Apple Intelligence Arrives in macOS 15.1 Sequoia, iOS 18.1, and iPadOS 18.1

I have just one question for now. I have the 2021 iPad Pro M1 11-inch. In Apple Intelligence it shows this after upgrading to 18.1.

image1028×700 59.5 KB

I don’t remember reading about a Waitlist. It’s not a feature automatically available to everybody who upgrades on a proper device?

Yup. Clean Up is available without the waitlist, but you can’t get Writing Tools and such until you’re accepted by Apple.

Apparently they are doing this so that the AI servers aren’t overwhelmed at once or something. Took a few hours for me to get accepted.

Thanks for the excellent summary, @ace .

You know what I find interesting about this? How insipid these features seem to be and how weak they are. You, yourself, in the summary keep talking about things like, well, the transcription was useful but it missed all the proper names, etc. What? How is that a good transcription?

Yes, better than paying someone to transcribe it, but, no, this means we have to have our proofreader’s hat on with floodlights on the text. How is this better? When you start getting into nuance instead of that-Python-function-is-wrong it makes these features seem more like quicksand than stepping stones.

Dave

Thanks. Joined the waitlist…

Apple Intelligence is not available if you use English (Australia) … can’t even access the Photos AI

Well, I use a Mac Mini 2018 and an iPhone SE (champing at the bit a bit for a new MacMini release) so these OS updates should be amusing . I’m well prepared because of Logic Studio which has great features but create a lot of gnashing of cpu when you’re using the lower end of hardware with it. I have read a lot of complaints out there about Apple’s slow approach to AI integration but I applaud it. Other AI platforms like to dump and discover bugs or problems later. Apple’s security emphasis is welcome and a good “first and foremost” stance to take.

The waitlist was something reported by people using the 18.1 beta during the summer after it became available. That said, I was surprised by it, too. I figured it would take a day but the next time I unlocked my phone there was a red badge on the settings app and I was in - maybe 30 minutes later?

I’ve since read a guess that the waitlist is the time it takes Apple’s private cloud compute servers ready for your Apple account.

That makes sense.

The choice of wording here is just unfortunate. After hyping and selling AI since June, and people buying hardware being marketed with that argument even though we all knew we were going to have to wait until the .1 update drops to finally get AI, now telling people there’s something like a “waitlist” as if this were some shady Kickstarter offer, gives the wrong vibe. Just tell people their account is being set up and that they would receive a notification once it’s done and they can use AI.

I did my first test with the writing tools and it worked really well for me. Honestly I likely wouldn’t have done this except to test, but last night while I was using my iPad with hardware keyboard to enter in my journal app I noticed after typing three sentences that I had inadvertently hit the caps lock key and I had typed all-caps.

The iPad Air 4 can’t do AI and it’s still on 18.0.1 anyway, so I selected and copied the text, pasted into Notes on my phone (using continuity - magic), and used the rewriting tools. I tried professional and then concise (which was better), but both removed the all-caps, properly capitalized the starts of the three sentences as well as the pronoun “I” and my wife’s name, and made it easy to copy and paste into the journal app (which I could have done with the iPhone version of the app, but instead I used continuity again to paste it back on the iPad.)

I was pretty impressed.

Yes you MUST set your language to “US English” or no go.

Unrelated to the AI but my Mazda Carplay is messed up now after 18.1 update. I can still play music once phone is connected but default screen in carplay is now apps, not the music app/song. Others on a Mazda list are reporting crashes with maps as well, in carplay.

App crashes, I hope! :-)

I’ve been having frequent disconnections with wired CarPlay in my Mazda since new iPhone and iOS 18.0. It usually happens right after I send a text with Siri. I get an error that says CarPlay is no longer connected. When I unplug and replug the phone back in, CarPlay is back working within 20 seconds.

Very annoying and utterly random. I can go for hours with it working fine, or sometimes it will do the error 2-3 times over a 20-minute drive.

Same braided cable I used with old iPhone 15 Pro Max, so I don’t think it’s the wiring. Something about the new phone/OS is causing issues.

Oops. Yes, not car crashes!

I’m getting really tired of hearing about AI in general and now Apple’s on the bandwagon with theirs. Well, guess what, I won’t be using it on my MacBook Air - not now and not anytime soon. To unleash something like this on the public by any company is way too soon IMHO. We’ve already seen on 60 Minutes and other reports the potential for using AI for harm and for nefarious purposes. Not to mention its current inaccuracies. If Siri is any judge on this technology, it has an awfully long way to go to be truly useful AND safe.

AI is just software algorithms. It is not, despite the hype and scare tactics of the press, actual intelligence. Your phone is not going to suddenly decide that it wants eliminate you as a user and generate fabricated content to blackmail you with, or whatever else the pundits think is right around the corner.

Like all other software, these “AI” applications are not actually intelligent. They do not do anything beyond what they’re programmed to do. And if you’re afraid that their developers have programmed them for nefarious purposes, then you should be equally afraid that those same developers have programmed their non-AI apps for those same purposes.

I’m no more afraid of Google AI doing damage than I am of the rest of Google’s software doing damage. And for the record, I do not trust Google to keep any of my data private, AI or not.

Likewise for Apple. If you don’t trust them when they say what their “AI” software will and will not do, then you shouldn’t trust them with any of their other software - like Safari, Pages and even macOS itself. It could all have hidden code doing things you don’t want.

We use Apple’s software because we believe it is not doing these things, but if you believe it is or might be, then you should be avoiding the entire ecosystem and stick with only software that you or people you trust have audited, which will probably limit you to open source software where such auditing is possible.

I see no reason why algorithms based on neural nets (what everybody mistakenly calls “AI”) changes any of the above.

I don’t think any reasonable person is worried about any of that.

But there is a lot of concern that these tools reinforce and cater to some of the worst human behavior, just like social media before it (which is also “just a bunch of code” and algorithms).

Just watch Apple’s recent AI TV ads (here, here, here, here, and here). The key use case Apple itself boasts about essentially comes down to deception intended for lazy people and egoistical posers who have zero work ethic and care little about their family members and friends. Who wants that kind of behavior enforced? We have more than enough of that going around as is.

What’s missing here is emphasis on how these tools can be used to support good behavior and suppress bad impulses. But none of that is apparently sexy enough or being discussed, hence so much skepticism.

If even Apple can’t come up with better showcases, no wonder folks are pushing back. And that’s before I, as a scientist and journal editor myself, even get into yesterday’s Apple “scientist” in a lab coat having AI write her journal article for her. We actually require authors who publish in our journal sign that they don’t do that. That has become quite common in my field actually. I hope Apple is just being tone deaf here and not actually deluded.

I’d add that the current hype and hoopla about AI is for a specific type of artificial intelligence, generative AI. Other types of AI have been used for years, in uses that have had many benefits for both individuals and organizations.

One example that I appreciate is that when I travel, I no longer need to notify my debit card and credit card issuers that my cards will be used outside of their normal geographic areas. The card issuers’ authorization systems are able to recognize, say, that a purchase of plane tickets followed by a payment for coffee at an airport followed by a rideshare charge at my destination means that a subsequent string of purchases in a previously unvisited city can be approved without any hassles. And speaking of traveling, Google Translate has been vastly improved by the use of AI over the last few years. Another example I value is financial markets using non-generative AI’s to detect insider trading and front running.

I’d also say that non-tech and non-business media interest in AI only began when journalists detected a massive potential threat to their jobs from generative AI’s. The same sort of thing happened with instant messaging. News outlets barely paid attention to IM apps until reporters realized that Twitter could serve both as a (lazy) source for stories and as a way to get content (instantly) published outside of the usual story assigned-edited-published process.

Do these agreements disallow using tools such as Grammarly (and now, Apple Intelligence) for proofreading? These tools use AI to suggest stylistic changes. I use them and often accept their suggestions in whole or in part. [I just took a Grammarly suggestion to rephrase a sentence in this comment.]

Proofreading is fine just like spellchecking was before.

It’s having software write entire sentences or paragraphs for you (and that’s what Apple’s silly video showed) that’s verboten. Well, strictly speaking, you can do that too, but then you have to specify that not you, but somebody else’s generative AI wrote your paper.

I would love it if Apple came up with an automatically updating, country-specific, description of what’s actually available now. I keep reading about iOS 18 features that are available, or should be available soon, or are available in the U.S., but not in Canada, etc.

I’ll read about a feature on a site like TidBITS, and think, oh, interesting - only to find the feature absent on my phone, even though we’re both running the same iOS, supposedly.

Not blaming TidBITS, this is completely on Apple. The lack of clarity on what I can actually do on my phone or computer has me ignoring articles on the new updates, as they usually lead to disappointment these days.

I remember when we all got the same system software, dadgummit! </old man ranting>

It’s gotten even worse with so many rumor sites, Apple pre-announcing features at developer conference (which now makes worldwide news and used to be only of interest to actual programmers), and so many different OSes.

I really like Adam running his “Do you use it polls?” because it reminds of me features I found exciting when announced, but the actual rollout was months later and by then I’d forgotten all about it. Now we’ve got so many features that require regulation/permission from various government agencies (AI, satellite service, hearing aids, etc.) that it’s really getting crazy!

This is pretty close: iOS and iPadOS 26 - Feature Availability - Apple (albeit hard to find). Frustrating, though, that it doesn’t seem to include Apple Intelligence, but maybe because it’s new, or considered beta. (See How to get Apple Intelligence - Apple Support).

Yeah, I feel badly when people are bit by such things, but it is really difficult to find the information reliably (it’s often there, but in very small print at the bottom) and even when I can find it, it’s hard to work into articles in such a way that is readable.

Yes! That’s one of the goals of those polls, and once things settle down a bit, I need to get back to more of them.

Thank you for the suggestion, just looked at it.

In theory, that kind of page might be useful, but as you point out, it isn’t up to date, and doesn’t have a useful “last updated” notation anywhere. IMO, any web page with news or info like this on it should have the date it was last edited, right up at the top.

The page also suffers from the “pick an item, and we’ll tell you if you can have it” issue that I get with clothing store web pages. Those folk don’t understand that I only want to see the items they sell that fit me, so let me filter by size first. Likewise, the Apple page really should have an option to filter by country first. Having to see if Canada is mentioned in each and every entry is backwards.

Yup, got bit by government regulatory thing - Canada hasn’t approved the hearing test/hearing aid functionality yet, so no joy there.

(I actually remember when Apple used to have computer-model-specific versions of the system software… Don’t miss those days )

A brief afternoon test proved to me, once again, that Siri is a sad joke. I have enabled it on my phone only because CarPlay does not work without it.

“Hey Siri” may or may not invoke Siri, depending on whether AirPod/Bluetooth is connected. Double-tapping to invoke Type to Siri never works on the Home screen, sometimes works in apps; it stopped working once I connected to Bluetooth in my car. Siri failed to start a phone call when I was wearing AirPods, nor could it invoke Walkie-Talkie on my watch (but offered to download a third-party walkie-talkie app—but only when I typed the request. Siri has never understood what I speak, even though text dictation has worked fine).

“Hey Siri, why don’t you work?”

“I’m sorry, I don’t understand.”

Obviously.

Any speculation as to why macOS doesn’t allow for a similar max charging setting as iOS now does? I’m not a huge fan of “optimized charging” because I find it cannot anticipate unusual use scenarios. A slider like iOS 18.1 that allows just not charging a MBP beyond 90% would most likely help with battery aging. Maybe in a subsequent update.

Imagine my disappointment when after all the Apple marketing hyperbole about Apple Intelligence and how this is going to change our lives and yada yada, I learn that even with all the AI super sauce in macOS 15.1 Siri is still dumb as a rock.

Me: Open Control Center settings.
Siri: Sorry, I can’t help you with that.

2nd try
Me: Open Settings
… Settings launches…
Me: Now, open Control Center
Siri: It doesn’t look like you have an app named “Control Center”.

I believe that was expected from this release - small improvements, not dramatic ones.

(A story from today, but I read others during the beta that said the same.)

Siri, so far, has been little more than a scripting language. No intelligence, artificial or otherwise. (FWIW, Amazon’s Alexa is exactly the same. I don’t know about Google’s assistant.)

It will be interesting if this new AI work is going to change its ability to understand language, or if it is only going to impact what it does when it executes the resulting commands.

The Economist had an article about this a few months ago, in which they argued that the use of LLMs in science writing might actually be a good thing in some ways. I’d be curious to hear your thoughts, though the article is behind a paywall, and anyway it would probably be off-topic to discuss this more.

How do you know this? Are you certain that bad actors won’t exploit the power of AI for nefarious purposes? I’m not. They’ve exploited everything else. Who’s going to stop them now?

I’d say that if a true AI - an actual artificial life form that is self-aware and intelligent - existed, then we would definitely have to be worried about it. And something akin to Asimov’s three laws of robotics would probably be necessary.

But we are quite a long way from being able to develop any such thing. Which is why I really object to the press using the term “AI” to refer to any algorithm implemented with a neural net.

It seems vanishingly unlikely on the iPhone because of Apple’s best-in-class security. If bad actors could regularly break into the iPhone, we’d see much bigger problems than we do now. As it stands, security vulnerabilities on the iPhone are limited to nation-states and major organized crime, and even then, once they’re used against high-value targets and discovered, Apple shuts them down with security updates.

Why is the iPhone the only possible point of entry? And why is breaking into our devices the only way that AI technology in the hands of bad actors can be used inappropriately? Even without AI technology, we are all vulnerable to intrusive exploits even if we never use the internet because the corporations and institutions we do business with do use the internet and many of them are routinely compromised by hackers eho wish to do them (and us) harm. We are all like people living in crime-ridden neighborhoods who think we’re safe because there are locks in our doors. We’re not safe at all!

It’s not, but it’s what you quoted from @Shamino in your reply. macOS isn’t as secure as iOS and iPadOS, but it’s good. Speculation about a bad actor somehow infiltrating macOS and turning Apple Intelligence to nefarious purposes borders on (dark) fantasy.

It’s not, but this topic is about Apple Intelligence, not AI in general.

This is also unrelated to Apple Intelligence (and massively overstating the case). If you wish to discuss the reality of online security for the everyday Apple user, please start a new topic.

In your opinion. I’m not thinking about Apple Intelligence as much as about the obvious fact that there are legions of very bad actors out there who get off on exploiting every vulnerability they can. To me it’s naive to think that people like this won’t try to use Apple Intelligence against us. Davis Shamino was speaking generally, and so am I. I stand by my comment that we’re living in a crime-ridden neighborhood thinking we’re safe because our doors are locked.

In my informed opinion from 34 years of full-time work exploring, writing about, and analyzing Apple technologies. I’m pushing back on this because encouraging paranoia about unrealistic threats makes it harder for people to take the real threats seriously.

You can imagine any bogeyman you like, but the idea that Apple Intelligence could be undermined for significantly malicious purposes is unrealistic. Even if we assume that all of Apple’s security measures—such as the signed system volume, sandboxing, and notarization—could be compromised, it’s difficult to see how the exposure of Apple Intelligence would pose a serious security or privacy threat because it operates entirely on-device or through Apple’s Private Cloud Compute. At worst, a compromise of Apple Intelligence could:

• Inject unwanted text into notification or message summaries?
• Make Clean Up replace removed objects with dirty pictures?
• Replace photos in Memories movies with something else?

It’s little more than digital graffiti, because that’s all Apple Intelligence can do. Any threat actor with the technical knowledge, skills, and attack vector to breach Apple’s security would be far more interested in exfiltrating confidential information like passwords or financial data, encrypting data for ransomware use, or some similar action that would have political or financial gain.

There are many real Internet threats, but as I said, this is not the topic to discuss them.

Whatever you say, Adam.

Just thought I’d post an update to my comments above. After a new USB-C cable didn’t solve the issue, I talked with dealer about updating the infotainment firmware on my car. They confirmed there was an update available and thought it would fix my problem, but since the car is out of warranty they wanted $170 to do a software update!

I thought that ridiculous and found several videos on YouTube showing how you can do it yourself. There’s some risk of bricking your system as this isn’t a supported update by Mazda (and I had to download the firmware file from some rando’s google drive account), but the great news is that it only took about 30 minutes and worked great! Problem solved, $170 saved.

I’ve never been anti computers in cars, but the mix of tech with the typical auto support/repair system is the worst of both worlds.