Apple barely talked about AI at its big iPhone 17 event | Hacker News Reader

Apple barely talked about AI at its big iPhone 17 event

theverge.com

Apple needs on-device AI to do chores for me with the apps I have installed. Apple has everything it needs:

* Apps are already logged in, so no extra friction to grant access.

* Apps mostly use Apple-developed UI frameworks, so Apple could turn them into AI-readable representations, instead of raw pixels. In the same way a browser can give the AI the accessibility DOM, Apple could give AIs an easier representation to read and manipulate.

* iPhones already have specialized hardware for AI acceleration.

I want to be able to tell my phone to a) summarize my finances across all the apps I have b) give me a list of new articles of a certain topic from my magazine/news apps c) combine internet search with on-device files to generate personal reports.

All this is possible, but Apple doesn't care to do this. The path not taken is invisible, and no one will criticize them for squandering this opportunity. That's a more subtle drawback with only having two phone operating systems.

I agree completely, it's really unfortunate how AI on apple devices has been going. The message summarization is borderline useless and widely mocked, meanwhile their giant billboard ads for it are largely stupid and uncompelling. Let me choose to give it access to my data if I want to do really useful stuff with on device processing. They've been leaning into the privacy thing, do the stuff that would be creepy if it left my device, generate push notification reminders for stuff I forgot to put in the calendar, or track my location and tell me I'm going to the wrong airport. Suggest birthday gifts for my friends and family, idk.

Edit: And add strong controls to limit what it can and cannot access, especially for the creepy stuff.

They're stuck on the privacy angle, because what it means is you can't call remote services. You'll always have access to more resources at a data-center than a phone. So, while the frontier of what's possible with purely-local models will keep advancing, it'll never exceed what's possible with remote models.

People care about extra privacy when the delta in capability is minimal. But people won't allow a massive discrepancy, like the difference between a 8B model and a 700B model.

> In the same way a browser can give the AI the accessibility DOM, Apple could give AIs an easier representation to read and manipulate.

Apps already have such an accessibility tree; it's used for VoiceOver and you can use it to write UI unit tests. (If you haven't tested your own app with VoiceOver, you should.)

I have used it actually! It's been years so the fact that it's an accessibility tree just like in a browser didn't come to mind immediately. Both Mac and Windows have such representations for native apps. The actual functionality apps and the accessibility clients support is something like a two-way negotiation. A lot of stuff that should be supported in apps, in theory, is not, just because no client supports it, etc.

> iPhones already have specialized hardware for AI acceleration.

This really is the problem. Why do I spend hundreds of dollars more for specialized hardware that’s better than last years specialized hardware if all the AI features are going to be an API call to chatGPT? I am pretty sure I don’t need all of that hardware to watch YouTube videos or scroll Instagram/web, which is what 95% of the users do.

"Do you want me to use ChatGPT to answer that?"

I think on-device AI will show up more front and center but in a few more years.

A big issue to solve is battery life. Right now there's already a lot that goes on at night while the user sleeps with their phone plugged in. This helps to preserve battery life because you can run intensive tasks while hooked up to a power source.

If apps are doing a lot of AI stuff in the course of regular interaction, that could drain the battery fairly quickly.

Amazingly, I think the memory footprint of the phones will also need to get quite a bit larger to really support the big uses cases and workflows. (I do feel somewhat crazy that it is already possible to purchase an iPhone with 1TB of storage and 8GB of RAM).

2TB microsdxc cards have been available for a year or so, and 1TB cards have been available for several years and are even quite affordable. They work in many Android phones including my cheap Motorola. So it's Apple's sky-high premiums that has made their 1TB phones surprising.

https://www.bhphotovideo.com/c/product/1868375-REG/sandisk_s... 2TB $185

https://www.bhphotovideo.com/c/product/1692704-REG/sandisk_s... 1TB $90

https://www.bhphotovideo.com/c/product/1712751-REG/sandisk_s... 512GB $40

> Apple needs on-device AI to do chores for me with the apps I have installed

Nevermind that—iOS just needs to reliably be able to play the song I’m telling it to without complaining “sorry, something went wrong with the connection…”

Honeslty I don't think I've ever had this happen, apart for when im in a tunnel on a train without service and streaming.

Apple is generally anti market hype. It is a smart PR move to avoid mentioning AI after the Apple Intelligence fiasco, their researchers leaving, and the bubble sentiment at the moment.

You are missing the point. Why was Apple Intelligence a fiasco? Because they failed to understand what users like GP wanted.

It failed to deliver on its promises, investors sued them from overstating AI capabilities.

IMO, it was the researcher team's fault, good riddance.

no, I placed this squarely on apple‘s shoulders. There are real use cases for new AI tools that are actually useful. Use cases that Apple is already invested into — Siri, text to speech & vice versa, etc. many of these have open source models that they could very easily be integrating into their product, even if they didn’t have a partnership with the premier AI research lab.

Instead, we got, what? An automated memeoji maker? Holy hell they dropped the ball on this.

It's not a smart move to avoid integrating the most important capability advance in computing in the past decade - LLMs. They do support small cases, like summarizing text. But there's scope to do more.

They've being doing some research on this: https://machinelearning.apple.com/research/ferretui-mobile

I didn't see this, thank you! They have a follow-up as well:

https://machinelearning.apple.com/research/ferret-ui-2

This is all possible, but an absolutely terrible idea from a security point of view, while prompt injection attacks are still a thing, and there's little evidence they will stop being a thing soon.

We can work toward closing security gaps with new technology, yes. It is necessary for large-scale adoption of LLM tech.

This is a non-story. This was a hardware event. Apple is releasing many new AI features as part of iOS 26 which will launch along side the new iPhones. AI is software. And yet, a number of the features are clearly powered by AI models such as camera enhancements, health monitoring and live translation. Also GPU performance continues to increase in the A19, with CPU remaining presumably fairly flat since no numbers were given, so that’s a win for on-device inference.

If Apple had an insanely great AI feature that truly differentiated itself from their competition, we all know they'd take a lot of time focusing on how their hardware enabled or enhanced that functionality.

The expectation is that Apple will eventually launch a revolutionary new product, service or feature based around AI. This is the company that envisioned the Knowledge Navigator in the 80s after all. The story is simply that it hasn't happened yet. That doesn't make it a non-story, simply an obvious one.

> This was a hardware event.

So was last year’s, technically, but that didn’t stop apple from making it all about AI.

I think this is the correct approach on a phone. I don't want AI front-and-center. I want it in the background quietly making everything better. To me, that's a much more useful form of AI.

I have a similar philosophy about home automation. Every few years I geek out and set up a bunch of crap and spend a bunch of money and waste a lot of time, and then it tends to fall apart fairly quickly and I repent of everything...

...except for the motion-activated lighting in our foyer and laundry room. $15, 15 minutes to install, no additional charges, no external services, no security issues, and just works year after year with 100% reliability.

"Any sufficiently advanced home automation is indistinguishable from a haunting."

And I want it turned the fuck off, quietly not doing anything with my personal shit.

I want to reach for my tools when I want to use them.

I'll argue that face recognition, event detection and share recommendations are nice features.

They are all done locally on your device for the last decade, at least.

[flagged]

that stuff is also essentially machine learning, just more parameters and better marketing

I don't think it's marketing's primary purpose to be off-putting?

Anyway, it's not the same thing: I'm fine with machine learning to give me better image search results, I'm not fine with machine learning to generate "art" or machine learning to generate text. Everyone has collectively agreed to call the latter "AI" rather than machine learning, so the term is a useful distinction.

it's a misleading distinction that's causing people to spiral out and think they're talking to an actual intelligence and is also being used to bamboozle lawmakers into allowing massive amounts of content theft

You're letting the hype men set the goalposts for you, then, as every ML thing has been retroactively rebranded as "AI."

Remember the term "smart" as applied to any device or software mode that made ~any assumptions beyond "stay on while trigger is held"? "AI" is the new "smart." Even expert systems, decision trees, and fulltext search are "AI" now.

> You're letting the hype men set the goalposts for you, then

Not really, I'm taking the hint. If they call a feature "AI", there's a 99% chance it's empty hype. If they call a feature "machine learning", there may be something useful in there.

Notice how Apple, in this event even, uses the term "machine learning" for some features (like some of their image processing stuff) and "AI" for other features. Their usage of the terms more or less matches my line of features I want and features I don't want.

Well, yeah, Apple is being reasonable now because Apple just got through a big bad PR thing with their recent failed attempt at "AI". Apple are currently trying, as much as possible, to avoid applying the term "AI" to anything.

But that's not true of any other actor in the market. Everyone else — but especially venture-backed companies trying to get/retain investor interest — are still trying to find a justification for calling every single thing they're selling "AI".

(And it's also not even true of Apple themselves as recently as six months ago. They were approaching their marketing this way too, right up until their whole "AI" team crashed and burned.)

Apple-of-H2-2025 is literally the only company your heuristic will actually spit out any useful information for. For everything else, you'll just end up with 100% false positives.

It's generally perfectly safe to ignore products whose main purpose is to ride some hype wave.

Who said anything about the product's purpose? Excessive use of the term "AI" is a decision of marketing departments, and happens entirely downstream of product design.

The same product could be produced five years ago or today, and the one produced five years ago would not be described as having "AI features", while the one produced today would.

(You can check for yourself: look at the online product listing for any mature "smart" device that got a new rev in the last three years. The Clapper would be described as an "AI" device today.)

> (You can check for yourself: look at the online product listing for any mature "smart" device that got a new rev in the last three years. The Clapper would be described as an "AI" device today.)

And I make an effort to avoid those with "AI" features where practical. I do not need an AI toothbrush.

Semantically, AI has always been a superset of ML. So it's always been correct to call machine learning AI.

All machine learning is AI, not all AI is machine learning.

Yet there's a very clear distinction between when companies use the term "AI" and when they use "machine learning".

Not to be too, um, dismissive, but one of the things we discussed in my 300 level class called _Artificial Intelligence_ in college 2 decades ago was regular expressions, so, that ship has sailed far over the horizon.

Yeah, the annonying movies IPhone does from my photo library is something I'd love to opt out of. I get that some people love this feature but I don't. And as you say, I could ask the phone to do that for me on request.

Disable “show featured content” in Photos.

I love this feature on both Photos and Google Photos

I think it is just generally the correct approach. It was not that long ago that ML (outside of research) was often talked about but it being ML was not the focus. The focus was on the actual benefits and what it did.

Of course this is going to be spun and turned into a negative, but I basically want ML to be invisible again. The benefits being clear, but the underlying tech no longer mattering.

Actually if Siri ever worked when I really need it, they could finally catch up with their promises made in 2011.

Right? Siri lets me down on a regular basis - and has for years.

Apple has so hardwired my neurons to expect "It just won't work" when it comes to Siri that if I cannot physically reach my device, I just don't bother anymore.

Indeed, a good AI Snitch runs in background on your phone. The era of intelligent spyware.

They did talk about ML in their image processing software for their cameras, though. And I don’t think they mentioned AI when talking about their babelfish AirPods, but it’s clearly there, as well.

They talk about "Apple Intelligence models" running on the phone. They dropped that name for their AI stuff in a few other spots, too.

Yeah. I think they talked about ML for the Watch workout feature as well (or was it the AirPods?) And I think for other features as well but I am not certain and I am not rewatching the whole thing.

They’ve been putting AI in a lot of places over the years.

That's because people don't trust AI that much... and most people don't differentiate between a really cool feature and AI so it's more confusing marketing than anything.

I think it's more because they're way behind on AI and they literally have nothing to actually tell us.

They’re not way behind on AI, they’re way behind on LLMs. Several features they presented depend heavily on AI.

whatever Siri is, they're way behind on it. Siri has gotten siriously worse, Siri has fallen behind Siri.

(of course can't speak to this new release, obviously)

No objection for me. I haven't used it since 2016 or so. I did not try it on the beta.

That’s not true. There were multiple AI features presented, just very well-known ones like the babel fish AirPods Pro. Nothing fancy, but it’s still AI and it’s useful.

As an Apple user, what are they behind on? What features are other makers shipping that Apple is missing?

(Genuinely curious, perhaps there are third-party apps I can use to bridge the gap.)

You need to spend a week to disable all AI features in Pixel. You need to spend just a day to disable all AI features in iPhone.

Apple fans would be going crazy with excitement if Apple unveiled Gemini 2.5 Flash Image (AKA "nano banana"). It's the most powerful image editing tool since Photoshop 1.0.

Pardon my ignorance, but is this part of Android or a standalone app?

Is it something that is not usable on Apple devices?

It powers part of the Google Photos app on Android. It can be used on Apple devices through Gemini, but it is not well known yet.

Probably because Google marketing is the worst. Wtf is “nano-banana.” This is straight out of a Silicon Valley HBO skit.

I can almost excuse them for not hyping this too broadly because of how disruptive to certain jobs it will be.

in some ways (hardware) they're ahead, their dominance in mobile chip power lets them do a lot more machine learning on-device

also hard to find a better laptop for running an LLM locally too

How are they behind, in terms of things normal people actually use? Apps like ChatGPT work just fine, and I have literally never heard an Android user I know personally talk about something they can do which isn’t available on iOS.

> That's because people don't trust AI that much

ChatGPT being the number one app is a weird way for people to express they don't trust AI: https://apps.apple.com/us/charts/iphone

And with Gemini being the 2nd.

Number one free app. Generative AI doesn’t sell phones.

Actually people trust AI too much!

If you've read the Steve Jibs biography, you’d know his obsession with creating the future products.

I feel like he’d be obsessively working to combine AI, robotics and battery technology into the classic sci fi android.

Instead, modern Apple seems to be innovating essentially nothing unless you count the VR thing and the rumors of an Apple car, which sounds to me much like the Apple Newton.

I'd disagree: his obsession was with watching everyone fail to create the products of the future, learning from their mistakes, and only then committing to a product strategy (e.g: iPod, iPhone, iPad).

Good.

I feel like they don’t want to get burned the way they did with the AVR/metaverse thing. Those headsets are worthless. Probably a lot of ptsd going on with people thinking how Jobs would have had their ass over it. I suspect they’re being very cautious with “AI”.

How did they get burned with AVP?

Late to a reply but the tech went nowhere. I’m not a financial expert but I don’t think they could have come close to breaking even on that product. I know no one that has it and there hasn’t been an article here talking about it in years (guessing here but I don’t recall seeing one).

Maybe some day Apple will realize they need to fire the San Francisco consultant they turned Siri into - offended at any request and absolutely useless to provide you a meaningful answer.

they just casually mentioned they have always been on the forefront of AI.... however that might have been a poor choice of words. If they would have just swapped it with "Machine Learning" and then referring to the ML chip and on device learning & analysis. A Win.

just using "AI" as term... they are so on the forefront that they sent your data to ChatGPT, otherwise you would be too ahead of the pack...

I have a Samsung A55 but I shudder to think that soon the AI crap will trickle down to midrange devices. Samsung is "all in" on AI.

This is what I like about Apple. They just use technology judiciously without making a big deal about it, and talk up the product instead. As it should be.

Let me know when Siri can do even the most basic things using natural language. It can't even properly answer things like "hey siri, is apple juice already on my shopping list" or answer most questions that require a web search. Summarize it for crying out loud, I asked you because I'm not right next to my screen to click search results.

THAT would make me take an upgrade. Until then, I'm just keeping this phone until it goes out of support.

You’re absolutely correct!

Good.

Reply on news.ycombinator.com