust all the details.
These two processes are so different, it’s important to understand how they work, which is nothing like a human.
But I'd also want to point out that the way you're characterizing an LLM planning a trip doesn't have any structure to it, which indicates that in your scenario you're not using any kind of harness. I've been amazed at how capable even 30 billion parameter models are when I put them inside of a harness that provides structure and task management. If you consider that scenario, especially with the ability to search the web and use skills, suddenly the LLM looks a lot more like what the human process looks like.
Where humans and (current) LLMs differ the most is their failure mode. A human friend could be bad at planning trips, but that's kinda predictable, we're used to it, we know how to catch that Exception. LLMs on the other hand still have failure modes that come across as really wacky, like, what are they smoking in Mountain View?
Which might actually serve as better evidence of different internal workings at a deeper level, than just parroting well-known superficial features of stochastic whatevertheysay.
I don't trust LLMs for this application lol.
You presented an LLM with an obviously bonkers goal, the LLM told you it was a bad idea at multiple steps, and this is somehow... a shortcoming of the LLM?!?
You said it yourself: you needed to "bully" the LLM into even producing this plan.
Please, tell me what it should have done instead. Be very specific!
A reasonable travel agent would have fired me as a customer. The LLM failed to do so.
It should have flatly refused.
I think the LLM should advise you of risk and lack of feasability but should otherwise answer the question, unless you're trying to do something plainly destructive to others e.g. weaponizing anthrax or something.
A reasonable travel agent would have fired me as a customer.
unless you can somehow make it have judgment
I think this is a great example of the unrealistic expectations people have for LLMs. No sane and sensible person would treat any single source of knowledge as infallible, for any consequential decision.
(Certainly, of course, you don't have to look very far for examples of idiots being overly trustful of LLMs, or Google, or GPS, or Wikipedia, or whatever. It certainly does happen and yes, I've heard all these arguments before about other technologies besides LLM. Replace "LLM" in your post with any of those other terms, and I promise you somebody made literally the exact same argument in 2003 or 2009 or 2014 or whatever)
Any reasonable person would consult a second doctor, or at least other sources of knowledge, after the doctor advises them of some irreversible course of action. Because we don't even expect highly trained and intelligent medical professionals to be perfect.
And yet, we get angry at LLMs for not having perfect judgement, even though their creators are extremely literal about how they can make mistakes.
They're obviously achieved in drastically different ways at a low enough level; LLMs obviously do not simulate neurons or any biological construct. (For the record, I'm absolutely not one of those people who thinks LLMs are "alive" or should be treated like they are)
Reminds me of the olllllld days of Pentium II's when people got N64 emulation working shockingly quickly using HLE techniques. If you weren't around for this, it was quite the shocker at the time. I think the analogy is doubly apt, because HLE emulation has some serious limitations... it gets you maybe 80% of the way there really fast, and for the remaining 20% you need to roll up your sleeves and do serious LLE.
https://en.wikipedia.org/wiki/UltraHLE
It takes the prompt and continues it based on weights in
the training data. If there is no data it picks the most
likely thing (maybe made up). If there is it’ll mostly
add things from that data. Maybe it’ll make tool calls and
pull in data that way too but you can’t actually trust all
the details.
Note that I'm not arguing that LLMs are particularly talented at this particular use case. I'm pointing out that humans are also pretty unreliable.
You're also doing that thing where you point out that LLMs can be unreliable (yes, they are) without acknowledging how flawed nearly every other source of information is: people, websites, etc. I'm not defending LLMs in that regard... I'm just saying it's not a differentiator.
Humans do not in fact do that, they reason based on a mix of past experience and emotion, consider what is good and what is not and then answer. These are completely different processes.
The difference becomes apparent when an LlM makes a mistake for example and then apologises obsequiously and repeats the mistake, or apologises and makes a different mistake. Or when they fail to count letters (one of many flaws monkey-patched by calling tools).
They don’t reason they don’t evaluate and they can’t count. This is so so far from human intelligence.
From finding areas with favorite activities for each parents, teens and kids to discovering the do-not-miss attractions and scheduling our vacation between them - it is invaluable. I've seen places I never knew existed in countries I've never been to before and speaking languages I did not speak.
Very few mistakes and lots more flexibility and understanding than the travel agents I used before. I do write long prompts though with lots and lots of info about our family and what we like to do.
Not yet good at finding, filtering by our criteria, comparing and booking available accommodation yet, but it's getting there.
It has a downside - I'll never do these pre-arranged trips where one is in complete luxury bubble, interactions with locals are the best part of experiences. What a waste of potential.
And yes its mostly compatible with kids, it depends more on specific location than mode of travel (ie avoiding malaria/dengue/etc. regions)
"I told my bumbling assistant to plan a trip for me and he got nothing right but I enjoyed it because the chaos introduced a certain spontaneity and whimsy missing from my life"
I am also under the impression that the LLM tech is plateauing before bringing the promised productivity. Great as a coding assistant, great a summarizing a text, translating, great a helping plan a trip...
But for the rest, e.g. act as a life assistant, it is still far off with no hope to reach the desired performance level.
I would not be surprisd to see OpenAI and the likes to start reverting to Siri v1 strategies, i.e. "if this then that" kind of agent routing.
Now people want to handle car rental. What are the relevant data that models were trained on for this kind of application? For Python code there is kirjillion examples on Github, for mathematical proofs there is endless stream of papers, books, etc. But for car rental? Mostly adds in the internet that want to trick you into a bad deal. So yes, LLM will be a disappointment, as it tries, well, to trick you into a bad deal. In addition, data are rather scarce so there will be a lot of hallucination, as it gets mixed up with yacht rental, bikes rental, ski equipment rental, etc.
The performance of specific tasks will depend on either those tasks having been included in the training (which Apple could work on), or added by ways of fine tuning, and context sourced from userland.
For any category of tasks, there's a ton to be gained still in terms of how context is populated more effectively (relevance) and efficiently (token use). See software engineering harnesses and the skills architecture of OpenClaw for example. SWE harnesses make all the difference in how well Claude Code and OpenAI Codex perform. OpenClaw can't do shit without loading skills from the filesystem into context JIT.
I'll be very curious to find out how Apple is feeding context in their new AI approach. Part of it appears to be an 'index' that my iPhone started building (visible in main Settings screen) after installing the iOS 27 Developer Beta.
If new Siri still sucks, well, it's sucked the entire time. The worst of it is the security aspect where the setting to let you use Siri without authenticating hasn't worked since they added it! (still broken, iOS 26.5)
Have a conversation with the average Ai power user (outside of tech / coding) and this is the level the conversation will be on.
The laundry list of object removal, spacial photos, better speech to text etc is always just the latest open models just being slapped in there and branded as Apple.
Ultimately the meat of this presentation was the work of people outside Apple.
Where I do want AI is for really complex queries, like "find me a time and money efficient itinerary through Europe visiting places I haven't been before. Present options and I'll tell you what I don't like about each of them then we'll narrow in on an optimal solution"
Now the big (BIG) caveat is that I used Claude Code on my Max 20x plan from within VS Code. I have a fairly decent harness that I'd built and was sure to prompt it to run several subagents, including one that grounded walking times with Google Maps directly.
I'd say this is FAR beyond what the average person would do ("Hey Siri, plan me a trip to Prague") but also it shows that the models can do it with the right harness and guidelines. This wasn't that hard for me to do, so it seems to be more of a feature buildout ("the travel expert" AI) with a few markdown files than anything.
All told: web search for grounding times/locations, map grounding for walking paths and times, an adversarial agent to keep the model(s) honest, and a little bit of prompting and you've got a really great travel planner.
In short: the average person won't do this, but if I can build it in a few hours any of the 100% of people working at Apple/OpenAI/Anthropic who are smarter than me can build it and bake it into Siri (or ChatGPT, Claude, etc).
A lot of these "problems" seem to stem from people just not wanting to interact with other people at all. Do we really want to become like Asimov's "Solarians"?
It sucks. This is why none of my local supermarkets have real checkouts anymore.
Nobody wants to make a phone call anymore because most calls are scams; phone networks are terrible and apps have replaced them, like a lot of legacy tech.
Supermarkets make more profit if they pass on the checkout labor to the customer. That’s the whole story.
These generations are disillusioned from decades of decline in our society that have root causes predating any of them.
This has nothing to do with the unwillingness of young people to use a phone to call a business.
> Supermarkets make more profit if they pass on the checkout labor to the customer. That’s the whole story.
It's half the story. The willingness of young people to accept it, and even prefer it, is the other half.
Well, I think Siri AI puts this notion firmly to rest. Yes, if you have unlimited tokens and well-posed problems you can solve open Erdos problems. However, if you have meaningful real-world computational and reliability constraints then you better just stick to "summarize my messages and find the dogs in my photos".
And this isn't just Gemini, I can burn effectively unlimited Opus tokens and still get garbage code out or be run around in circles without very diligent oversight.
The genius part is that the menu is interactive, so you can add items to a shopping cart, which then results in a local language text you can show to waiters asking them for your full order.
It was a great sample of how even a little bit of ux can go a long way.
No, this isn’t the same as planning a multi-day vacation. But it is plainly useful today, and it feels very close to handling more complex tasks like that.
Maybe the difference is the model and the harness. At this point, I’m starting to think some people are either gaslighting themselves about how useful these systems are, or overgeneralizing from one narrow setup. Gemini, for example, seems especially weak at agentic behavior.
The wholesale dismissal just feels strange coming from the HN community I’m used to.
It's just not compelling to say that an AI can do an easy task quickly. This is still worth zero dollars to me.
I installed iOS 27 yesterday.
I asked it: please notify me when the temperature goes above 80F (so I can close the windows).
Siri responded: it'll be 99F today in Phoenix.
...
I just asked Siri a few weather questions and named the city where I live, nailed it. My favorite digital device is my Apple Watch and if Siri improves over the next hear or two, that will be great for me.
Presumably you might be able to task it with planning an itinerary with specific dates and bookings in mind, and then ask it to complete the task…sort of. The big gotcha i think is payments. Obviously you wouldn’t want to enter your credit card details into an llm lol. perhaps it would be ok if you had a saved card on file with your favorite airline, etc? Or maybe chrome has a feature to autofill a credit card for quick entry? Not sure.
Still…it’s a messy unsolved problem and we’re definitely not there. I wonder how this tech will look in 10 years from now?
AI as a "product" is about sucking up data for corporate interests first, then providing functionality to common people last with probably a few other steps in between.
Marketing departments have to twist themselves into pretzels and invent customers that don't exist in hopes to sell AI to people who look at those fake customers in the ads and go "Gee, I wish that was me!". People who casually book trips to Japan to shop for vintage clothes generally don't exist in such large numbers that they justify entire product stacks.
Here's what I need AI to do. Open an app, perform an action in said app, close app. Maybe open multiple apps and do things in other apps that are contingent upon data from one of the other apps.
Here's what AI can do. Poop Emojii with glasses....
It’s like they were trained on corpuses of box ticking material, like iso 9000 documentation, or security certifications. And now they know how to describe what they should be doing, but they never actually do anything.
100% of the time turns of all the lights in children's bedroom. Alexa has no problem with this.
Disappointing to say the least. Completely useless, I was going to get an Android this year on upgrade cycle. Will check this out first.
idk if most of my home assistant automations have actually saved me time since i def had to debug them, but the level of satisfaction when they do work is def worth it for me, since i created (and debugged) them haha
I'm in the middle of remodeling a new apartment and all my switches are smart. I won't even have physical switches for some fixtures like window rollers.
This is a fascinating example. You would initially assume it's just inconvenient versus flipping a switch--this can't be labour-saving by default.
The only way it makes sense is if you're doing it remotely; from another room, you're putting your children to bed. That's even weirder though, because you're taking what should be a moment of connection/care and trying to automate it. I guarantee your children would value you taking the time.
It's a use-case that is either inefficient or inhuman, and I find it really odd that it's one that you value.
Why waste time and effort just picking apart what someone else does with their free time? I can only assume becasue they disagree with the issues relevance, but that only goes to show the intent of the person replying. They dont care about the tech issue and just want to show why they think they are better than the person with the problem.
Human condition i guess!
Considering the uses and impact of tech is part of talking about it; you can't limit discussion just to the wires.
You could have simply asked without denigrating the commenter.
I do think that the sententious tone-policing of yours and other comments is injurious to discussion.
> That's even weirder though, because you're taking what should be a moment of connection/care and trying to automate it.
"weirder" has a pretty negative connotation, hard not to see its usage as denigrating.
"what should be a moment or connection" is an assumption on what the moment actually is. All OP mentioned is that they want to turn off one light in one bedroom. Your comment invented a narrative about what that must mean and then made statements on the narrative that exists only in your mind.
I don't think anyone is trying to "tone-police" you I think people just disagree with your take.
It's difficult to have a discussion if a bunch of unrelated people jump in not to add but to quibble about phrasing. If you don't want to discuss anything, just don't involve yourself.
In any case, OPs reasons for wanting to turn off any light in any of their rooms are unknown to us.
Maybe they took their kids out to breakfast and realized they forgot to turn the lights off while they were driving. Good thing they bought those smart lights that can be controlled with siri! Oh no! It doesn't work the way it was advertised!
There's no reason to imply OP is a bad parent just because they want to turn off a light remotely.
There's no reason to hunt for a poor parallel to shut down discussion.
So you tell me, what are we discussing?
BTW, the very HN guidelines you link to say that if you think something violates the guidelines, flag and move on without engaging.
I also have a "go to bed" scene that turns on a couple lights so I can see the stairs and turns off most lights around the house.
I don't really need AI to do it, I can just use the app, but Alexa usually gets the job done and I don't need to look at my phone.
I would have preferred Siri because one less provider but it's just unbelievably bad for this day and age.
"Siri, start a stopwatch" - runs the App "Stopwatch" without starting it
Such errors happen maybe 50% of the time. You can never just ask something without double-checking afterwards.
Set a timer for 50 minutes…always get a timer for 15.
There's never going to be a situation where a heavy Google Assistant user switches over to Apple for Siri. Anyone who would have switched from Apple to Google for their assistant likely would have done so by now. Siri just isn't a very important feature. It doesn't bring people to Apple's platform nor does it steer them away. It might bother users that it sucks, but it doesn't bother anyone enough that it hurts Apple's bottom line. Frankly, continuing to pour money into that bottomless pit does more damage. I wonder why they do it.
With Google Assistant (old assistant) I could say "Hey Google, play daft punk" and it would start playing Daft Punk on Spotify.
With Gemini (new assistant) it says "sorry I cannot play music, but here are links to services where you can find Daft Punk albums".
Fortunately at the moment you can still toggle between them. I guess not for long though.
So now I just use the Google home app and it works as expected.
If I specify “Shuffle playlist _____ in Apple Music” somehow that works right, even though it’s still using Apple Music in the first example when it plays the wrong music.
We’ll see if they managed to unfuck it with the new Siri update, or knowing LLMs perhaps they’ll make it non-deterministic so sometimes it works and sometimes it plays music you didn’t ask for.
Pretty bad if simple one-word commands to system APIs don't work, we had better voice control capability with "Speakable Items" on classic Mac OS.
Orders of magnitude less than the literal trillions that others have?
Huge call if so, given that missing the bus on AGI (if AGI happened) is a universal existential risk, but it turned out to be the right one.
I'm not sure it is, if your primary business is selling hardware. People are still going to need a 5" wide screen and battery in their pocket in order to do things.
What I want Siri to be able to do today is the same as when it launched with the iPhone 4S about 20 years ago: Just set alarms, calendar invites, tweak device settings, and look up answers on the web. The first three it could already do prior to the Siri revamp, the latter is a really nice nice-to-have for iOS 27... but beyond that, I don't believe that AI has many jaw-dropping areas of advancement within the use cases of consumer electronics. B2B applications of AI is where the money and the wow factor is really at.
I think that we are now in the realm of diminishing returns regarding these chat assistants.
What do you mean exactly? Audio conversation only? If so I don't see it very practical for most of the things
Voice interfaces can be silent.
The home and widgets screen can be customized to the point you don't recognize it as iOS
> devs customize the UI
Have you used Spotify? It completely ignores Apple UI and does its own thing cross platform. If you mean let devs customize the OS' UI, why would they? UI consistency is one of Apple's core strengths (or so it was before the 26 releases).
This is a truly damning comparison. In Star Trek, the massively powerful ship's computer is mainly ignored in favor of touchscreen interfaces and the natural language voice controls on the computer are mainly used for making tea and occasionally asking a question, which the computer often can't answer or answers incorrectly. All real work is done using other interfaces.
Is it available in China at least or is this another “50% of the userbase gets nothing new in the OS update” year?
Edit: https://x.com/wongmjane/status/2064052590992916840?s=46
Lol
What's really happening is Apple unilaterally withholding features while making vague noises about regulation as bargaining chips in talks with EU regulators where Apple is trying to weasel out of punishment for breaking anti-monopoly laws.
It's legit to be skeptical on the privacy front, but giving deepseek access to my entire phone. Or the TrumpAI at some point in a dystopian future seems... not great.
Opening up an API does not mean that everything on the phone is accessible to anybody.
So, where developers comply, all of that content is now accessible to those alternative implementations.
It’s not full read/write of the phone, and it’d exclude obvious secrets like passwords, but it is quite far reaching access.
I don’t know what sort of restrictions they can put on the alternative implementations. Can I vibe code one and have it live in a week? or is there a minimum bar?
The way I see it: If a user willingly (1) installs another AI app like deepseek and (2) willingly gives it access to 'full phone and app data' with a warning screen or setting of whatever that seems... like a good thing?
I may not agree with those users that it's worthwhile providing their full private data to [some AI startup X] or [Some Chinese or US AI company that will hover up as much for their own use] but if the EU forces Apple to provide this as an option, that sounds good to me.
The whole point of the regulation is that the data on the device is _the user's_ data and if Apple can have its AI services work with the user's data, competitors should be able to do the same.
From my (admittedly European) perspective it looks like Apple is just throwing a tantrum here.
One reason is that the data on a user’s phone isn’t solely owned by them. Some of it is shared with other people, or “belongs” to someone else: chat, email, shared documents, photos of people, contact information, etc.
In a corporate environment, this is more explicit: you have access to company information, so the IT department controls what apps you can install / run, because individual EEs won’t always make the best choices.
Second, I think app developers are more likely to share more data, if they know that the shared data doesn’t leave the user’s control. And that (presumably) makes the feature work better. If I’m developing an app, I’ll think twice about indexing any sensitive data, if I don’t know where it was going to end up.
Don't get me wrong, just like you I personally would also prefer LLM-integrations with a privacy-focused provider and I think Apple is a good party to get that from (assuming they're using good models and keep their privacy guarantees here...)
But in the end you're still often 'sending data to an LLM provider', and the EU enforcing them to also let that be competing LLM providers still doesn't sound like a bad thing to me.
If Mistral would give the same privacy guarantees: great! If a company wants to use their enterprise OpenAI subscription: great! Etc. etc.
Let's allow for some competition here and not force a specific LLM-provider onto users just because they like the Apple hardware and software ecosystem.
I saw that. Maybe you’re unfamiliar with Apple’s Private Compute Cloud? It’s intended to allow cloud computation on data without making the data available to anyone, which I think backs up my interpretation that apple’s stance is “no one should have this data, not even us”
This is from https://security.apple.com/documentation/private-cloud-compu...
We designed Private Cloud Compute with core requirements that go beyond traditional models of cloud AI security:
* Stateless computation on personal user data: PCC must use the personal user data that it receives exclusively for the purpose of fulfilling the user’s request. User data must not be accessible after the response is returned to the user.
* Enforceable guarantees: It must be possible to constrain and analyze all the components that critically contribute to the guarantees of the overall PCC system.
* No privileged runtime access: PCC must not contain privileged interfaces that might enable Apple site reliability staff to bypass PCC privacy guarantees.
* Non-targetability: An attacker should not be able to attempt to compromise personal data that belongs to specific, targeted PCC users without attempting a broad compromise of the entire PCC system.
* Verifiable transparency: Security researchers need to be able to verify, with a high degree of confidence, that our privacy and security guarantees for PCC match our public promises.
- - - -
Second, according to their press release ([1] and a sibling comment elsewhere in this chain), they’ve been trying to find a way to allow interoperability without giving full access to everything. Unsuccessfully, so far. So it’ll be interesting to see where it goes, but I’m sympathetic to their current stance.
[1] https://www.apple.com/newsroom/2026/06/due-to-dma-siri-ai-de...
Surely there's something better we can do than say "the average user is a dumbfuck better consolidate all control with Apple".
Couldn't you make a more believable straw man, please? The "Nigerian prince wants to send you billions" is really tired. Try something more emotional! Hackers will steal your kid's photos and post them on pedophile forums or something. This will resonate better with uninitiated and allow to easier lobby monopolistic practices. Good luck!
Which is fucking stupid, and Apple will never, ever throw open the gates to something so dangerously braindead. Their entire reputation depends on it.
And China is kinda self-explanatory.
Facebook got roasted for this, but now the EU wants the same open data policy from every big tech company.
These are the same to you?
Personally I would consider withholding products or features from the EU, not because I want to "steal everyone's data", but because they're a pain in the neck for a small business to comply with.
Personally I think EU policies single handedly ruined the web. Every time I'm shown a cookie policy I am a tiny bit more angry at the EU. If I were Apple I would drag my feet way more. They probably only do it because shareholders would force them to go after the Euros if they didn't try.
Apple’s performative DMA outrage is getting more pathetic by the iOS version.
On Android you can just have two accounts logged in at the same time, so there's no need to switch regions in the first place.
>EU users will be able to access Siri AI on macOS 27, visionOS 27, and watchOS 27.
https://www.apple.com/newsroom/2026/06/due-to-dma-siri-ai-de...
Apple cares greatly about their brand yet this has hurt their brand like nothing else in the past decade
but Apple isn't known to make grand promises and then not keep them, is it..? usually they just deliver what they say they will
yet i've been reading about "well they promised AI Siri two years ago and Siri still can't set an alarm right" in every thread even remotely related to the topic
i don't remember reading this much about anything else. it seems to have soured people quite a bit, at least in my internet bubble
Also, even when you DO get AI into products, consumers might not like them. The overuse of copilot led to a barrage of Microslop jokes, for example
And the Mac Neo is a best seller. Yep they really hurt themselves?
The only thing hurting Apple right now is memory like everyone else out there all because of the AI data center fiasco.
I don't see strong evidence the average consumer is demanding 'AI features' in everything. I mean even amongst the technically inclined this is often bemoaned, anecdotally.
why don't they just wait and not ship any AI junk at all? instead of promising a Siri AI rework, which then doesn't deliver? or Image Generation stuff that feels wildly put of character and generates tasteless and often downright creepy images?
not to mention that all of the new AI stuff they announced won't go live in China and the EU for a while.
why not do exactly what you proposed and wait it out? instead they seem to be trying to deliver AI stuff and just unable to.
there's also reports that apple execs held a secret emergency "oh shit what do we do about AI" type meeting.
they very much didn't intend to be this behind
and usually they don't care about glaring omissions like this, either: iPadOS was lacking a calculator for yeeeeaaaars.
they repeatedly said they'd ship a calculator if they can do something special for its introduction, and only then.
so why did they lose their hesitancy to ship mediocre stuff here?
I'm not even entirely against genmoji. It's just odd for Apple to do.
I mean, that was one of many things, and I'd argue the least interesting by far. If the Safari extension creation thing is decent at all, that's a seriously cool addition. There's some real value shown in this most recent WWDC. Whether they actually release it this time is another question...
So while they could win, it’s pretty hard to get hyped about it before we see real-world tests.
Why? What strengths and structural advantages do you think thy have?
What black swan situation could arise that Apple cannot counter?
If there’s truly an existential threat to its device business, Copy Well
Apple would never willingly pay Nvidia for GPUs anyway.
Why absorb supply chain pricing pressures and volatility when you can pass those costs directly to the consumer?
Now [relevant parts of] Siri AI is restricted to iPhone 17 / iPhone Air and more recent models.
People who believed Apple and bought an iPhone 16 to use with Apple Intelligence are getting the shaft.
https://clarksonlawfirm.com/lp/apple-intelligence-false-adve...
And there was a lawsuit and those users will be compensated.
This new update for 17pro is no longer misleading
“Apple has agreed to pay some iPhone buyers a collective $250m (£184m) to end a lawsuit accusing the company of misleading people about new artificial intelligence (AI) features and capabilities.”
Screenshotted in case they change it https://imgur.com/a/n1I3z8g
It’s also a text from 9:14am the same day.
If this is the second slide in your marketing slideshow, you clearly have nothing better to show.
Ok? But if I search for something just show me the thing I searched for (the proof). There is literally no need to repeat it in a slightly different tense. Who is that helping?
But also:
a) do you trust the llm to get it right 100% of the time? Because i’m gonna always read the original message to make sure.
And b) just excerpt the message “mike says i love it here at the snow … but i need snowboots” if you’re so desperate to shoehorn LLMs into everything, that’s just as easy a task for them as summarizing is.
But for whatever reason - because this stuff is mostly useless, because they’ve all got ai psychosis, or because of general dysfunction in the corporate structure - they can’t come up with a better example.
More likely, they just wanted an image that shows it can search your text messages for context. Most people still aren't familiar with AI, so having a very simple example image[¹] for people who may not have used an AI search tool before is useful.
[¹] I'm sure you and I can both think of more complex questions to pose to the Skinner box, but we're both familiar with AI and know what to expect.
I don’t think that’s true. Who are these blissfully unaware masses, and how do I join them?
I guess it depends on your definition of “familiar”, but it’s in the news every single day, and you can’t even google something without being slapped in the face with it. I feel like the only people left unfamiliar are people who have no access to the internet - and by definition they aren’t buying iPhones
"Try describing something different for the shortcut."
I guess I shouldn't be surprised that it still doesn't work.
All my automation shortcuts can be easily explained in pseudo code under 5 minutes, but it took me ages to put them together because that weird UI/UX forcing me to drag-and-drop squares around to manipulate data structures. Programmers hate it, non-programmers can't understand it, it is not designed for anybody.
AppleScript even had "dictionaries" declaring their commands and everything, would have been perfect way teach LLMs how to automate applications.
Seems like the logical next step
Siri seems to rarely get better and sometimes actually get worse.
This has been a problem on iOS since the dawn of time and has nothing to do with AI
I can go ask Gemini questions that require it to get information from several emails at once like “which x vendor had the lowest price”. Im assuming it can do the same with my texts or, if not yet, it will soon. I had zero such faith with Apple.
I will wait until the fall and see if this looks like the germ of something actually useful before deciding if I’m going to switch.
Pixels have had it for years but I think it's in every android phone now; from the app switcher you can press and hold and the system will OCR the text and allow you to copy it.
Because so many things _still_ don't make that easy to do on mobile.
I switched to iOS this year and I’ve been learning that the grass is not much greener. I do miss uBlock Origin. Maybe my next stop is GrapheneOS or a similar degoogled ROM…
Can you give a couple of examples?
>The Passwords app alerts you to weak or compromised passwords and can update them on your behalf without the hassle.
Finally, I hope this works well. Personally one of the worst things to deal with.
I don't really believe in Apple being that quality team.
Why?
“I don’t prompt Claude anymore. I have loops running that prompt Claude and figuring out what to do. My job is to write loops”.—Boris Cherny
They have no expertise in this area and their software quality as never been worse.
AI could potentially help solve those unpopular site/app/whatever edgecase.
Unfortunately not for other fields like email, notes etc…
IMHO the perfect password app could just keep all previous versions of any field until the user deletes the history.
git + somesite.com.gpg
https://github.com/FiloSottile/passage (or: forked using AGE instead of GPG)
There's a 0% chance it will work. Most websites I've seen have one or all of:
* Force you to use email or SMS as a "second factor" to unlock changing password even if you know the old password
* A stupid idea of password complexity usually requiring one of a finite set of 5-8 "special characters" which is often only revealed after you've chosen a password that doesn't have them. Or in some cases even banning characters other than the ones they check for. There's a standard for this where you put a regex on the password field, which a good password manager will always use, but the kind of idiots who think limiting the entropy of passwords to increase security is the correct way to do things almost NEVER implement this.
* A maximum password length, even as short as 16 characters in many cases
* CAPTCHA etc.
Any effort spent on this would be better spent elsewhere, including even educating other companies on how passkeys should be used.
Also, the Venn diagram of "memorable" and "reasonably secure" really only intersects in the region of "Correct horse battery staple" phrases -- and the problematic sites I'm talking about nearly always limit length, which thwarts that type of password terribly. What is the purpose of maxlength on a password?? These shouldn't be stored in any form other than a hash, so unless long enough to pose a DoS threat during the hashing process, length is truly none of their business.
Apple has detectors for codes sent via email or SMS, if your email account is one that is configured with the OS mail client.
> A stupid idea of password complexity usually requiring one of a finite set of 5-8 "special characters" which is often only revealed after you've chosen a password that doesn't have them. Or in some cases even banning characters other than the ones they check for. There's a standard for this where you put a regex on the password field, which a good password manager will always use, but the kind of idiots who think limiting the entropy of passwords to increase security is the correct way to do things almost NEVER implement this.
An AI agent can read the failure message and craft a new password
> A maximum password length, even as short as 16 characters in many cases
Same deal
> CAPTCHA etc.
While there's always the complex solution of scanning the image and trying to detect what is going on or slide the puzzle with enough of a curve to act like the motion of a human limb, there's also Private Access Tokens, supported by both Cloudflare and Google-provided captcha systems now IIRC. The OS uses an anonymous system to assert a single bit that there's proper browser chain-of-custody.
> Any effort spent on this would be better spent elsewhere, including even educating other companies on how passkeys should be used.
There are proposals as well to provide API to do upgrades from passwords to passkeys as well automatically. Nobody said the feature has to always use AI - but it may help the feature be robust enough for people to seek it out and try it.
I don’t think I’ve seen a single category of UX fail as hard and as often as auth screens do. It’s like at some point after 2015-2017 developers were struck with mass amnesia and forgot how to build decent login UIs.
Apple cannot compete in AI and has to adopt Gemini
Google is a really amazing company.
Nest devices are garbage, it has been like a decade since their phones were competitive with other vendors, ChromeOS was barely hanging outside of education center and now it's a zombie walking since Apple released their cheap Mac Nano, Gemini is a joke as a product compared to Anthropic and ChatGPT. The only things worth something are Chrome, Gmail, Workspace and their Cloud.
If I want to use AI, I want to be able to select the exact messages / photos which I want to send to it. Otherwise I expect the device to keep the data protected. I don't need any of these features either; I can remember if someone sent me a cookie recipe.
E.g. "the user asks if their Bitcoin private key is unique, let's make a web search".
I'm curious how the pricing will work. Would it be free up to some limit and then some subscription pricing? I can't imagine it can be free unlimited usage given the price of serving these models.
Meta also realized this and attempted multiple times to build their own hardware but they've given up each time. They started as early as a partnering with HTC in 2011 to make a Facebook phone.
Quite frankly, I'm kind of excited to see what OpenAI can build. I think an AI-first phone could challenge iOS and Android. It's a new paradigm and if OpenAI gets it right, it'll be very hard for Apple and Google to pivot.
I personally think chat + code is the future of apps. For example, I find myself wanting to do many things inside ChatGPT instead of traditional app because I can tell it to do things that are simply impossible on a static app UI. For example, I have some data I want to send to an app but before I do, I want ChatGPT to clean the data in some way first. And then after the data is uploaded, I want ChatGPT to pull some data off the API and make charts that I want to see.
I imagine a world where very intelligent models run at 10k tokens/s, app building is extremely standardized, and it simply builds any app you want inside the OS. IE, if you want a dashboard of your health data, you ask it to build it almost instantly exactly how you want it. I'm already doing something similar today but it's slow and not easy to do for non-engineers.
Incidentally, that’s what’s preventing Apple from rolling out their OS-privileged AI in the EU, as the EU mandates equal access for competing AI products. It will be interesting how this plays out.
If OpenAI makes their own AI-phone, do they have to let Anthropic and Deepseek run their models on it too?
provided it gets big enough, yes. the EU's position roughly is "if this hurts an entire market just to benefit you, and lots of people use / rely on it, then you gotta allow it"
If I want Anthropic to have the same level of access that Apple grants itself on _my_ phone, I should be able to do so.
If people want to buy open hardware, then just buy those. If they don't exist, make them yourself.
2. No, I don’t have the skills to. But if it’s so important to you, you can surely build it yourself.
It takes real time to drag along five ecosystems. That is the main reason it’s taking Apple longer than they’re so-called competition? Noticed that Google and Microsoft only do bits and pieces. Microsoft has no mobile and Google at its heart is an ad company the processor is six years behind.
though monopolistic practices are unequivocally bad and used to get struck down. one may argue this is just another instance of disallowing monopolistic behavior.
the DOJ used to have sharper fangs than what the EU is doing now
> Apple would have to give any virtual assistant direct access to users’ private data — and the ability to directly control other installed applications — as soon as Siri AI is made available in the EU
I was working in cellphone sales at the time and I can tell you no one wanted that phone back then even when Facebook was massive. An easy to hit facebook button was not a value add anyone was begging to exist.
Although with how many phones now have stock forced installs of Meta apps perhaps they won their con in the long game.
Similarly no one really wants a physical AI device, and attempts at such are pure techbro hubris on the companies part.
Similarly no one really wants a physical AI device, and attempts at such are pure techbro hubris on the companies part.
I want an AI assistant that I can use truly hands free. I keep my phone in my jacket when I'm riding my motorcycle. I want to be able to start, stop, adjust, and check details in route guidance. I want to be able to ask what the weather is like ahead on my route. I want to be able to ask it to start looking for a sensible place for me to stop for fuel and/or food without making me do a big detour.
Actually I would also quite like better driving directions, since I can't look at the directions on a screen.
The audio setup is not the problem.
The problem is when I say "Hey Siri, cancel route guidance," it will say something like "…You must unlock your iPhone to do that."
The entire point is that I can't take my gloves off, get out my phone, and unlock it, at speed, on a motorcycle.
They need software companies to give a damn and improve apple maps/gmaps/etc.
Or perhaps pull over and adjust your routing like most of us do, or maybe bolt a sacrificial phone with maps on it to a holder. If you don't want to risk your main phone.
The entire point is I don't want to have to stop in order to receive new information.
I don't know how to make that clearer, sorry.
Riding on a motorcycle is already dangerous enough with the average land tank driver on their phone. Talking to your AI assistant while riding at speed sounds like pending split focus disasters waiting to happen.
Are you a bot?
I’d go further and claim if controlling the hardware is a existential threat for these companies, their ability to deliver or not should be seen as an indictment of the entire “llm formulation of ai” era we’re currently in. If llms have the potential they claim, then empowering them to design a best of breed phone should be table stakes.
EDIT: To provide meaningful chat functionality they have to either eat up the cost or charge a subscription for it. This will be first time they charge for Siri - a product that doesn’t garner any positive reviews. This gets even more interesting to watch
It seems like revisionist history to say that; lots of people were sold on iPhones years ago because of Siri. They have one of the few business cases for voice assistants, which are notoriously difficult to actually monetize, that actually makes any sense, since "selling iPhones" is meaningful and "selling a subscription" would be nice on top of that.
> Aga: have you heard of Calanthea? It’s a plant.
Really groundbreaking use of AI!
wow! we are in the future!
I'm almost sure that sometimes searching the same thing will give you the result and sometimes it won't.
I'm sure they customized some of it, but this looks basically like Gemini integrated with iCloud instead of Google Workspace.
The text responses had Gemini's verbosity. Asking ChatGPT to show me iconic dishes from both Brazil and Morocco (Apple's example), is much cleaner, less verbose. Quick list of dishes and links to the recipe. Gemini just spews a wall of text and bullet points and goes on and on with fluff. Tons of "What this dish is" "Why it works" Same with its frequent use of tables, which I see less of with ChatGPT.
Each Siri demo they did in the keynote had that hallmark verbosity I typically get with Gemini without prompting it to not do that.
In my mind the Gemini LLM defines the bounds of capability and capacity, but any actual functionality or usefulness (or lack of) comes from Apple’s Siri harness.
Apple’s entire software stack has a branding problem.
> and you'll realise that no, this isn't some HN-specific peculiarity.
I think it is tbh
Hmph.
That said, I'm THRILLED they claimed to "fix" the border radius snafu of Tahoe. Go ahead and push that now with the next Security fix. We won't mind at all.
Using Siri essentially required me to use my hands anyway, so what's the point of voice?
I'd very seriously consider moving away from iPhone to a device that treats voice AI as a first class citizen (presently I mapped the 'double back tap' to open grok voice chat, and triple back tap to end it, which is a wonderful improvement over not having these, as you can do that pretty easily, even while driving etc).
I don't understand. Are you so little in control of your own self that you now need an embedded AI to do something for you that was already a complete waste of time and energy? Just delete the apps.
I'm sure you guys have tons of others to list.
This feels like it could be solved with a list of permissions that the user has to turn on when using 3rd party AI.
Apple already:
1) requires developers to submit ID to publish an app on the appstore (at least I had to after ~1000 downloads to be able to publish an update)
2) has strong kernel enforced memory integrity and disallowes arbitrary code execution (unless explicitely approved for games like roblox, jitting not allowed tho has to be interpreted).
3) reviews every app update.
It's really just Apple being angry about the EU's DMA endangering their golden goose (App Store revenue) and using any meaningful new functionality as a bargaining chip.
They've done staggered geo launches for other features in the past many times, both before and after the DMA was passed, and in this case there's even another great reason to not want to globally launch all at once (AI inference server capacity). If they can at the same time market it as part of their ridiculous turf war against the European Commission, I guess they just have to take the opportunity.
It just reads like arrogant foreigners throwing a tantrum over our laws.
At the very least to the extent that the whole setup limits national sovereignty.
Your comment comes across as though you expect us to believe EU citizens are a homogenous whole, who happen to align with your perspective on this matter.
> At the very least to the extent that the whole setup limits national sovereignty.
That's how anything grouplike works indeed.
> I struggle to believe there isn’t a significant fraction of EU citizens who are frustrated with the EU’s laws.
Sounds like something that'd have polling data coverage?
Across the gamut of regulations the EU has, it's not really the ones that apply to Apple that draw much ire.
They complain about things their competitors are able to implement with no problems at all, and then once their tantrums are not sustainable anymore, they’re apparently able to solve all the problems as well.
For a company that sells itself as a design + engineering firm, they seem to have very little confidence in their ability to design or engineer their ways around constraints pretty much everyone else is able to.
iPhone mirroring for example. Seems like practically 100% chance that if they put that out in the EU they'd be facing lawsuits for not making it work with every Android and Kindle and digital pregnancy test on the planet. And making it an open API right out of the gate is a much bigger undertaking than just making it work with your own devices through a proprietary API that you're free to break at any point and just update your devices to work accordingly.
Yes, it is harder to support APIs compared to… only having a completely closed system.
It would take a trillion dollar company to have the kind of resources that kind of thing would take.
The fact Apple thinks it should be exempt from rules that try to impose some fairness and choice, simply because “it’s hard” is the kind of thing that makes me feel they should be forced to be broken up in some fashion that limits the conflict of interest. They can be a petty platform owner who extracts rents from every developer OR they can offer things like apps and services on top of a platform. They are proving themselves too evil to be both.
If only there were a way to design a system in a way that didn't even require you to know who's on the other side of an interface. Some kind of... protocol specification maybe.
Imagine the things we could do, like making a phone call to a different brand phone than yours, or a web server running a different OS than the client, maybe even one that the other side has never heard of!
What is the purpose of that?
Nah, that just shifts the goal posts. If they did that, developers would be whining about "scare screens", as we have already seen when Apple put app installs behind a permission prompt.
They're already up in arms about the requirement from Apple (and Google) to know who is behind the apps that slurp up all your data.
The DMA maximalists won't be happy until Apple releases an anonymous service to automate setting up a Kafka topic to send each iOS user's PII to whoever wants to receive it.
Because we applied the learnings from computers and phones have ended up way more personal?
I'm extrapolating (there is less detail in that press release than I expected from your comment), but this sounds to be like it would be the thing that enables such a "list of permissions". I would be curious to know exactly what this agent entailed and why the EU did not approve it.
I want Claude and OpenAI to be able to compete with Siri on an equal footing. Just like Apple maps has to compete with google maps. Competition pushes companies to make better products.
In my opinion, Apple is doing the right thing for users. It’s not like they have a huge revenue stream here. Yes, there will be some features or usage that require iCloud plus or whatever to cover incremental cost, but I genuinely believe that they don’t want services creeping in that break their trust with users or their privacy-first reputation.
Apple’s decision (users will have a less powerful product because we’re not vacuuming up their data and using it for profit) is exactly the kind of thing the EU should want. No country has appropriate data privacy guidelines for AI (yet) so opening up choice can’t provide alternatives.
(To be clear, I’d be fine with Anthropic here, but am fine with this state. Maybe because I’m so used to Siri sucking that I’ve given up hope.)
The right thing for users would be to allow user choice, and for Apple to compete fairly.
Apple allowing third party access doesn't automatically mean user data gets hoovered up by OpenAI, Anthropic, etc. It just means users now get the choice, if they want to make that choice. Users could stay with Siri/Apple if they care about what Apple is offering, or choose to accept the risks and terms of service with other third parties.
The EU isn't saying "you must preinstall every competitors offering" its "you must offer the ability for others to hook into the same APIs to be able to offer their own assistant on par with the first party option."
The user still remains in control by virtue of their own choice.
So some government official will scour the entire API surface of iOS and decide which ones Apple needs to expose to third-parties? They have already decided App Store and Payments APIs need to be made available. Now it looks like they also expect off-device foundation models need to be made available to third-parties.
What about making Apple Watch specific APIs in iOS be made available to all third-party watch makers so any one can bring any smartwatch and use it just as effectively as the Apple Watch with an iPhone? What about all the AirPods specific APIs that lets Apple offer a better experience with AirPods than a generic bluetooth earbuds? What about Apple Pencil? And so on... If you go down this path, the list is endless.
Don't threaten me with a good time? All of those seem like great policies. The fact that I cannot use an apple watch with an android phone is ridiculous, and vice versa as well.
At some point this is just a debate about vertical integration. Apple can deliver better experiences with it, but of course it limits user choice.
Many people want fully modular, open systems, which is lowest common denominator.
I can see both sides of the argument, but I am so skeptical of regulators deciding what can be integrated or not. If modularity is better for consumers, why don’t they prefer modular systems?
At the very least I think there should be a very clear tradeoff; right now the EU seems to think they can regulate their way to all of the benefits of vertical integration while outlawing vertical integration. I don’t see how anyone could look at that with a straight face.
How did we go in less than two comments from providing access to APIs that are already present, implemented and actively used by Apple (who in their holy wisdom deem us mortals not worthy to access these the way we choose) to a completely different hypothetical of requiring actively building support for another companies hardware?
Such slippery slopes really aren't helpful, nor in any way comparable to what the DMA actually intends or states.
Because there aren’t any to choose from?
“Smartphone” has become a mandatory thing everyone is required to use to function in society without major friction.
Businesses hate supporting a ton of distinct platforms, as proved by the developer marketplace killing Windows Mobile through refusing to ship apps for it.
This suffocates any third entrants just like the FPTP voting system suffocates third political parties.
So what modular OS are people supposed to choose?
That's irrelevant to the discussion. SoC is not a digital platform in any way under DMA. It's not a platform at all.
I’ve read all of it, multiple times, and been grilled by EU regulators (vicariously, via corporate lawyers).
It still boils down to general guidelines that it’s impossible to know if you’re violating before the fact, and they will not even approve/reject proposals in advance. It’s basically “go read the act yourself, and ship what you think is compliant, and you’ll know whether we interpret the words the same way by whether or not we fine you.”
Good times.
Companies want to know exactly where the line is so they can figure out how to comply with the letter of the law while doing as much as possible to get around the spirit of the law. This has been demonstrated over and over again. It isn't the job of the regulator to help companies with this process.
I’m more of a rule of law person myself. If there’s a law that must not be broken, and breaking it results in penalties, it seems insane to me to not specify it in advance.
Sure, big tech is largely evil. Arrest ‘em, find them, IDGAF.
But pretending that DMA and related regulations provide enough information to ensure compliance is willfully ignorant. The regulations are designed to allow selective enforcemen.
Yeah we already have that. We have the words 'SLOW' on the road that ask you to slow down from the current limit for the hazard ahead, but pulling you up on this is officers discretion.
The DMA doesn’t have the objective measures. It’s all discretion, all subjective, all post hoc.
Which, cool, some people like the idea that police target those people and need flexibility to make life harder for undesirables in ways they would never do to high status people. I don’t personally like that, and I don’t think tech regulation should work that way.
The exact rules ultimately don't matter, because the EU is after outcomes. If the current rules don't lead to the desired outcomes, they will keep changing the rules, until they get what they wanted. (Or until their goals change.)
It's competing at the wrong level.
The iPhone is a toaster. Nobody's up in arms about whether the toaster takes other manufacturer's crumb tray. It's a television, and nobody's demanding QLED and OLED be swappable. It's a console. Xbox doesn't play PS5 games. It's fine.
There's no real line between hardware / firmware / software / malware ... For what Apple offers consumers, every layer of whateverware should be trusted.
Drawing imaginary lines based on the embodiment or substrates for logic gates is mistaken.
There are lots of phones. Lot's of different philosophies. Stop taking away consumer right to pick a philosophy and design for an end to end experience. It's fine.
It’s not an either/or thing, it’s about preventing so called gatekeepers from anticompetitive behavior via favoring their own accessories and services while simultaneously preventing any others from possibly competing.
There’s no valid reason at all a third party smartwatch shouldn’t be able to integrate to the same level as an Apple Watch. No reason third party Bluetooth earbuds shouldn’t be able use ADWL for automatic device switching, etc.
Want to still use only Apple? Great, nothing says you can’t. But at least it would be user choice and there would be actually competition which would lead to better products for all.
Can’t believe I lived to see the day that people on HN start defending vendor lock in and closed platforms as a good thing. Have all the hackers retired?
It's already user choice. The problem is too many users like the lineup. And too many who aren't going to use it, don't.
The EU is just a disaster with these laws. They can't innovate so they do this.
Good. They are making an operating system. User choice and competition matter. I know Apple would prefer to allocate more resources to Liquid Glass animations and burying more UI elements inside “…” menus, but I personally think I don’t need any more innovation above the OS level from Apple. Especially because 80% of their changes to the application layer in 10 years have just made their platforms worse.
Let them ship a stable platform that allows applications to do tons of useful things, even when you don’t accept a mega-package of apps and services all from the first-party vendor that locks you in.
If Apple built houses, you would have to jump through hoops every time you use a microwave or lamp you didn’t buy from Apple.
That's...literally the point of these regulations. Sounds great to me.
Apple is also restricted in the sort of consent prompts they give the user. That could matter when a non-technical users is prompted by a third party app to effectively allow unfettered access to all user personal data on the device.
Sometimes when you look at the functional requirements for a feature it turns out to be a bad idea. In the EU, functional requirements can come after-the-fact from regulator interpretation of the DMA. Until Apple determines what those requirements actually are going to be, releasing a potentially harmful feature is irresponsible.
This is not complicated. Even in the US, every other industry is regulated to your benefit, you're just used to it and haven't realised. Digital technology obviously needs to be too. And yes, you have to do it properly.
There's nothing "fair" about this at all. It's a group of luddites in the EU who dislike how successful American companies are.
Again, the EU is stifling innovation with these backwards-looking rules. No wonder they have no innovative companies.
However, even if your premise is correct, it does not matter.
In the end, trying to manage such products (require massive investment, have network effects, offer significant gatekeeping and rentseeking opportunities) is extremely problematic.
On one hand, the market cannot do it properly: There are tons of externalities, and, like e.g. building out rail, the absolutely gigantic barriers to entering the market means the existing players merge into a monopoly or oligopoly.
On the other, the product is too complex and too dependent on continuous evolution to officially turn it into a state-controlled / state-run monopoly (the solution many countries have deployed to solve e.g. how rail, or medical insurance, or road networks, end up in a terrible state if left up to the market).
So what is one to do?
The current crop of mostly US led large companies seem to have gone with a 'just trust me, bro!' argument, with some 'AI is so important you cannot put up any roadblocks at all!' sprinkled in.
And yet these companies time and again prove that they can't be trusted. Which is obvious and logical: Companies must conform to the law, but are otherwise amoral. Or rather, their 'moral' compass has nothing to do with human moral compasses: They must earn money for their shareholders, in whatever legal way they can find that is most efficient, paying as much attention to future company growth and health as its shareholders desire. That isn't just 'what they are incentivized to do' - that is what they are legally *required* to do.
And yet you've gone with a motif of 'but apple is the one company that is doing it right so lets just trust them.. bro'.
There *is* a solution:
Use the fact that the state has powers of persuasion that companies simply do not have. The threat of law, and the monopoly on violence.
Essentially, a state can simply tell a company: The populace have spoken and they value X (say, privacy). They value it a lot. You will deliver. At low cost. This is not a request, it is a demand. If you don't want to or can't, then we shall write laws to regulate you and then *everybody loses*.
Conceptually this works, in a weird game of chicken / madman theory: If the corporation in question believes that society will regulate them into oblivion unless they comply with society's demands even if this means society incurs a great cost, then the corporation *will comply*.
This has happened before. There is no actual law in the US that a movie gets a rating, and the movie industry pays for and manages the ratings of its movies entirely as an internal affair. And yet, in general, movie ratings are stellarly well run compared to what a government run institution would have done.
The reason *is* that threat. The movie industry decided to police itself because it was quite clear that if they did not, the government would have, at great cost to the movie making industry (and at significant cost to society as well, in the form primarily of much worse films).
For some reason that isn't entirely clear to me, CEOs of large corporations that deem themselves 'IT companies' do not understand this part. They will fight tooth and nail to fight every law, and especially in the US, perhaps due to extremely dire and long-term distrust by its populace in its own government, many of its citizens incorrectly side with its corporations on this idea, even though time and again corporations prove that they have no allegiance other than to the almighty dollar (which, to be clear, is not a complaint. That is how society has set them up. My only complaint is that e.g. you seem to have forgotten that this is how it works).
Hence, given that the system works on, in essence, fear / coercion, the only right answer is to do an attitude adjustment, find a massive club, and beat a whole bunch of IT companies into absolute pulp until the remaining CEOs understand.
And before you make a note about the brash, medieval nature of that comment - it is already clear that these CEOs who think they are God's Greatest Gift To This Planet, are already meekly running, tail between their legs, to kiss the pinky ring of a personalist wannabe emperor president. They are _clearly_ motivated by such fear and _clearly_ cannot be trusted to rise to the occasion and be a new form of benevolent leadership for the citizenry.
I wish they were. It'd be so much easier.
Google can easily argue that if Apple gets to rule over a walled garden, zero-API ecosystem where no one else can compete, then it’s right that they can too, regardless of how privacy-respecting they are or aren’t.
The device won't be able to ask for significantly more permissions than Apple asks for their own model for regulatory reasons, nor will it be able to convey the seriousness of granting the permission (e.g. immediately give unrestricted access to the vast majority of personal information/documents stored on the device).
But Apple also architected their system to justify not having constant permission prompts for access to sensitive data. And for regulatory reasons they also can't mandate that competing models have the same architecture.
The regulators and Apple (along with hopefully other AI companies) will need to work together to determine longer-term stable path forward.
1) ask for permission explaining the scope
2) warn you about the dangers with a confirmation / nevermind option
1) Acme AI requires access to your email provider in order to execute this request. Grant / Deny
2) You're about to let Acme AI read and send emails on your behalf, this might be dangerous due to X and Y. Do you want to continue? / Nevermind.
1) Asks for access to a service
2) Asks for a specific use-case of the service
2 is permission to act, but you might want to deny access to some parts such as sending email and scope to summarization
I will wait and see what people find out about it before passing judgement. It's quite possible that it isn't possible to have an API to use other companies' AI instead of Siri AI. Are there any equivalent API hooks on Android?
Our family uses Siri with a HomePod a lot, and it's already much better than it was a couple of years ago where it could basically set timers, tell you the weather. Now it answers questions ("when did the Knicks last win an NBA championship") with decent answers, instead of "I'll send the web results to your phone". But it's still far behind voice-chatting with Claude in the Claude app, so very much looking forward to this upgrade.
I will say though that proper voice transcription in Claude -- or any of these agents -- sucks. If it can't understand the question properly, then it can't provide the right answer. It works okay for me, but not for my kids, not when speaking quickly or in incomplete sentences (as people tend to do), etc.
In "English" later this year...
We've heard that before, haven't we, Apple? I feel the right way to fix the trust issues would be to announce this when it's actually done. Like, here's Siri AI, and you can download and use it, right now.
I believe we also heard that a couple years ago.
15 years ago they had the balls to run Siri live on stage: https://youtu.be/6rL9EL2LlrA?is=5yMQxs0C2VAC5Lwz
The responses came in very fast though, so I’m sceptical that the latency is representative (or that they didn’t cherry pick results, but they looked LLM generated). We shall see though.
I’m writing AI apps these days, and even pulling Gemini 3.5 flash on Google Cloud takes longer to get a multi-step response.
Obviously the video is not representative, and there are fast models on fast hardware. But if this takes 2 minutes it’s not very compelling to users.
After getting it in my hands, it's the same. At least 4 times slower for similar basic Siri responses. My guess is they are doing less local and more server-side generation to start as the on-device models might not be good enough yet.
It’s really disappointing to see the on-device models being limited to so few devices. And this was after the iPhone 16 and 16 Pro were marketed so heavily with supporting their now failed effort at AI.
iPhones have 12gb, current Neo has 8gb, the next gen Neo is speculated to have 12gb (as it'll be based on a later iPhone chip).
All the iPhone 16/Pro owners have been waiting for Apple Intelligence features announced from that WWDC 2 years ago. They didn't get delivered and now won't ever be delivered with on-device intelligence due to the 8GB RAM limitation.
At first I thought it was the usual planned obsolescence. Then I realized it may be a true technical limitation. I suspect an embedding model is required to run on device in order to make several of the features work. Embedding models are small compared to LLMs, but, depending on their capabilities, could be the memory driver.
It's shortsightedness on Apple's part, being both the first to the game, and the last to the game. Go back any number of years in the past decade. For at ;east 10 years Apple was telling everyone how great their local models are, and how amazing desktop-class their "neural" chips are. And yet here we are with them somehow incapable of running anything except on a few recent top-tier models.
> Apple’s most powerful on-device model and the features it enables, like expressive voices and more advanced dictation, […]
On other devices, I think there’s still on device support (just not with the “most powerful model”), for these devices:
> Apple Intelligence and Siri AI in iOS 27, iPadOS 27, macOS 27, watchOS 27, and visionOS 27 are available on iPhone 16 models or later, iPhone 15 Pro, iPhone 15 Pro Max, iPad mini (A17 Pro), MacBook Neo (A18 Pro), iPad models with M1 or later, Mac with M1 or later, Apple Vision Pro, Apple Watch Series 9 or later, Apple Watch Ultra 2 or later, and Apple Watch SE 3 when paired with an Apple Intelligence-enabled iPhone nearby.
This is from the footnotes on https://www.apple.com/newsroom/2026/06/apple-introduces-siri...
I do wish they’d been more clear about what the “advanced features” are :(
I still don't think Siri can do that ::angry::
i was waiting for siri update and bummed it is not supported on 15.
when i use my native language (i mostly do it when in carplay) to search songs etc.. it gets it wrong a lot of times.
+ a more integrated into things like imessages, whatsapp etc..
The stock price definitely didn't like it though.
It's not even funny, it's not smart. It's like if they released MS Siri and said it's Mac System Siri.
For pedantry's sake, they were saying "AI = Apple Intelligence" last year as well, so it's not like they just pulled it out of their butts now that popular opinion has turned against AI.
I think it just feels uncreative? Siri as a brand has some value, but if you want it to feel like a watershed moment where old Siri is "behind us" finally, just give it a new name.
This doesn't follow for me. They can trivially allow it to still respond to the old wakeword. They should absolutely change the name in the event they can finally make it useful, because "Siri" is (in my mind and many others') a synonym for "hapless idiot." "Thanks, Siri" has been uttered hundreds of time in my house and my car, and 100% of the time it's sarcastic.
Many of those people will speak a language that’s not English, or live in the EU or China where it’ll still be “Siri”, not “Siri AI”.
“Do you have the new Siri?”
“Yeah I updated… but she still seems so dumb”
“Oh yeah… well that’s Siri for you I guess”
Horrifying for marketing folk, I would presume. You’re just setting people up to confirm that Siri is always useless and improvements are invisible.
AI is a technology, not a product. Consumers don't care about technologies, they care about what the product does versus what they currently have.
I think Jobs was an asshole, but one good thing I can say about him is that he understood the difference between technology and products. Imagine if they had called it the "iPod HDD."
Siri and Voice Control were both usable during the same time and it feels like it could work here too.
Totally agree that AI is just an implementation detail though. IMO that new product name should NOT have “AI” in it at all.
To prove my point, I opened a random date on the Apple website matching today's date to compare. 16 years ago, June 8 (1) Apple released the iPhone 4. There's still no room for jokes about that release, and from this perspective, calling their AI 'Apple Intelligence' feels weak compared to what they used to deliver.
I agree that some years ago Apple was the strongest in marketing, their brand team had been setting the bar for tech, but I simply can't say that anymore.
1. https://web.archive.org/web/20100608073904/http://www.apple....
Your 15 Pro Max supports Apple Intelligence. Newer phones can answer more questions without going to cloud infrastructure.
It's the iPhone 16 line that feels a bit shitty not getting the latest and greatest since it was advertised as "built for Apple Intelligence"
Apparently the 17 Pro is the only currently released iPhone that will get the best local model. Which I suppose makes sense considering it has 12GB of RAM compared to the 16 Pro's 8GB.
At this point, “Siri” has a pretty strong cultural association with being underwhelming or unhelpful. Even if the new version is dramatically better, convincing people to give Siri another shot may be harder than launching the same technology under a new name.
Feels like a missed opportunity to reset expectations.
But that's a big If!
You clearly never used Siri before
It's beyond my expertise, but it's publicly available if you're curious.
Apple knows this which is why it is taking years to test and iron out the kinks. But somebody somewhere will make it hack a social media account, or give over somebody elses credentials, or generate illegal child images etc.
Pho is a pretty bad source of fiber.
It sucks that we're skipping over such good tools like cronometer.com to figure out what we're actually eating and going straight to hallucination, adding more confusion to nutrition.
I think a lot of it is the old "perfect is the enemy of good" with Apple trying multiple times now to announce this big basket of all these AI features supposedly coming all at once instead of just regularly shipping new useful AI integrations every month. There was so much easy useful shit that was immediately apparent as soon at OpenAI dropped that first big voice mode years ago coupled with basic app integrations. Particularly in the context of the AI labs that are operating in that lane almost too much where it seems a new model or mode comes out every two weeks.
With the most recent 'Apple Intelligence' function, it took a while for Apple to grant the ability to disable/enable each feature, then a bit of time for the respective MDM Software developers (Jamf, etc) to provide toggles.
Can't wait for unexpected password updates and naughty mails accidentally sent to my boss...
https://github.com/finnvoor/yap
I tried it and was pretty impressed. That said I haven't heard anything yet about them switching to this for the text input voice dictation in iOS but it would be really nice.
> Your data is never stored
> Used only for your requests
> Verifiable privacy promise
Apple is cooking. Although at that point might as well bring the cloud features to more devices. Yeah it costs more but also locks users in harder.
(It’s been driving me crazy there’s no “AI this” button to discuss whatever is on my screen.)
Note: I have MS 365 personal or whatever it's called this week so I'm not sure how Copilot acts for a completely free user.
> double-tap-the-bottom-of-the-screen feature that pulls up siri
It’s disabled if not using Apple Intelligence, and can’t tap screen while talking to Siri (it dismisses instead).
Now they’re gating features to the M3 I’m not convinced wouldn’t work on expensive Apple Silicon predecessors… am more convinced the double tap disable is intentional.
- Siri please suggest an organization for this folder - Siri over my last work in this app can summarize what i am struggle ?
Pro active:
- Hey, the last hour you exchange 30 mails from the same subject, i look with your team ai and all have same struggle, based your key points in communication is X,Y,Z, this an mail for final align
- Your and your partner don't have quality time in last day, i see has the seat available in your favorites restaurant for next hour do you want made an appointment ?
87% US teenagers own an iPhone. ~35% teens own an Apple Watch.
The interface for creating them manually has been so bad for so long, it feels clear to me that LLM-driven shortcut orchestration was always the endgame. Apple built up their ecosystem of composable "tools", and then trained an LLM on how to call them.
The result, IMO, is the first OpenClaw/Hermes competitor that's feasible for use by the general public.
Everyone with a paid Claude or ChatGPT that they're struggling to use to the fullest is going to have very little reason not to swap over to an upgraded iCloud+ plan (if they don't already have one). I suspect we're going to see mass cancellation of $20/mo plans very soon.
OpenAI's timing for removing their temporary increased usage limits is looking pretty unfortunate...
I have shortcuts set up to count the hours I log in my work Google calendar and copy them to my clipboard to help me prepare invoices.
So while I've already been sold on what Shortcuts can do, getting the general public to see the possibilities is probably gonna be a challenge.
1 is performance. It's slow. You can run one within the app and literally watch execution flow from one block to the next. Absurd, for the CPU power at hand.
2 is reliance on developers to deliberately implement hooks and "intents" when the developers of at least half of apps including most "big company" apps do not care to bother, often because 95% of their app's surface is actually cross-platform stuff.
Example: There are no shortcut actions for Google Calendar, and Gmail only has one real one which is a generic send email. No "search email" etc.
I'd rather see Apple lean into "computer use" to allow it to use any app that displays things on the screen, but IDK how you make that safe.
But fundamentally, the real difference is they have now bought and white-labeled Gemini to replace all the stuff they failed to make 2 years ago.
People outside HN will begin to expect they can do anything with a computer in the same way they expect to be able to say anything in a support chat. Using pre-LLM automated chat feels like a joke. You enter the chat expecting to having a conversation and instead you get a GUI phone tree.
This is exactly how it feels to use any of the AI tous from Big Tech and others.
We have entered the era of deeply personal computing. There are so many incredibly personal features that no mega corp could ever be expected to build. Now that lay people can build things, let them!
There's a podcast that I listen to which is translated to a bunch of different languages. It's exclusively on Spotify with no RSS feed. I have a cron that checks daily passes it to an LLM and notifies me as necessary. I did not code a line. I only set up an OAuth endpoint.
Enabling your customers to do things like this will make them incredibly sticky too! So please for the love of God, let me glue Siri into whatever I want.
I can't wait to take a photo of a cricket ball and ask it what it is, ffs.
These people need to get out, touch grass, watch trees swaying in the breeze, and put their phones down before they lose toonmany neurons.
I still look at older MacOS screenshots and think a lot of it looks better, but directionally they are improving Liquid Glass.
I really enjoy this lag. Apple with the whole DMA made iPhone completely dull to my eyes. Previously? Updated yearly. Now? 3+ years without replacement and probably will stick to it for next 2-3 years.
Sure maybe in US Apple is fun. But in EU it's.. boring (and not like a Golang boring, just boring)
- Take a selfie
- Create a reminder
- Call Vicki
- Rotate photo left
- Create a new event (do you create an old event?)
- Send an email
- Resume my podcast
- Create a note
- Add photo to album
...can't iOS do literally every one of those things already? What the fuck is happening?
What do you mean?
That’s what I expected from Siri but you can get in from ChatGPT .
Genius way to sell more phones
Really they are just selling on device Ai
> Siri AI coming in English later this year.
Strange way to phrase it, but okay.
> Siri AI will be available In beta later this year and requires an Apple Intelligence–enabled device set to a supported language. Available in English to start. Siri AI will not initially be available in the EU on iOS and iPadOS.
Ah okay, not EU enabled. The only reason for this, in my tinfoil hat, must be for data farming.
Wait... don't tell me... there is an App for that.