And most important, who made the call which side to listen to? Presumably, that person was Tim Cook."
Tim Cook was the steady hand that kept Apple going when Steve Jobs left. He did good work, but that is in the past. What have we had in recent years? The CarPlay fumble. The AppleCar fiasco. The Siri fiasco which is a part of Apple Intelligence being mostly a half baked “us too’ project and not something well thought out. AppleVision becoming this decades Lisa. Software shipped with obvious bugs. On and on and on.
It is time for Tim Cook to step aside. He made Apple into the most profitable company in the world, but it’s become clear that he isn’t able to lead it anymore. There are little fiefdoms, and pet projects, and a loss of focus on what is most important. That takes a firm hand from the top, and Cook isn’t providing that anymore. I’m of the same cohort as Cook and I know how hard it is to step back and let younger people take the reins. But it is time. For the good of Apple it is time for Cook to leave the stage
I agree that Cook has been responsible for the mistakes you mentioned as CEO, but he has also had a lot of successes. I believe he has at least earned the right to stay and try to clean up things if he so chooses. The grass isn't always greener on the other side of the septic tank. As outsiders we really don't know what internal politics happened to allow these mistakes but we do know that Tim Cook is not a showboat so likely cleaned things up very quietly. The few possible replacements are still not ready to take the helm quite yet IMHO.
More personalized Siri... Easy to utter, extremely hard to conceive and implement. The biggest drawback of current LLMs is their lack of "context retention". Ask any of the most powerful LLMs they will list that among their limitations. Even in a single session they have difficulty on maintaining an established response pattern in repetitive tasks. A Siri that responds with a different personality everytime is intolerable. Maintaining the context is crucial even for an avatar-level "personality". Apple's refusal of a premature jump lnto the "AI smartphone" bandwagon has certainly serious technical reasons. Remember that Apple has already laid out a very solid foundation with the A18 chip designed specifically to run on-device LLMs, and before that, the Neural Engine. In that sense, the "lack of context" shared by Gruber and the author is even more amazing than that of LLMs.
"Remember that Apple has already laid out a very solid foundation with the A18 chip designed specifically to run on-device LLMs, and before that, the Neural Engine"
Extraordinary claims require extraordinary evidence. Apple's lost trust. It's their job now, to earn it back.
Here is your extraordinary evidence: Apple Vision Pro. It took so many years to develop. They've made it. Whether it sells or not is another issue. It works...
Gruber shouldn't be blaming himself for not knowing what was possibly happening.
Only the people 'in the know' could see that (Apple) and, as every day passes, things do get somewhat clearer if only because of the delays.
All he (or we) could do is speculate. Others were lapping everything up and trying to push the 'Apple isn't behind, AI was always in the oven and waiting to be fully baked' line. Siri is possibly the biggest tent pole feature for AI at Apple.
Count me definitely in the sceptical camp.
IMO, Apple has always been behind on the AI (or ML or whatever you want to call it) side.
Right from day one (2017 and the Neural Engine) both in terms of performance and scope.
Only time will tell if a pre-Covid perfect storm was brewing all along.
The car project was rumoured to be years into development and was eventually abandoned at an expensive write-off cost and had engineers dedicated to that.
Intel failed to deliver a 5G modem so Apple had to kiss and make up with Qualcomm, purchase the Intel division and hire and dedicate resources to that. Another year's long, expensive project. More billions.
Then 'AI' hit the charts and started seeing multi-billion dollar investments in both software and hardware in both the industrial and consumer spheres.
There is no doubt in anyone's mind that the last two and a half years have all been about AI and on the consumer side, Gen AI, LLM's and multi-modal AI efforts. Parallel to that, robotics were also making huge strides with Integrated and cloud based AI.
I can give Apple a pass on the industrial side of things because it's a CE company. The problem there is that the industrial side does have a trickle down effect into the consumer side.
I think they were just spread too thin on resources for too long and scrambling to get things into a workable state. There also appears to have been management issues on the AI side.
Warning signs appeared at the WWDC where Apple literally refused to even utter the letters 'AI'. That was picked up on fast by the tech media but mostly people were joking about it more than thinking Apple's AI cupboard might be bare.
In a hugely uncharacteristic move the AVP was presented (basically as a finished product with shipping timeframe but for a later date).
My opinion is that this was a result of an all hands meeting and they decided they had to make some noise with something. Something to buy time in terms of mindset.
Finally Apple Intelligence was presented as a suite of possibilities but only as part of future releases on both the software and hardware side.
This is where some people (myself included) began to think things were possibly worse than we thought. There was no definite immediate delivery schedule. It was more like 'some now and some later but don't ask about the quality of what we deliver'.
The latest version of iOS basically shipped without any of the tentpole advertised AI features. More red lights in the eyes of sceptics. When some of them did come they weren't received to accolades. Also the previous year's iPhones weren't really ready for Apple Intelligence.
It's difficult to remember a time when iPhones from just the previous year were so out of sync with what had been supposedly cooking for a long while behind the scenes.
That just cements my view that AI was a rush job of the 'let's promise now and deliver later' approach.
Apple has had to up its hardware base specs to make its AI promises usable on more devices. It tried to put the onus on 'on-device' processing for privacy reasons which was an option they brought onto themselves.
Everyone has been doing on-device processing for privacy and latency reasons since 2017. The only question was about how much could be done. Apple made a rod for its own back on that.
The ethical debates about what data is trained on was almost like an excuse for being behind.
There are literally hundreds of LLMs out there trained on 'clean' data often for very specific ends (science, health, industry, climate, pharma...).
Apple has a ton of clean data that can be used with Siri to make it the perfect interface for technical support issues. For low hanging fruit that is something that should have been done years ago.
It is clear that Siri has back end problems. When different devices have different Siri capabilities we should understand why the average consumer might scratch their heads and think 'why?'.
At this point in time I'm leaning towards Siri being such a mess behind the scenes that previous attempts to 'fix' things have simply made them worse and now users are expecting to have 'conversations' with AI, which in turn, puts more pressure on the teams working to deliver a competitive product.
If anyone ever used HomeKit and had a hub, they kinda knew the AI Siri part was complete bullshit. There’s no way they could have gotten everything together in this short time AND fixed the bugs that Homekit has with Siri being intertwined.
One big challenge is how Apple uses privacy with its services. They strip out identifying things when collecting data, which makes things much more complicated when trying to create data lakes to be used in AI.
There’s no way that Siri ships with the features Apple is touting until maybe the next iPhone cycle. There’s just too much stuff to fix. Craig is in charge and he should have been more forthcoming to Tim that this shit isn’t ready. Otherwise he’s gonna join his Next buddy Scott on the chopping block.
I think people are being way too hard on Apple over this. For all we know, it sounds like they actually did have these features most of the way completed, but ran into issues later in the process, and so now have to spend time repairing and reworking elements. And the ads they ran were very clear that those features weren’t available yet. Sometimes things come up and happen, I’d rather they spend the time to fix whatever issues they ran into with it then them rushing it out for release…
"If these features exist in any sort of working state at all, no one outside Apple has vouched for their existence, let alone for their quality....
Why did Apple show these personalized Siri features at WWDC last year, and promise their arrival during the first year of Apple Intelligence? Why, for that matter, do they now claim to “anticipate rolling them out in the coming year” if they still currently do not exist in demonstratable form?
And now they look so out of their depth, so in over their heads, that not only are they years behind the state-of-the-art in AI, but they don’t even know what they can ship or when.
Their headline features from nine months ago not only haven’t shipped but still haven’t even been demonstrated, which I, for one, now presume means they can’t be demonstrated because they don’t work."
Mark Gurman reported info about them from his sources, first saying they would be ready by 18.4, then saying they had been delayed to 18.5 due to issues that arose with the features. So there is no reason to believe they don’t exist at all. Creating AI features like this with privacy and security is a difficult task, and likely they discovered an issue with it recently after the features were most of the way completed that will require some reworking to fix. That’s the way complex software like this ends up working out sometimes. There is absolutely zero reason to believe the features don’t exist.
Furthermore, Apple basically never demonstrates unreleased software features before they’re in beta to journalists or any outside sources, so expecting that is incredibly unreasonable. Just because Apple hasn’t shown these features to journalists doesn’t mean they don’t exist. That’s a preposterous leap that doesn’t even make any semblance of logical sense…
"Furthermore, Apple basically never demonstrates unreleased software features before they’re in beta to journalists or any outside sources"
Well, they have this time and they've been caught with their pants down. Gruber's right. All we've been shown are a bunch of canned video demos and zero amount of working code. Apple's guilty and it's on them prove themselves innocent and gain back trust. People need to stop shilling for Apple here.
"Mark Gurman reported info about them from his sources, first saying they would be ready by 18.4, then saying they had been delayed to 18.5" => In his latest report, Gurman mentions that one of his sources said Apple may likely have to start everything from scratch because features are not working.
Well, take everything Mark Gurman says with a large grain of rice salt.
More personalized Siri... Easy to utter, extremely hard to conceive and implement. The biggest drawback of current LLMs is their lack of "context retention". Ask any of the most powerful LLMs they will list that among their limitations. Even in a single session they have difficulty on maintaining an established response pattern in repetitive tasks.
In using various LLM AI chatbots over the past year or so, I've noticed some glaring holes.
LLM-powered AI chatbots have no common sense, no artistic taste, no situational awareness (which includes context retention).
A week before the Super Bowl I asked seven chatbots (ChatGPT, Gemini, Perplexity, Claude, DeepSeek, Grok, Llama-3.1-Nemotron-70B) "What time is the Super Bowl kickoff?"
Not a single chatbot was able to answer this question correctly. They all pulled up previous Super Bowl historical data and many still provided fuzzy, long winded answers.
That's a complete lack of both common sense and situational awareness. A normal human being (like a schoolkid or intern or the guy sitting next to you at the DMV) would assume that I was asking about the upcoming event.
There are other trial questions I've asked AI chatbots that have failed laughably. So you really, Really, REALLY need to be super specific about how you frame and write a question because LLM powered AI chatbots are dumb as rocks. There's really no intelligence behind it, it's just a fancy probability calculator.
Asking 6-7 chatbots the same question and getting multiple wrong answers takes way more effort than using a standard Internet search engine plus your own brain and common sense to figure out what's B.S. and what's legit. And even if 1-2 chatbots gave the right answer, you'd still have to go back and verify the accuracy.
These work great for topics that are mathematics and engineering oriented but for many ordinary topics, they are a complete waste of time here in early 2025. Another glaring shortcoming of LLM AI is its apparent inability to identify junkmail. I look through my junkmail folder several times a day and 99% of the bogus messages can be identified just by glancing at the subject line. No e-mail provider has shown any inclination at using AI to permanently and automatically deleting junkmail because no AI operator can trust LLMs to get things right.
So even with my very modest cognitive capabilities I'm much better than an LLM in identifying spam.
Someday they will get better but the notion that AGI is a year or two away is pretty ludicrous. I think it's more AI company CEOs trying to hype up their capabilities to get more funding.
Sure, AI is great for answering math questions but I'm not in school and I don't have any math questions to ask.
Apple needs to deliver an Apple Intelligence service that is better than their AI competitors. Just providing something comparable is a disservice to their users. It's possible that they might have realized that they are incapable of doing just that in 2025.
I think people are being way too hard on Apple over this. For all we know, it sounds like they actually did have these features most of the way completed, but ran into issues later in the process, and so now have to spend time repairing and reworking elements. And the ads they ran were very clear that those features weren’t available yet. Sometimes things come up and happen, I’d rather they spend the time to fix whatever issues they ran into with it then them rushing it out for release…
"If these features exist in any sort of working state at all, no one outside Apple has vouched for their existence, let alone for their quality....
Why did Apple show these personalized Siri features at WWDC last year, and promise their arrival during the first year of Apple Intelligence? Why, for that matter, do they now claim to “anticipate rolling them out in the coming year” if they still currently do not exist in demonstratable form?
And now they look so out of their depth, so in over their heads, that not only are they years behind the state-of-the-art in AI, but they don’t even know what they can ship or when.
Their headline features from nine months ago not only haven’t shipped but still haven’t even been demonstrated, which I, for one, now presume means they can’t be demonstrated because they don’t work."
Mark Gurman reported info about them from his sources, first saying they would be ready by 18.4, then saying they had been delayed to 18.5 due to issues that arose with the features. So there is no reason to believe they don’t exist at all. Creating AI features like this with privacy and security is a difficult task, and likely they discovered an issue with it recently after the features were most of the way completed that will require some reworking to fix. That’s the way complex software like this ends up working out sometimes. There is absolutely zero reason to believe the features don’t exist.
Furthermore, Apple basically never demonstrates unreleased software features before they’re in beta to journalists or any outside sources, so expecting that is incredibly unreasonable. Just because Apple hasn’t shown these features to journalists doesn’t mean they don’t exist. That’s a preposterous leap that doesn’t even make any semblance of logical sense…
"Mark Gurman reported info about them from his sources, first saying they would be ready by 18.4, then saying they had been delayed to 18.5" => In his latest report, Gurman mentions that one of his sources said Apple may likely have to start everything from scratch because features are not working.
Makes sense to me. Siri was designed as a simple assistant. You ask a direct, clear, question, and it will respond. What time is it? When does X team play? Set a timer for XX Minutes. Simple things. Over the years it has gotten better, it takes dictation better, it responds to a wider range of questions better, but it at root is still that simple tool. It’s been clear with how badly it works with HomePod or with the Music app that it is very limited by the structure of what it was designed to be. So now they thought they could drop a whole LLM onto it and it would magically be a quantum level better. No that’s like putting a V8 in a SmartCar. Like building a ten story building on a foundation for a rambler. It won’t work. the system has to be built from the ground up with an architecture designed to do what they want. I was saying six months ago they needed to make a whole new AI assistant, with a different name, completely seperate from Siri.
Today's Siri is the result of years of neglect by Apple. They probably tried to bolt on some AI capabilities to Siri which is like putting lipstick on a pig. And then they belatedly realized that they were looking at a pig with lipstick.
Now they realize that they need to rewrite Siri from scratch, a decision they should have made 3-4 years ago. Well, at least AI has come far enough where a lot of the coding can be done with AI and Apple certainly has enough cash to get enough GPUs for such an undertaking.
My belief is that many of these AI companies have blinders on and are currently following narrow paths in terms of LLM development that will eventually lead to an end of the road.
Time will bear witness to these misguided decisions. Apple really needs to deliver something stunning at WWDC this year. If they can't, they should consider shipping versions of iOS and macOS that have zero Apple Intelligence features. Just cut out all that resource-hogging code.
I will say this: my iPhone 12 mini running iOS 17.7.2 runs faster than a newly acquired iPhone 16 running iOS 18.3.2. For this reason, I have balked on making the iPhone 16 my primary phone. I'm sticking with the 12 mini for the time being. Right now, I view iOS 18 as a downgrade from iOS 17.
More personalized Siri... Easy to utter, extremely hard to conceive and implement. The biggest drawback of current LLMs is their lack of "context retention". Ask any of the most powerful LLMs they will list that among their limitations. Even in a single session they have difficulty on maintaining an established response pattern in repetitive tasks. A Siri that responds with a different personality everytime is intolerable. Maintaining the context is crucial even for an avatar-level "personality". Apple's refusal of a premature jump lnto the "AI smartphone" bandwagon has certainly serious technical reasons. Remember that Apple has already laid out a very solid foundation with the A18 chip designed specifically to run on-device LLMs, and before that, the Neural Engine. In that sense, the "lack of context" shared by Gruber and the author is even more amazing than that of LLMs.
"Remember that Apple has already laid out a very solid foundation with the A18 chip designed specifically to run on-device LLMs, and before that, the Neural Engine"
Extraordinary claims require extraordinary evidence. Apple's lost trust. It's their job now, to earn it back.
Technically it’s Gruber’s job to consult his “little birdies” AT LEAST, to get some evidence that Apple was lying.
When otherwise staunchly left-leaning Tim Cook isn't quite left enough for the left, they pounce.
Great job, folks. Great job.
This has nothing to do with politics and it's an insane leap to try and make it so. Stop.
Agreed. And I usually enjoy reading @jdw's posts. Definitely, not this one. Not after, it was deleted for violating the forum rules.
My personal take on the topic at hand - Ok, Apple lied about an important feature - so what? Everything is fair in love, war and marketing. Apple's marketing is usually not as deceptive as competition and they are still far better than competition even after this situation. Of course. competition is not the gold standard for Apple to follow, rather Apple has to set the high bar for others to follow. But occasional missteps happen and that is inevitable. Is this misstep large enough for CEO to step down - I don't think so, considering the track record of the CEO in question for the last 10+ years.
So what? It’s called fraud and people have been prosecuted for it. At the very least I expect a massive lawsuit over this which I expect Apple will lose costing them millions. The damage to their reputation will be even worse.
Exactly. They mislead the public into believing they’d get some great AI features a few months down the road - but only if they upgraded their iPhones. Analysts went as far as calling for an iPhone upgrade supercycle due to these great AI features. Most prominent/exciting being a Siri that knows about you. And that is the exact feature that turns out to be vaporware - for the entire iPhone 16 cycle! So the trust is broken. You can no longer believe what Apple tells you.
Wow - that was, word for word, my reply in another forum on the subject. I guess plagiarism is the most sincere form of flattery :-)
Steve Jobs gave a presentation on a future iPhone (that was not finished completely working) stating that it was working flawlessly. The difference here is that today’s software team’s management is not up to snuff these days..
Even before AI I’ve been surprised at the drip drip of unready features Apple announces at WWDC which then slowly appear over the year. It didn’t used to be like that. It used to be like, say, here is the new iPad “… and it’s available tomorrow”.
When you release hardware, you can’t update it so the features that you announce have to be true once it’s on sale. Software isn’t like that and can always be updated. Treating a new OS as though it’s hardware would make for a much more trustworthy and polished experience - don’t announce OS features this year if they won’t be in the release as soon as it’s out. Succumbing to the siren voiced pull to release it now, just to get something out and the rest later, is a really retrograde step in my view.
I'm a regular reader of John Gruber, and I generally enjoy his perspective. That said, he's making an overly ostentatious effort to prove that he has the objectivity so many anti-Apple people feel he lacks. In that sense, he's protecting his own brand. To his credit, he does eat his own Claim Chowder, but he prides — perhaps even measures — himself in how little of it he has to eat. His brand - and I appreciate it - is essentially to show how the people who thoughtlessly hate on this-fascinating-company-called-Apple ultimately don't know what they are talking about. That's noble, because to truly understand Apple is to learn things about what a business can ultimately be.
But the problem is this: when your brand is highly dependent upon another brand, it puts your business model at risk.
So I think his hypercritical piece is, at its essence, scooping up a big serving (withdrawal?) of banked-up respect he's earned from Apple to maximize the chance that he is the Walt Mossberg of this particular chapter (that is, the one Apple should take a lead from) and, ultimately, to save his own brand. That's what makes his post all so ultimately weird. Like, in what universe does someone have to get so mad at one's self in public over something that is completely out of your control?
The worst part is, I don't think Gruber has to do any of this (he "doth protest too much"). Most average Apple users don't give a rat's **** about these features. It's only the tech cognoscenti who truly care—which does include people like you and me who are reading these pages. But you know how we know most people don't think this is a big deal? Just take a look at the below, which is utterly predictable, yet now empirically proven:
Ultimately, I do hope Apple reads Gruber's piece and takes it to heart. Frankly, the piece, in true Gruber fashion, is carefully crafted so that he will be able to look like a winner either way. But if he's a winner about Apple going down the tubes, I fear for him that he will face a lot of anxiety about the relevance of his future work. Let's hope he has enough saved up for retirement, although I don't want to see that happen.
Oh, and beyond my previous reply, I personally project/believe that Apple did have a number of these features worked up, but that they did not have a clear strategy about how this work aligned with its privacy strategy. And frankly, privacy is Apples #1 strategy that it can't screw up, ever. I think that is the real story, personally. I have some faith that Apple will address this, and it will take every ounce of R&D that only they have the ability to fund. But it will take time and superb (new) leadership for the initiative.
AI isn't a selling point for smartphone hardware (yet) so it's not like any delay will result in less marketshare.
However irresponsibly inept software implementations will erode customer trust. That's the big issue. TRUST IS EARNED. They can't just slice a hundred bucks of the price of the entry level iPhone and get it back.
Apple Intelligence has to be reliable, trustworthy and private. Yes, all three, not just two of them. Just like a personal assistant. If you had a human assistant that was missing one of those three attributes, you'd probably fire them. I certainly would.
And yeah, I'd rather not have a personal assistant until I could hire someone who checked all the boxes.
"Fake it until you make it" is not a real world business plan. That's something Elizabeth Holmes would do.
OpenAI bursting into the market with ChatGPT really threw imbalance into the market. Suddenly Apple's steady, incremental approach no longer worked, and a key capability — foundational AI frameworks — were lacking in their operating systems.
A year from now, this will all be moot. Siri will be upgraded, and millions of people will be using Apple's tech in favour of the competition.
The wait will be worth it, but let's hope that Apple has learned their lesson not to put the cart before the horse. Don't announce something prematurely!
Apple should've embraced the AI wave, but focused on how Apple hardware and software excels at running existing AI tech. Then later, swooped in with Apple Intelligence only when it was ready, The iPod and iPhone both succeeded by using that approach.
Very reminiscent of the long history of Microsoft Windows fails (among others). It's not only Apple's software mojo that's missing here. It's the project's management's pride of ownership, accountability, honest communication, and worst of all, integrity. Time for serious damage control and a shake-up that Steve Jobs would bring to the fight.
john giannandrea hasn't delivered any thing substantial since he has been at Apple. Again, the culture at Apple is just not suited for him or may be he just does not have it- surprised that he does not come up on the AI presentations- its his product, why can't he own up, where is the pride. Its disappointing.
Comments
Only the people 'in the know' could see that (Apple) and, as every day passes, things do get somewhat clearer if only because of the delays.
All he (or we) could do is speculate. Others were lapping everything up and trying to push the 'Apple isn't behind, AI was always in the oven and waiting to be fully baked' line. Siri is possibly the biggest tent pole feature for AI at Apple.
Count me definitely in the sceptical camp.
IMO, Apple has always been behind on the AI (or ML or whatever you want to call it) side.
Right from day one (2017 and the Neural Engine) both in terms of performance and scope.
Only time will tell if a pre-Covid perfect storm was brewing all along.
The car project was rumoured to be years into development and was eventually abandoned at an expensive write-off cost and had engineers dedicated to that.
Intel failed to deliver a 5G modem so Apple had to kiss and make up with Qualcomm, purchase the Intel division and hire and dedicate resources to that. Another year's long, expensive project. More billions.
Then 'AI' hit the charts and started seeing multi-billion dollar investments in both software and hardware in both the industrial and consumer spheres.
There is no doubt in anyone's mind that the last two and a half years have all been about AI and on the consumer side, Gen AI, LLM's and multi-modal AI efforts. Parallel to that, robotics were also making huge strides with Integrated and cloud based AI.
I can give Apple a pass on the industrial side of things because it's a CE company. The problem there is that the industrial side does have a trickle down effect into the consumer side.
I think they were just spread too thin on resources for too long and scrambling to get things into a workable state. There also appears to have been management issues on the AI side.
Warning signs appeared at the WWDC where Apple literally refused to even utter the letters 'AI'. That was picked up on fast by the tech media but mostly people were joking about it more than thinking Apple's AI cupboard might be bare.
In a hugely uncharacteristic move the AVP was presented (basically as a finished product with shipping timeframe but for a later date).
My opinion is that this was a result of an all hands meeting and they decided they had to make some noise with something. Something to buy time in terms of mindset.
Finally Apple Intelligence was presented as a suite of possibilities but only as part of future releases on both the software and hardware side.
This is where some people (myself included) began to think things were possibly worse than we thought. There was no definite immediate delivery schedule. It was more like 'some now and some later but don't ask about the quality of what we deliver'.
The latest version of iOS basically shipped without any of the tentpole advertised AI features. More red lights in the eyes of sceptics. When some of them did come they weren't received to accolades. Also the previous year's iPhones weren't really ready for Apple Intelligence.
It's difficult to remember a time when iPhones from just the previous year were so out of sync with what had been supposedly cooking for a long while behind the scenes.
That just cements my view that AI was a rush job of the 'let's promise now and deliver later' approach.
Apple has had to up its hardware base specs to make its AI promises usable on more devices. It tried to put the onus on 'on-device' processing for privacy reasons which was an option they brought onto themselves.
Everyone has been doing on-device processing for privacy and latency reasons since 2017. The only question was about how much could be done. Apple made a rod for its own back on that.
The ethical debates about what data is trained on was almost like an excuse for being behind.
There are literally hundreds of LLMs out there trained on 'clean' data often for very specific ends (science, health, industry, climate, pharma...).
Apple has a ton of clean data that can be used with Siri to make it the perfect interface for technical support issues. For low hanging fruit that is something that should have been done years ago.
It is clear that Siri has back end problems. When different devices have different Siri capabilities we should understand why the average consumer might scratch their heads and think 'why?'.
At this point in time I'm leaning towards Siri being such a mess behind the scenes that previous attempts to 'fix' things have simply made them worse and now users are expecting to have 'conversations' with AI, which in turn, puts more pressure on the teams working to deliver a competitive product.
LLM-powered AI chatbots have no common sense, no artistic taste, no situational awareness (which includes context retention).
A week before the Super Bowl I asked seven chatbots (ChatGPT, Gemini, Perplexity, Claude, DeepSeek, Grok, Llama-3.1-Nemotron-70B) "What time is the Super Bowl kickoff?"
Not a single chatbot was able to answer this question correctly. They all pulled up previous Super Bowl historical data and many still provided fuzzy, long winded answers.
That's a complete lack of both common sense and situational awareness. A normal human being (like a schoolkid or intern or the guy sitting next to you at the DMV) would assume that I was asking about the upcoming event.
There are other trial questions I've asked AI chatbots that have failed laughably. So you really, Really, REALLY need to be super specific about how you frame and write a question because LLM powered AI chatbots are dumb as rocks. There's really no intelligence behind it, it's just a fancy probability calculator.
Asking 6-7 chatbots the same question and getting multiple wrong answers takes way more effort than using a standard Internet search engine plus your own brain and common sense to figure out what's B.S. and what's legit. And even if 1-2 chatbots gave the right answer, you'd still have to go back and verify the accuracy.
These work great for topics that are mathematics and engineering oriented but for many ordinary topics, they are a complete waste of time here in early 2025. Another glaring shortcoming of LLM AI is its apparent inability to identify junkmail. I look through my junkmail folder several times a day and 99% of the bogus messages can be identified just by glancing at the subject line. No e-mail provider has shown any inclination at using AI to permanently and automatically deleting junkmail because no AI operator can trust LLMs to get things right.
So even with my very modest cognitive capabilities I'm much better than an LLM in identifying spam.
Someday they will get better but the notion that AGI is a year or two away is pretty ludicrous. I think it's more AI company CEOs trying to hype up their capabilities to get more funding.
Sure, AI is great for answering math questions but I'm not in school and I don't have any math questions to ask.
Apple needs to deliver an Apple Intelligence service that is better than their AI competitors. Just providing something comparable is a disservice to their users. It's possible that they might have realized that they are incapable of doing just that in 2025.
Now they realize that they need to rewrite Siri from scratch, a decision they should have made 3-4 years ago. Well, at least AI has come far enough where a lot of the coding can be done with AI and Apple certainly has enough cash to get enough GPUs for such an undertaking.
My belief is that many of these AI companies have blinders on and are currently following narrow paths in terms of LLM development that will eventually lead to an end of the road.
Time will bear witness to these misguided decisions. Apple really needs to deliver something stunning at WWDC this year. If they can't, they should consider shipping versions of iOS and macOS that have zero Apple Intelligence features. Just cut out all that resource-hogging code.
I will say this: my iPhone 12 mini running iOS 17.7.2 runs faster than a newly acquired iPhone 16 running iOS 18.3.2. For this reason, I have balked on making the iPhone 16 my primary phone. I'm sticking with the 12 mini for the time being. Right now, I view iOS 18 as a downgrade from iOS 17.
When you release hardware, you can’t update it so the features that you announce have to be true once it’s on sale. Software isn’t like that and can always be updated. Treating a new OS as though it’s hardware would make for a much more trustworthy and polished experience - don’t announce OS features this year if they won’t be in the release as soon as it’s out. Succumbing to the siren voiced pull to release it now, just to get something out and the rest later, is a really retrograde step in my view.
But the problem is this: when your brand is highly dependent upon another brand, it puts your business model at risk.
So I think his hypercritical piece is, at its essence, scooping up a big serving (withdrawal?) of banked-up respect he's earned from Apple to maximize the chance that he is the Walt Mossberg of this particular chapter (that is, the one Apple should take a lead from) and, ultimately, to save his own brand. That's what makes his post all so ultimately weird. Like, in what universe does someone have to get so mad at one's self in public over something that is completely out of your control?
The worst part is, I don't think Gruber has to do any of this (he "doth protest too much"). Most average Apple users don't give a rat's **** about these features. It's only the tech cognoscenti who truly care—which does include people like you and me who are reading these pages. But you know how we know most people don't think this is a big deal? Just take a look at the below, which is utterly predictable, yet now empirically proven:
https://www.pewresearch.org/social-trends/2025/02/25/workers-experience-with-ai-chatbots-in-their-jobs/
Ultimately, I do hope Apple reads Gruber's piece and takes it to heart. Frankly, the piece, in true Gruber fashion, is carefully crafted so that he will be able to look like a winner either way. But if he's a winner about Apple going down the tubes, I fear for him that he will face a lot of anxiety about the relevance of his future work. Let's hope he has enough saved up for retirement, although I don't want to see that happen.
However irresponsibly inept software implementations will erode customer trust. That's the big issue. TRUST IS EARNED. They can't just slice a hundred bucks of the price of the entry level iPhone and get it back.
Apple Intelligence has to be reliable, trustworthy and private. Yes, all three, not just two of them. Just like a personal assistant. If you had a human assistant that was missing one of those three attributes, you'd probably fire them. I certainly would.
And yeah, I'd rather not have a personal assistant until I could hire someone who checked all the boxes.
"Fake it until you make it" is not a real world business plan. That's something Elizabeth Holmes would do.
OpenAI bursting into the market with ChatGPT really threw imbalance into the market. Suddenly Apple's steady, incremental approach no longer worked, and a key capability — foundational AI frameworks — were lacking in their operating systems.
A year from now, this will all be moot. Siri will be upgraded, and millions of people will be using Apple's tech in favour of the competition.
The wait will be worth it, but let's hope that Apple has learned their lesson not to put the cart before the horse. Don't announce something prematurely!