(…) Apple Intelligence was both the worst and best thing to happen to Apple as far as the AI bubble goes.
It was a mass-radicalization of its users against AI – an attempt to cram a barely-functional series of add-ons that nobody asked for in the most-obtrusive way possible, with summaries that were almost immediately turned into memes and a new Siri that was, somehow, even worse than the old Siri.
And I think that told Apple to pump the brakes. It’s barely spent anything on capex. It’s barely done anything with AI. Despite headline after headline claiming it’s “falling behind,” nobody can really explain what it is it’s falling behind on or why it matters. People hate Apple Intelligence, and I think Apple knows that, and so they’re going to jingle the keys for the markets by putting “AI” on stuff without ever really putting their back into it.
My guess is that Siri sucks because Apple made the calculations years ago of what it would actually cost to make a strong AI product and decided to not participate. They won’t say that publicly because it would require that they publicly admit that Siri sucks and they would never do that.
Siri gets a massive overhaul in ios27. It’s actually useful for understanding your data or general purpose questions now.
How are they pumping the brakes on Apple Intelligence when the next major update for iOS will integrate it into almost every part of the iPhone?
They pumped brakes on their own LLM, they are now going to use Google one.
Yuck
I think letting the others work out the kinks before making big investments makes sense. What is the benefit of being first vs. additional sales? Look at Microsoft with their OS LLM integration. They are backpeddling due to cost and user resistance. Their users can always use other general purpose LLMs.
That would mean that they would be starting when other players are well established in the LLM space.
I think they are waiting, or maybe even working on, the next paradigm in AI. There are a few new approaches, like JEPA, that are a lot less resource intensive, and which are way closer to actual intelligence.
LLMs are at a very advanced state of diminishing returns, and are a bit of a dead end. They are not going to get much better. They need much higher resource investment for minimal gains, and are inherently bad in many use cases and scenarios.
Thanks for the link. Latent embedding is nothing new, need to read up on it a bit to see how it works. True artificial intelligence needs some representation of ideas and concepts… Language provides this and has plenty of training data, but is verbose and ambiguous. Using a compressed/latent representation totally makes sense, but I wonder how it’s trained.
I don’t consider LLMs to be AI, but it’s a step along the way. As you say though, it feels like a dead end.
But not before giving the LLMs access to some of the most personal data stored on people’s devices, betraying their long standing promise of caring about privacy.
Explain this.
From a regular user’s view, my understanding is that my personal data remains on-device and leaves only with my express action or consent.
Your consent is still required, but Apple is making promises about anonymity and privacy they simply can’t guarantee when sending data to a 3rd party server.
Apple is using its years of marketing as a privacy forward company just to put one of the worst technologies for privacy a tap away from uninformed users. (E.g. my parents who will tap ‘OK’ on every popup thrown at them.) After years of headlines of how dangerous consulting LLMs for health concerns can get, Apple launches ChatGPT integration in Apple Health.
Additionally Apple has made promises to be carbon neutral by 2030, impossible if we start to measure the true cost of Apple Intelligence. Unless of course we just blame that on Google or OpenAI. You know, like Foxcon was responsible for the suicide nets and horrendous conditions in the iPhone factories.
For more than a decade Apple has trained its employees on supplier responsibilities, ethical code of conduct and marketed their privacy and environmental causes. None of these values are compatible with OpenAI or Google, who are now running Siri and Apple Intelligence.
I get your concern and I feel the same way. However, you’ve used concerns as fact when even the link you’ve provided counteracts your argument.
Guess I’ll wait to see if Apple’s promises match their previous ones they fulfilled when they used ChatGPT for AI. I’ve still found no evidence to the contrary.
I agree my initial comment was oversimplified, but what you say is not fact but only concern contradicts Apple’s own statements about what data they share with 3rd parties under the Apple Intelligence product.
If putting “OpenAI must process your information solely for the purpose of fulfilling your request and not store your information or any responses it provides unless required under applicable laws. OpenAI also must not use your information to improve or train its models.” on their website is enough of a guarantee for you that OpenAI won’t do what they please, then good luck.
Explain this.
Your notifications were also supposed to remain on-device and not be funneled to the government.
People fall head over heels for something they have no way of verifying but are so proud of repeating marketing talking points.
I didn’t make any claims so I don’t need to explain anything to you.
Are you assuming we’re arguing?
Next OS release (around September) will be the built-in Google Gemini partnership. Apple isnt rolling their own AI, and instead will be outsourcing it to a third party.
They already did this with ChatGPt and were still able to partition my data.
The explanations/writeups I’ve seen for the next OS suggest the same behavior, although I see concerning parts about how much the AI will be in use on your device.
Do you have more information on that? I can’t find anything but conspiracy theories but I’m eager to have actual information to back those accusations.
So far, that’s my understanding as well. That anything that touches personal data remains on device by default and only leaves the device on your explicit commands. Of course I may be missing something, but this feels like Apple eating their cake and having it too, even if the cake isn’t as good.
What accusation? Apple themselves said they are partnering with Google Gemini. Not all AI queries will be on-device, as some tasks will need more power. Itll then be sent to their “Private Cloud Compute”.
The accusation that Apple will give your data to Google et al is one you made. You state that queries will be sent off-device but don’t recognize that this is with consent and exactly how it worked with ChatGPT integration.
It’s a valid fear but not a valid fact.
I havent yet made any claim that Apple is sending your data to Google.
Apple acknowledges these issues per their Private Cloud Compute newsroom release:
Cloud AI security and privacy guarantees are difficult to verify and enforce. If a cloud AI service states that it does not log certain user data, there is generally no way for security researchers to verify this promise — and often no way for the service provider to durably enforce it. For example, a new version of the AI service may introduce additional routine logging that inadvertently logs sensitive user data without any way for a researcher to detect this. Similarly, a perimeter load balancer that terminates TLS may end up logging thousands of user requests wholesale during a troubleshooting session.
It’s difficult to provide runtime transparency for AI in the cloud. Cloud AI services are opaque: providers do not typically specify details of the software stack they are using to run their services, and those details are often considered proprietary. Even if a cloud AI service relied only on open source software, which is inspectable by security researchers, there is no widely deployed way for a user device (or browser) to confirm that the service it’s connecting to is running an unmodified version of the software that it purports to run, or to detect that the software running on the service has changed.
It’s challenging for cloud AI environments to enforce strong limits to privileged access. Cloud AI services are complex and expensive to run at scale, and their runtime performance and other operational metrics are constantly monitored and investigated by site reliability engineers and other administrative staff at the cloud service provider. During outages and other severe incidents, these administrators can generally make use of highly privileged access to the service, such as via SSH and equivalent remote shell interfaces. Though access controls for these privileged, break-glass interfaces may be well-designed, it’s exceptionally difficult to place enforceable limits on them while they’re in active use. For example, a service administrator who is trying to back up data from a live server during an outage could inadvertently copy sensitive user data in the process. More perniciously, criminals such as ransomware operators routinely strive to compromise service administrator credentials precisely to take advantage of privileged access interfaces and make away with user data.
We are still waiting for a full write-up on how Private Cloud Compute can protect against this, but I’m sure Apple will release more information as we get closer to September.
That misrepresents/oversimplifies a complex situation.
Its even messier than that when you start looking at the upcoming CEO transition and the current White House administration taking donations from certain C-suite Apple executives.
The model runs primarily on-device, when the on device model can’t answer a query, it encapsulates the context and sends it to an ephemeral session running in apple’s cloud.
I have concerns about the model but privacy is not one of them.
Your request and attachments like documents, photos, or contents of the document, as well as limited data associated with the request, such as current time zone, country, device type, language, and feature being used when making the request, will be sent to ChatGPT to answer your request and enable ChatGPT to provide you accurate and relevant results.
When using the ChatGPT extension, your IP address is obscured from ChatGPT. Your general location, which is approximated by matching the IP address of your internet connection to a geographic region, is provided to ChatGPT for purposes such as enabling ChatGPT to prevent fraud and comply with applicable law.
https://www.apple.com/legal/privacy/data/en/chatgpt-extension/
Yeah I agree with you there, no one should have used the ChatGPT extension with Siri back then.
Are you talking about the plans for iOS27 when you talk about the encapsulated requests?
I’ll have to read a bit more into that, if anything is available yet
Yes, I was talking about how they described how Apple Intelligence will work at WWDC.
Their posted information is more vague on how it is supposed to work but doesn’t ring any alarm bells.
Your link is for the current OS only, but I found details about iOS27 in the newsroom.
There are a few subtle differences that are probably crucial. Both iOS26 and 27 use Private Cloud Compute. For 26 they specify that it runs on Apple silicon servers (not necessarily owned by Apple) and that no personal data is shared with Apple. For SiriAI in iOS27 it says that data is processed on servers using Private Cloud Compute (without specifically mentioning Apple silicon) and state no data is shared with Apple or others. Which would imply it runs on Google’s servers who are the ‘others’ with potential access.
For me that is too much ambiguity, but I look forward to people more capable than me investigating and testing these claims thoroughly
I think things would look much the same for Apple. I think they will sit on the sidelines and watch everything burn. I could see them doing some choice acquisitions as things begin to collapse, but I could also see them do nothing.
I think Apple is in a very weird place at the moment. The Vision Pro was a dud, but it was also the most interesting and future-forward thing I’ve seen anyone put out in a while. If they were smart, they’d tread water and sink as much money into making that as small as humanely possible – no matter how long it takes – because the entire AI bubble is a result of everybody running out of hypergrowth ideas, mostly because we’re flat out of new interfaces.
I want to be clear that what they want to do with the Vision Pro requires it to basically be weightless and invisible and never need adjustments. When it works, it’s genuinely awesome. But that’s a load-bearing when. One slight movement means the whole thing goes out of focus. I can’t even use mine anymore because it needs an update that requires you to wear the thing the whole time. So much promise, released too early, shoved out the door by a CEO on his way out.
