The iPhone's Last Stand?
https://stratechery.com/2026/the-iphones-last-stand/Of all the things they could build, why must they pick this future...
Certainly the copilot button in ms paint did nothing to attract the clawbot ecosystem to windows
They have no ground to make up on AI, and changing their operating system to center on AI would piss off every iPhone user I know outside of tech, and probably half of them within tech.
[1] https://www.macrumors.com/2026/05/13/apple-q1-market-share-g...
Apple doesn't leak much but there has been coverage of this:
https://spyglass.org/apple-ai-fail/ (April 2025)
https://www.theinformation.com/articles/apple-fumbled-siris-... (paywalled)
https://appleinsider.com/articles/26/06/07/one-fateful-meeti... (2 days ago)
things there don't seem to be going well
From what I can tell, this means any app that is using their foundation model is sharing against a user's pooled usage quota & Apple takes all the revenue upside if a user chooses to upgrade their plan.
Why would an app developer choose their model which has a minuscule 32K context window, might get throttled because of usage in another app, and doesn't share revenue over any frontier model vendor (where you can package/pass along token costs to your customers)?
Same reason developers continue to use Apple payments even when they have to shell out 30% of the revenue.
I can see Apple, setting App store rules around declaring AI usage, or could start labeling apps not using their models with strong language designed to amplify the increasing user concerns around AI and so on.
The product strategy has to be better the product itself does not have to be objectively better for the developer for them to have to choose it.
Its definitely the case that Apple said that some AI capabilities will rely on cloud hosted models; and that some of these capabilities will be metered; and that "some paid iCloud+ plans" include increased usage. Its unclear what the full extent of capabilities that rely on cloud models are. We know that Spatial Reframing is a capability that relies on cloud models and will be metered; this was stated on-stage. The typical conversations with Siri are, supposedly, a mix, possibly based on how complex the query is? We know there are two cloud-hosted models (what they call internally "Cloud" and "Cloud Pro"); are both of these metered? Does it depend on the use-case (e.g. is Spatial Reframing metered because of its use-case or because of which model it uses?)
A lot of this would be easier to figure out if there was, somewhere, a "meter" you could see drop every time you used a metered feature; but there isn't. There is zero indication of metered AI usage anywhere in the settings that I have found (I am on the 27 dev beta). There is also zero indication that any of the iCloud+ plans include metered AI usage. It is also of note that Apple very specifically said on-stage that "some iCloud+ plans" include increased usage; implying that the $2.99/mo plan likely does not.
Apple has also announced that developers under a certain revenue number can use their foundation models in private cloud compute for free [1]. This implies: third party developers will pay for it. In other words: it might be the case that your shared iCloud+ pool of AI usage only applies to AI capabilities from first-party Apple apps, while in third-party apps the developers are expected to front the bill. Which further reinforces your point: Why would anyone choose to integrate with these cloud models?
Strange situation, and unlike Apple; they seem very frantic.
[1] https://techcrunch.com/2026/06/08/apple-bets-cheaper-ai-will...
While I don't necessarily disagree with their vision but if implemented like "Copilot for Windows" I don't see me or anyone wanting to go anywhere near it.
Apple being slow is just fine, at least they didn't launch "Copilot for Mac".
Sometimes the lack of certain feature is the feature.
Is this not what Apple Intelligence is?
Sometimes it's a curse. Apple might be a 2-3 trillion dollar business right now, if they didn't refuse to sign CUDA drivers for their ARM servers.
you will be surrounded by an ecosystem of
devices, none of which stand alone, but are
more like portals to interact with your agents
And then a minute later I could ask it "Is copilot done yet?" and it replies "No, looks like it is still working on it". And then a minute later I ask again. It replies "Yes, it finished. It changed chart.py and styles.css. Do you want me to tell you what specific changes it made to the files?".
But it looks like none of the chat agents with voice interface have such a connector at the moment? An SSH connector would be the most useful. But a "GitHub Codespace connector" or something like that would also do.
I wonder if that will be a missing piece for long. If so, I would build an agent with voice mode and ssh connector myself. But I guess it should come out from the big guys any moment now?
A verbal diff sounds practically useless. Does it first read out the entire left-hand base, and then read out the entire right-hand target? Does it say loudly "REMOVING ... ADDING ... "? How would it read out something like Struct->Field? This seems lower fidelity than a visual confirmation, and I just don't think that voice commands make sense with this kind of work.
I think agents are scary and complicated and dangerous enough that it is genuinely scary to give an agent an instruction like go buy this ticket. It’s okay and apple can easily simplify and eventually win. The mainstream hasn’t really started using agents yet and no one has come close to delivering a platform that will get them there.
These ones also seem really weird because the baseline is most often someone using the iOS app to do the same thing, and the agent demos are usually slower in addition to being riskier. One of the Chrome demos had someone buying groceries at pretty hefty markup, which seemed to be targeting a narrow demographic of people who a) don’t worry about paying 50% more for produce and b) can spend time writing a prompt but not 30 second opening an app and just doing it with zero chance of getting scammed.
I don't expect to go for that, but other people might. Especially if AI stuff continues to improve.
I would actually be willing to use AI to purchase groceries if it could provide me with some assurance that it would choose the items better than a shopping cart.
As for now, I'm only willing to purchase non-perishable goods that are difficult to screw up online.
Might be a service idea for AI.
Once again, early 1990's General Magic looks prescient.
They were working on smartphones with agents capable of completing remote transactions before we had wireless data networks.
> General Magic: The Greatest Tech Company You’ve Never Heard Of
https://m.youtube.com/watch?v=tuFl4WEXBrk
> allowing end-user equipment with limited capabilities to upload Telescript programs to servers to allow them to take advantage of the server's capabilities. Telescript could even migrate a running program... transfer it to another Telescript engine (on a device or a server) to continue execution, and finally return to the originating client or server device to deliver its output.
https://www.wikipedia.org/wiki/Telescript_(programming_langu...
I came here to talk about this, like some other commenters did, too :) I think that this _is_ a predominant view amongst most of Silicon Valley but I think it's kind of a local maxima view... Easy to agree with, easy to see that it's a functional idea, but... people... (i.e. consumers) do lots more than just waste time on their phones even though I bet that's a huge amount of what people are doing across the US right now.
I guess the thing that _is_ true about this nugget is the "at scale" part. It's hard to find things _at scale_ that people would pay for on a phone. So the phone sort of falls back into this easy to monetize thing via advertising. But I think people (qua consumers) probably can clearly be a sustainable market for way more than attention harvesting (or dopamine fracking!) but it requires a lot more effort to think of things that you can build a market out of there. So people sort of lazy-back into attention harvesting via ads.
You might re-title the article instead, "The iPhone holds its ground", and it would be a more realistic title. But perhaps garnering less clicks.
I've always thought Ben Thompson is strong on enterprise and b2b topics but super weak on everything consumer related, he simply doesn't seem to understand consumer behavior (he has zero empathy or ability to project his mind into the average person's mind)
E.g. Ben was sure iPhone air would be a massive hit because he himself loved it. (It's struggled as people don't like the smaller battery life).
Ben was sure the Vision Pro would be a huge hit because he himself loved it. (It was a total failure as the average person doesnt want to pay huge amounts for a ridiculous looking dork helmet).
Ben raving about Meta's hand controller which he was sure was going to be the future of consumer electronics (The Neural Band). He was discussing how you could use it while your hand is in your jeans/pants pocket. Not quite thinking about how this would look while you're sat on the subway with someone sat opposite you.
Ben discussing how the future of watching sports is in VR. Not considering how weird it would be to go to a friends house to watch the game and everyone has their own VR headset. Also not considering the fun of watching sports is doing it with other people.
Basically, he has a huge issue with extracting his own liking of techy products to the average consumer who are basically nothing like Ben Thompson.
I know people are desperate for a Siri that works. The convince of just talking to your phone is priceless. If Apple gets this right, this is a huge deal – which it seems they are on the right track.
People are still talking to Siri for basic stuff like timers and alarms because it works, doesn't need an app, works when phone is locked or even away from you. If this works for more complex tasks like texting and general questions Apple will have the upper hand over Meta and Google in this new way of using computers/internet.
Apple also took a very clever approach for Capex and general AI strategy. Everyone knows that the best intelligence will eventually become a commodity and Apple decided to step aside from this expensive experiment. That's worth pointing out too.
The Air was interesting because everyone I've seen hold it, loves it. But, everyone also loves battery life and the best camera more. The Air is proof of that (similar with the mini lovers).
Just because you like something, doesn't mean it will succeed. These people will more likely using some sort of industry knowledge to form conclusion which conforms with their bias.
On the flip side, just because you hated something doesn't mean it will fail. There are plenty of Apple haters who will write things that seems to make sense but completely misses the mark every single time.
iPadOS also did not receive any product specific updates because I think Apple understands that device well: it’s also a consumption device with a bit more productivity capability. They know they can ship a full macOS on iPad, as witnessed by the lower performance A18 chip in the Neo running the full OS, but what’s the point? Using a desktop UI with a touch interface is terrible. So you’d need a mouse and keyboard. By the time you get that accessory, you’ve already exceeded the cost of a Neo or MacBook Air. There’s also no size, weight or space difference between a fully accessorized iPad and MacBook Neo, Air or 14” Pro.
I think Apple will be fine regardless of whether this new Siri AI stuff actually works well or not. I think deep down they don’t really care because they don’t have to. All of their devices are perfect clients that can interact perfectly fine with cloud inference. And their devices are such a joy to use. That’s what Apple is good at.
Now the confusing part is the new Microsoft hardware project. Is Solara a laptop? Tablet? 2-in-1? Phone? They already have a great hardware run with Surface, so I wonder if this new project is a more powerful local inference push?
For stuff that can't be run on phones, some of it will be run on Apple's servers, which I'm assuming Apple is eating the cost of for the time being.
Stuff that needs heavy reasoning or external knowledge will be processed by google, in exchange for $1 billion a year. However Google already pays Apple $20 billion a year for google to be the default iOS search engine, so you could view this as just changing to google paying $19 billion a year instead.
Some of it is free on-device. Some of it is free & rate limited per day. They mentioned in the WWDC infomercial that users with iCloud+ (the storage tier subscriptions, Apple likes to throw random things in with that) will be able to get more uses per day. And some of it developers will pay for.
And even if the assumption turns out to be wrong, they can just scale down and serve dumber and cheaper models. Shrinkflation is not a novel idea.
- iPhone Air to cram everything into a small space
- Vision pro - a new OS for looking at things and interacting
- Better Siri and AI that works with voice
- Smart local model / routing to big models in the cloud
- integration with wearables (air pods and watches)
People spend several thousand on Lasik so they don't have to wear glasses all the time.
I don't see glasses as the ultimate form factor that everyone uses.
lol no you did not. A whole lot of nothing.
Just not doing it for me. Think I'm gonna stop reading anything he says.
Edit: missing words, thinking faster than typing
He was disappointed in the Apple vision pro for just being an entertainment device (it seems like you two agree there?)
And then the interviews by media of tech should be viewed as an iterated game. He can ask interesting questions for an analyst, but he (and Nilay) do depend on access and that fundamentally constrains what types of questions they can ask if they want continued access
> Just not doing it for me. Think I'm gonna stop reading anything he says.
Pretty sane take tbh
But when it comes to anything around consumer behavior, individuals, etc, i.e. the average family in America, he is often completely and utterly wrong in all his takes and predictions. In fact, so wrong it's often laughable, and amazes me that he is so confident in his predictions.
Also, in the podcast I've noticed that he talks almost every podcast about his "hits", i.e. his times in the past where he predicted something accurately. But never, ever mentions the times where he was completely wrong. He's like the dictionary definition of confirmation bias (or survivorship bias).
It's like he's gotten overly confident (or a little arrogant) as he's become more of a tech celebrity, to the point where he thinks he's some sort of Nostradamus now and doesn't recognize his weaknesses or failures. And I've personally stopped listening to the podcasts as much as it's getting a little tiresome.
BTW, I also noticed how often he is wrong on deep tech topics, e.g. his explanation of IP addresses and routing in one podcast. It's like he thinks his business knowledge + Claude is enough for him to authoritatively discuss how technical systems work, and he often is mistaken...
Yeah might as well cancel your subscription if you’re not gonna read it
The fuck I will.
There is no cloud, only other people's hard drives.
-- your increasingly crusty and curmudgeonly old Unix admin who is not paranoid because they are out to get you
Yes the chassis had to be designed, but that can be used in common for future iterations. That's much harder for phones where the chassis is very tightly coupled to the specific circuit board design.
The 17e already is the cheap iPhone and it's $599. Putting it's internals in a different shell is one thing. Designing and building a half price internal board is quite another, especially as it would either require an entirely new SOC, or mean continuing production of a legacy SOC thus taking up valuable die production pipeline capacity.
Even if they did use an older SOC. Now they'd have to continue supporting that anaemic underpowered SOC with OS updates for years to come, and these future OS updates would have to run well on it.
I don't see it happening.
The case of smartphones is completely different: Android is actually a good OS and there's plenty of excellent devices and high quality brands in the mid range.
With the right set-top-box, you tvs become just another display for whatever you want to send their way. We are extremely close to startrek tng style spaces. I'm sure there is a github repo out there that may even be there.
Why do you think apple has the appletv, the apple watch, the phone itself is essentially a pocketwatch with an addressable display.
I think we have all the parts it's just now about knitting them all together and building up a ux that works for normal people.
I disagree strongly here. The chatbot is the furthest thing from sufficient for the average consumer. Take the newly announced feature that groups your compromised passwords together and offers to agentically change them all for you. Really cool! Could you do that via a chatbot interface? Sure. Would the average consumer? No.
I don't trust microsoft with hardware anymore, nor do i trust them with software.
first paragraph begins the article upon 2 very big and flawed statements:
> Apple fans would, for years and years, sneer at Microsoft’s penchant for talking about products that may or may not ship, deriding them as vaporware.
maybe some would, but as a whole I would say this is not a common thing
> After Apple’s bungled 2024 launch of Apple Intelligence and new Siri, however, vaporware is fair game
no it's not
I didn't know about Project Solara so learned a new thing from the article, but I got the impression that it's not as big as the author tried to make it seem, felt very distant and forced.
> The reason is obvious when you think about it: enterprises are paying for their employees’ time, so of course they are willing to pay for tools that make those employees more productive
Is that why there are billions dollars wasted in useless Microsoft subscriptions and services?
> consumers, on the other hand, are mostly looking to waste time, which is why attention-harvesting advertising is the only software business model that works at scale for consumer services.
What a callous view of people. Who's your benchmark? TikTok addicted kids?
> What they do want to do is watch short-form video
Yeah, it seems so.
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
brother, we are all walking around with a supercomputer in our pocket thats capable of accessing the sum total of human knowledge and yet we're still stuck with people who think the earth is flat.
You could be right, but I don’t think it’s going to be Microsoft that’ll be the leader here.
Saving time (==saving money) is something you can sell to companies. But above all, they are willing to spend on saving their managers time. The higher up the hierarchy, the better. If that involves wasting a lot more time for the underlings, then so be it. The underlings aren't the ones making the purchasing decisions after all
but they will gladly take the productivity hit from that time sink because it gives them teh ability to track employees. they'd rather know that everybody is working at 80% productivity than release that burden and just trust them. it's either this or filling out frustrating timesheets.
It's not a "callous view," it's reality. Social media, entertainment/streaming/media, gaming, and porn make up the vast majority of minutes spent on the internet, and it's not even close.
For what its worth I wish Apple would care more about those of us that want to use AI to actually do work and not these weird contrived examples asking if focaccia can be made gluten free. And I personally couldn't care less what Microsoft does as I'm lucky enough to never have to use their products outside of Github.
I honestly think the use of the word "consumer" is intentionally dehumanizing, a way for corporate figureheads to ignore the humanity of the people they interact with directly or indirectly. This in turn makes them numb to the markets and institutions they are degrading.
Microsoft is still simply one of the very best at enterprise dealmaking.
I wonder if the bigger question is what happens to us.
Convenience is great, but if we optimize away every moment of reflection, tradeoff, and decision-making, we risk becoming passengers in our own lives. The goal shouldn’t be to hand over our judgment to increasingly capable systems. It should be to use those systems to help us think more clearly and act more intentionally.
The future I want isn’t one where AI lives my life for me. It’s one where it helps me live it better.