Not suffering from FOMO is a delightful thing. You can enjoy life while others scramble after the “greatest” and latest. There are far more important things to worry about than whether you get ML models running on your phone.
I can't really judge at this point, but I really doubt there is hardware limitation for that, e.g. iPhone 15 Pro isn't so much powerful than iPhone 15.
Even when it is - they could utilize their server side AI solution, but for them greed comes first.
Given Apple’s general approach, I suspect it’s on-device and these are the only devices that can currently achieve a good experience. I’d like to see them make the devices work together - having an iPhone, Mac or iPad run the model but access it via a watch or HomePod.
6GB RAM on the iPhone 15, 8GB on the iPhone 15 Pro.
That 2GB really could make the difference: they have a 3B language model, which is going to be memory hungry. There needs to be enough memory left for the rest of the phone and apps to function.
> 6GB RAM on the iPhone 15, 8GB on the iPhone 15 Pro.
> That 2GB really could make the difference: they have a 3B language model, which is going to be memory hungry.
Even when it is - they could utilize their server side AI solution, but for them greed comes first.
What double standard are you talking about? Are you suggesting that you have to pay for recent hardware for local privacy otherwise you use ChatGPT and its not private?
As many have already pointed out to you, its most likely the earlier hardware simply wont run their local model.
The partnership with OpenAI includes an agreement that they will not track/log apple users, and even then it is a fallback option (after apples OWN cloud model) that users must specifically agree to.
You haven't watched or read what they announced then. Almost all queries are processed locally or in the "Private Cloud" on Apple chips in Apple data centers. For queries that are better answered by ChatGPT, the phone will explicitly ask you if you want to forward it to ChatGPT.
"Bring ChatGPT to Siri" is extremely misleading. It's less integrated with Siri than Google is.
Calling people who disagree with you "fanboys" doesn't help elevate the conversation.
That said, whether you think me a fanboy or not, your openAI criticism is wide of the mark. The chatgpt/openai integration is apparently going to be 100% opt-in and you will be prompted for consent every time data leaves your device. So much so that the economist said it risks becoming "ChatGDPR"[1].
Good, I don’t want hallucinating AI eating my battery.
I’ll upgrade when the latest OS no longer installs on my iPhone 11 and even then I might not get the latest and greatest - better to wait for my carrier to offer a previous-generation iPhone for peanuts and use that to upgrade.
One of the most interesting things about Apple's design here is that by heavily restricting the ways in which their on-device LLMs are used they've massively reduced the scope for hallucination.
It's far less likely to cause problems when limited to tasks like summaries, copy editing and short text rewording.
Hallucination is most likely to cause headline-grabbing mistakes when you use LLMs to answer questions about the world or generate larger volumes of text from a short prompt. Apple effectively outsource that to a clearly branded ChatGPT integration, neatly assigning any blame for hallucinations to OpenAI.
100%. It is likely that they will. The current on-device stuff is going to (obviously) be gated to the fastest devices presently.
I imagine they will roll out to other languages and devices relatively soon after. This article assumes too much about the future, and seems to be pushing an “Apple wants to sell you more hardware” narrative as if a hardware company wanting sales is somehow scandalous.
Despite that, I don’t think that is what is happening here. I imagine server models will be available across all/most devices not long after they do the initial launch.
Apple specifically avoids building solutions which put them in the position of being able to do lawful intercepts. I am convinced it is an explicit checkbox in their development process now.
If you’ve built the solution for commercial purposes you obviously don’t have any real moral concerns that can allow you to refuse a lawful request, while “build us a backdoor to unlock this iPhone” or “build us a backdoor to disable advanced data protection mode/e2ee mode” would be compelled speech.
The things they have complied on are fairly minor stuff like making airdrop not work with non-friends for more than 10 minutes at a time.
There are obvious privacy problems with streaming every user interaction to the cloud, which is why they’ve already avoided this for things like photo tagging etc. Maybe you could do anonymized inference similar to the way maps work (let’s say, apple hosts a session for an anonymized user identifier but they don’t know who) but it’s still very intimate information regardless.
On device really is better… not to mention the power concerns. Like I just do not get these objections at all, other than whining because it’s apple, and I think the same people would be whining 100x harder (with good justification) if apple was really steaming every user interaction to the cloud. We’d be hearing about the sham of apple marketing on privacy while sending your sexts to the cloud etc.
same thing for the whining about copilot. Like yeah if it’s on-device it ultimately is going to be a file on disk somewhere (omg SQLite, we hate that!!!). The whole disk is protected by bitlocker, and the file is additionally restricted tightly to the system user (not just admin, the kernel privileged user). The file has to be open during normal operation, so “double encrypting” the file an additional time and then storing the encryption key inside the PC is security theater, because the file has to be opened anyway, and everyone knows this and would immediately spout off if anyone did it in any other circumstance. Yes, if you force your way to system user you can read the file, because the system user needs to read it, but at that point the system user can also just keylog you and screenshot you directly.
The complaints people are raising aren’t technically cogent, but they don’t need to be, because they’re not technically based in the first place. People don’t want the feature and they find reasons to complain. You can 10x it anytime apple is involved.
It is almost like they only came out with this to appease wall street.
I'll just remind you about one of the features: You heard of Memojis, now you have Genmoji!
Generating a squirrel on a jetski emoji on the fly will save you loads of time and make your life better. There aren't enough stickers in your chat app.
Pick from three googley eyed cartoon styles to show off your unique personalities and creativity.
I'm quite disappointed with latest WWDC - for developers they haven't thrown any bone or carrot. I was hoping they would allow for developers to run prompts locally even if not on iPhone then at least on M1 iPads/Macbooks. Sort of like preinstalled LLM with ollama. There are many ideas that could be useful for user but hard to pull off because of either business model and/or latency.
That's the plan. Money needs to come somewhere, Apple has perfect the upgrade game (slowing phones down, new features, ...) when essentially every phone is the same since some years (I have a Mi 11 Ultra and it's the first phone I'll never upgrade before it breaks).
It's not about running an LLM on a device, it's about the hardware - which is quite good and long lasting. In both cases, the hardware is good. But apple has been caught to slow down their phones so the incentive to buy a new one is higher (like the scandal slow down because of the battery usage of new features.. etc..)
The OG just stated to keep his Mi 11 ultra until it degrades, because he/she/it is happy with it and there won't be such games like making it slow or bring you somehow into situation that you need to buy a new iPhone, like now apple does - buy a new if you want to have LLM..
Also, what LLMs are running on your apple device that is not within the 100-90%=10% of all iPhones and other iDevices?
OG can use android apps for that, if he needs LLM or other features of AI. It's even possible to side load them. Which is not possible and not sure for Apple. You can ask a question:
If a feature is built in into iOS, is it possible to write an app for iOS and having this accepted for AppStore? Look up the terms for developers. It's written: no.
So having apple introduce their AI solution, which is built in and sitting on top of openAI... No other app will be accepted that's doing the same . Nor will the apps have deep integration into the OS, because Apple doesn't want it. They also won't change their minds on it.
So this is the point of OP. It's not to be answered with in a sense "so? Then tell me what LLMs are running on your Mi11... Because there aren't and Apple now introduced local LLMs.." are you an Apple fanboy/fan girl?
Nope. Would that invalidate the facts? Are you an Applephobe?
They stated that apple phones are essentially the same as they were years ago, when it is very clear that the hardware is considerably more powerful. Their devices are now capable of on-device inference of a 3B parameter language model.
The commenters device is not capable of such a thing, as it is old hardware, which is counter to his position that there is little difference between the new and old phones.
Nope, I'm also not a applephobe. I like their devices and software, but I don't like their politics. I think we're aligned here in our thinking may the best product prosper!
Your question wasn't that clear. Also, I think Mi11 will be capable of doing such things, because Google is also eager to have their Models running locally. I remember a news from January of this year, that Google brings their Gemma 2B locally on devices. Other LLMs are also eager to achieve that. It's not like you stated, Mi11 is an old hardware and nothing will run on it.
The next thing to consider is, is really 3B model is running on (only new) devices or is it just a mockup because openAI is in behind doing some magic?
What about your comparison, that Apple is capable to do 3B and Mi11 not because it's old hardware.. we just learned, that over 90 percent of all Apple devices can't do 3B.. isn't it the same like to say over 90 percent of all androids can't run 3B? ... Yes.. because their not built to do so. In Future you can assume, all of the new devices of certain price will be able to.. in fact, I didn't see the statistics, but more/less than 10 percent are capable right now and will get the software update soon for ...
When it's about RAM, Mi11 have 8gb of RAM.. the Mi11 pro has 12gb RAM. And both have sd slot for upto 256 GB.. so plenty of space.. Snapdragon 888 is also not the baddest one. Let's wait what community will bring us.
But with 90 percent of apple old hardware also can't run it, so there also won't be a community to hack on the old hardware.. so what?
Currently none b/c I think every NPU you can currently buy will be rapidly old tech. It's like iPhone 1 (had one), iPhone 2 (3G, had one) [...], iPod 1 (hard drive, had one), iPod 2 (had one) [...], or MacBook Pro 1 (white, had one), MacBook Pro 2 (aluminium , had one), [...] all rapidly changing over the first generations.
After some iterations of something truly new, as (local) LLMs, things will slow down.
Some very cynical comments on this thread. Maybe, just maybe the older hardware isn't really capable of running these features properly and it's not some evil conspiracy to "force" users into upgrading?
It’s the amount of RAM that matters. It’s supported on the M1 but not on the A14. Both have the same CPU core design and neural engine, including the same number of neural engine cores. The A14, however, has 4 or 6GB RAM while the M1 has at least 8GB.
As the models used can be quite large this probably means the A14 doesn’t have enough headroom for it give a good UX.
The iPhone 14 Pro also had a 16-core NPU. The biggest difference is that Apple claims that the iPhone 15 Pro is "up to 2x faster" for machine learning than the iPhone 14 Pro.
I don‘t get what they really complain about here? I mean first it’s Apple. They didn’t introduce a new paid service but a service which will be free to iPhone users this fall. And I strongly assume it will be iPhone 15Pro, iPhone16 and iPhone 16Pro by that time. So people can either upgrade to an older pro model or get the newest hottest device.
Second: When VW produces a new car nobody complains that the million drivers won’t get an update automatically. And even if we would say VW introduces a new AI parking system or whatever would it mean everybody would get it. I think the server side solution could work on older devices but that is rarely what Apple does. Also they specifically split the whole AI presentation from all the others features coming to iOS, macOS etc. not like millions of users won’t get anything with the next update.
>When VW produces a new car nobody complains that the million drivers won’t get an update automatically.
Terrible comparison because
1, Car makers doesn’t release new cars every single year
2, For most people a car is
the 2nd biggest investment after their house which is nowhere near a smartphone in cost (that most people get subsidized by their carrier)
It’s about the fact that the current install base (which also goes into the millions) won‘t get stuff for free.
If Apple is doing it for pushing sales (Mac users with an m1 will get it without hardware upgrades and these devices are over 3 years old by the time of release) is anybodies guess.
> smartphone in cost (that most people get subsidized by their carrier)
In almost all cases, this is nothing but an intransparent form of installment payments, coupled with a carrier lock-in that helps keep prices high (because most people or even families can only switch ever other year or so).
Maybe it’s still very common in the US, but many people have moved on from this.
There's always something to complain about. If the AI runs on servers then oh no there's no privacy. If it runs locally then oh no it won't run on my old hardware.
Given Apple's general emphasis on privacy and selling hardware, this seems like the right path for them. Personally I'm all for it.
It makes perfect sense to me that Apple would gatekeep which devices are allowed to offload the AI workload to their private cloud. Older devices would need to offload pretty much 100% of the AI workload to the cloud because those old devices don't have enough power. That's expensive.
Apple is not charging the user for this. The user has to pay for it with their device purchase.
It's purely about compute power[0]. The A17 Pro has 35 TOPS of Neural Engine performance, A16 Bionic has 17. A15 has 15.8.
The M-series have less performance, but I think they're compensated by being in iPads where they can use more power (battery) without impacting people's lives too much.
Yes, 25% more RAM is significant. As you said, the M1 has significantly less performance, but it has enough memory. We can infer that RAM is the deciding factor from that.
I was thinking the other day that it’s taken OpenAI quite a bit of time to bring out GPT-4o for free to users. Even paying users have limits of how much they could use GTP-4 and GPT-4 Turbo.
For Apple to build out a dedicated server farm to accommodate even 50% of all their current users would be extremely challenging. I think they need users to run an LLM locally just simply because they wouldn’t be able to handle the amount of compute it would require to handle every request server-side.
I’m very interesting to see how local AI will impact the amount of RAM that comes standard with their Macs and iPhones. The 12 year old Macbook I still have in a cupboard somewhere has 16 GB of RAM. That was max at the time, but today 128 GB is max while they still offer 8 GB. If they had done that 12 years ago they would have been offering 1, 2, 4, 8 and 16 GB of memory.
> For Apple to build out a dedicated server farm to accommodate even 50% of all their current users would be extremely challenging
fyi there have been rumors that apple is looking to scale out Mac Pro production for their own server farms to support this. Since ultra is a super niche product normally this obviously is going to be like, several orders of magnitude of increase here.
Obviously there is the conventional wisdom that apple doesn’t do server and the server market doesn’t do apple, and they may continue to prefer commodity/off-the-shelf server hardware for the bulk of their iCloud infra, but things may be different with the bottleneck on AI hardware at the moment.
>A huge set of media companies have shifted to using AV1
Such as? Most YouTube videos I am watching are still VP9 at 1080p/1440p, and there's no reason to watch 4K on phones (you still can, but lower battery life is your own choice in that case).
> Only the iPhone 15 Pro and Pro Max — out of the 24 models compatible with the new iOS 18 — will be able to run Apple Intelligence.
Prices are above 1200 EUR in Germany. Err, no. I'll stick to ChatGPT and Claude.
---
Perplexity.ai:
The iPhone 15 Pro in Germany costs:
For the 6.1-inch model:
- 128GB storage: €1,299[1][2][3]
- 256GB storage: €1,399[1][2][3]
For the 6.7-inch Pro Max model:
- 256GB storage: €1,499[2][4]
These prices are for an unlocked, SIM-free iPhone purchased directly from retailers like Coolblue, MediaMarkt, and the Apple online store in Germany. The prices include VAT but no carrier contract.[1][2][3][4]
But why would anyone buy a phone upfront when you can get them subsidized by the carrier with data and talk? I end up paying the same after 22
months but I also get more. Last phone I bought upfront was 20 years ago when WAP was still a thing.
Because then the carrier charge about 10x what they'd charge a SIM only customer for the data, and it usually ends up costing more for the handset going through finance than buying it outright
I’m in Hungary not Germany and there is no financing if you buy it from carrier. Price / 22 months (or 24) and that’s it. There is no APR or such. In fact I get it cheaper because I also get data and talk.
Not many of us are in Hungary. In the U.S., financing the phone almost always makes more sense (especially lately, in light of inflation).
There are some plans here like yours, however. I'm on Google Fi and it's the same. Like your setup, by paying in monthly installments (with no finance fees) I also received device insurance at no cost. I would have purchased the phone outright up front, but it made no financial sense to do so.
I don't understand any of it, except consumers here seem only to think in terms of how much they can afford per month, rather than how much they can afford (overall). Decades of cheap credit, I guess.
In the US, subsidizing phones was a scheme that locked customers into expensive "buy out" scenarios where large, incumbent carriers could act like cartels.
Customers could get a "$1000" phone for effectively free on contract (the discount contingent on completing the 2 or 3 year contract, but with an option to "upgrade" the phone and renew the contract every year). To exit the plan, the "discount" was revoked, and the remaining value of the phone, which was often 80% of its MSRP, had to be paid off in full.
These phones were also required to be "carrier locked" (unable to be moved to another network, of which there are many in the US), installed with a carrier firmware that limited features, forced to run carrier apps and software, and were often stuck on the major OS version they were purchased with. Even high-end phones became effectively useless before their functional lives ended.
I think the EU didn't put up with this, but the US only much more relatively recently established similar rights and protections for cellular customers.
Because not everyone has access to those sorts of plans.
In addition such plans, when they are available, invariably lock you in for a set period committing you to regular payments. Either that or the “free phone” is paid back out of the term of the contract thus still requiring you to pay full price at the get go.
It may "leave out" more than 90% of "current" users, but there's still the upgrade cycle a lot of us are going to be walking right into this fall.
As an iPhone 14 Pro owner that's been stuck in a carrier agreement for the past couple of years, there hasn't been any "incentive" for me to upgrade, and I'm naturally at my upgrade point now. I'll likely be going to a 16 Pro no matter what.
Microsoft needs hardware for their “AI pc” too, this is nothing new and the only alternatives are server-side (which people will complain about even more) or extremely slow performance which will chew up battery (which, guess what, people complain about even more).
People are looking to find fault with apple and are going to construct a reason to be upset no matter what. It’s literally every single thread that mentions apple at all, just wildly unfair attacks and special-pleading and arguments from perfectionism that no other vendor is ever held to.
We really, really need to come to terms with these problems around apple discourse, it is so bad-faith and so noxious every single time. I don’t mean mod action but as long as people keep tolerating it and encouraging it and especially the whole “apple sheeple” “people who buy apple are doing it because they’re morons chasing blue bubbles” etc needs to not be on a site like this, and that is going to require people in the communities to stop tolerating it.
I think it’s a poor decision by Apple if Apple Intelligence running on Apple’s Private Cloud Compute is not available on even a one year old phone like iPhone 15/15 Plus or a two year old iPhone 14 Pro/Pro Max.
Since this service is launching as beta later in the year and is limited to US English, it seems like just a beginning. Anybody buying an iPhone 16 or iPhone 15 Pro just for this may be in for a rude awakening when they’re left out next year with iPhone 17 getting better features on AI.
I agree with Mark Gurman when he says — in this piece [1] titled “ Apple’s Push to Infuse Devices With AI Will Take Years to Pay Off” — that the current plan of release isn’t going to fuel an iPhone sales supercycle for the iPhone 16 series.
92 comments
[ 2.5 ms ] story [ 147 ms ] threadNot suffering from FOMO is a delightful thing. You can enjoy life while others scramble after the “greatest” and latest. There are far more important things to worry about than whether you get ML models running on your phone.
I can't really judge at this point, but I really doubt there is hardware limitation for that, e.g. iPhone 15 Pro isn't so much powerful than iPhone 15. Even when it is - they could utilize their server side AI solution, but for them greed comes first.
That 2GB really could make the difference: they have a 3B language model, which is going to be memory hungry. There needs to be enough memory left for the rest of the phone and apps to function.
Even when it is - they could utilize their server side AI solution, but for them greed comes first.
You were saying above that you’re glad you don’t have it and now you’re complaining that you don’t have it?
And a reminder that Phi1.5 comes in at 2.8GB.
Getting a performant model even while using aggressive quantization is going to require more RAM.
iPhone 16 ProAI may end up having notebook levels of ram (12 / 16GB)
Probably they could offload more to their servers but it wouldn’t be the same experience definitely.
As many have already pointed out to you, its most likely the earlier hardware simply wont run their local model.
The partnership with OpenAI includes an agreement that they will not track/log apple users, and even then it is a fallback option (after apples OWN cloud model) that users must specifically agree to.
"Bring ChatGPT to Siri" is extremely misleading. It's less integrated with Siri than Google is.
That said, whether you think me a fanboy or not, your openAI criticism is wide of the mark. The chatgpt/openai integration is apparently going to be 100% opt-in and you will be prompted for consent every time data leaves your device. So much so that the economist said it risks becoming "ChatGDPR"[1].
[1] https://www.economist.com/business/2024/06/11/hey-siri-help-... (requires a subscription)
I’ll upgrade when the latest OS no longer installs on my iPhone 11 and even then I might not get the latest and greatest - better to wait for my carrier to offer a previous-generation iPhone for peanuts and use that to upgrade.
It's far less likely to cause problems when limited to tasks like summaries, copy editing and short text rewording.
Hallucination is most likely to cause headline-grabbing mistakes when you use LLMs to answer questions about the world or generate larger volumes of text from a short prompt. Apple effectively outsource that to a clearly branded ChatGPT integration, neatly assigning any blame for hallucinations to OpenAI.
I imagine they will roll out to other languages and devices relatively soon after. This article assumes too much about the future, and seems to be pushing an “Apple wants to sell you more hardware” narrative as if a hardware company wanting sales is somehow scandalous.
Despite that, I don’t think that is what is happening here. I imagine server models will be available across all/most devices not long after they do the initial launch.
If you’ve built the solution for commercial purposes you obviously don’t have any real moral concerns that can allow you to refuse a lawful request, while “build us a backdoor to unlock this iPhone” or “build us a backdoor to disable advanced data protection mode/e2ee mode” would be compelled speech.
The things they have complied on are fairly minor stuff like making airdrop not work with non-friends for more than 10 minutes at a time.
There are obvious privacy problems with streaming every user interaction to the cloud, which is why they’ve already avoided this for things like photo tagging etc. Maybe you could do anonymized inference similar to the way maps work (let’s say, apple hosts a session for an anonymized user identifier but they don’t know who) but it’s still very intimate information regardless.
On device really is better… not to mention the power concerns. Like I just do not get these objections at all, other than whining because it’s apple, and I think the same people would be whining 100x harder (with good justification) if apple was really steaming every user interaction to the cloud. We’d be hearing about the sham of apple marketing on privacy while sending your sexts to the cloud etc.
same thing for the whining about copilot. Like yeah if it’s on-device it ultimately is going to be a file on disk somewhere (omg SQLite, we hate that!!!). The whole disk is protected by bitlocker, and the file is additionally restricted tightly to the system user (not just admin, the kernel privileged user). The file has to be open during normal operation, so “double encrypting” the file an additional time and then storing the encryption key inside the PC is security theater, because the file has to be opened anyway, and everyone knows this and would immediately spout off if anyone did it in any other circumstance. Yes, if you force your way to system user you can read the file, because the system user needs to read it, but at that point the system user can also just keylog you and screenshot you directly.
The complaints people are raising aren’t technically cogent, but they don’t need to be, because they’re not technically based in the first place. People don’t want the feature and they find reasons to complain. You can 10x it anytime apple is involved.
I'll just remind you about one of the features: You heard of Memojis, now you have Genmoji!
Generating a squirrel on a jetski emoji on the fly will save you loads of time and make your life better. There aren't enough stickers in your chat app.
Pick from three googley eyed cartoon styles to show off your unique personalities and creativity.
They all look like Pixar and smarmy Dreamworks face nightmare fuel memes from over a decade ago (https://tvtropes.org/pmwiki/pmwiki.php/Main/DreamWorksFace).
This was an actual part of the demo. I'm not being factitious.
> It became clear that people wanted to know our views on generative AI. So we decided to embrace it and call it Apple Intelligence.
Wouldn't it be better it was available and you turned it off I'd you didn't want to use it?
The OG just stated to keep his Mi 11 ultra until it degrades, because he/she/it is happy with it and there won't be such games like making it slow or bring you somehow into situation that you need to buy a new iPhone, like now apple does - buy a new if you want to have LLM..
Also, what LLMs are running on your apple device that is not within the 100-90%=10% of all iPhones and other iDevices?
OG can use android apps for that, if he needs LLM or other features of AI. It's even possible to side load them. Which is not possible and not sure for Apple. You can ask a question:
If a feature is built in into iOS, is it possible to write an app for iOS and having this accepted for AppStore? Look up the terms for developers. It's written: no.
So having apple introduce their AI solution, which is built in and sitting on top of openAI... No other app will be accepted that's doing the same . Nor will the apps have deep integration into the OS, because Apple doesn't want it. They also won't change their minds on it.
So this is the point of OP. It's not to be answered with in a sense "so? Then tell me what LLMs are running on your Mi11... Because there aren't and Apple now introduced local LLMs.." are you an Apple fanboy/fan girl?
Nope. Would that invalidate the facts? Are you an Applephobe?
They stated that apple phones are essentially the same as they were years ago, when it is very clear that the hardware is considerably more powerful. Their devices are now capable of on-device inference of a 3B parameter language model.
The commenters device is not capable of such a thing, as it is old hardware, which is counter to his position that there is little difference between the new and old phones.
Your question wasn't that clear. Also, I think Mi11 will be capable of doing such things, because Google is also eager to have their Models running locally. I remember a news from January of this year, that Google brings their Gemma 2B locally on devices. Other LLMs are also eager to achieve that. It's not like you stated, Mi11 is an old hardware and nothing will run on it.
The next thing to consider is, is really 3B model is running on (only new) devices or is it just a mockup because openAI is in behind doing some magic?
What about your comparison, that Apple is capable to do 3B and Mi11 not because it's old hardware.. we just learned, that over 90 percent of all Apple devices can't do 3B.. isn't it the same like to say over 90 percent of all androids can't run 3B? ... Yes.. because their not built to do so. In Future you can assume, all of the new devices of certain price will be able to.. in fact, I didn't see the statistics, but more/less than 10 percent are capable right now and will get the software update soon for ...
So, how to counter that arguments? :)
Edit: Not from January, but from a month ago:
https://medium.com/@samvardhan777/running-lightweight-llms-o...
But with 90 percent of apple old hardware also can't run it, so there also won't be a community to hack on the old hardware.. so what?
After some iterations of something truly new, as (local) LLMs, things will slow down.
As the models used can be quite large this probably means the A14 doesn’t have enough headroom for it give a good UX.
But I did just find this, which confirms what you’re saying. https://www.apple.com/apple-intelligence/
The iPhone 14 Pro also had a 16-core NPU. The biggest difference is that Apple claims that the iPhone 15 Pro is "up to 2x faster" for machine learning than the iPhone 14 Pro.
Second: When VW produces a new car nobody complains that the million drivers won’t get an update automatically. And even if we would say VW introduces a new AI parking system or whatever would it mean everybody would get it. I think the server side solution could work on older devices but that is rarely what Apple does. Also they specifically split the whole AI presentation from all the others features coming to iOS, macOS etc. not like millions of users won’t get anything with the next update.
Terrible comparison because
1, Car makers doesn’t release new cars every single year
2, For most people a car is the 2nd biggest investment after their house which is nowhere near a smartphone in cost (that most people get subsidized by their carrier)
They do. The 4 numbers that typically start with 20 or sometimes 19 indicate the model year of a vehicle which does change periodically.
If Apple is doing it for pushing sales (Mac users with an m1 will get it without hardware upgrades and these devices are over 3 years old by the time of release) is anybodies guess.
In almost all cases, this is nothing but an intransparent form of installment payments, coupled with a carrier lock-in that helps keep prices high (because most people or even families can only switch ever other year or so).
Maybe it’s still very common in the US, but many people have moved on from this.
Given Apple's general emphasis on privacy and selling hardware, this seems like the right path for them. Personally I'm all for it.
https://www.theverge.com/2024/6/13/24175985/apple-intelligen...
It makes perfect sense to me that Apple would gatekeep which devices are allowed to offload the AI workload to their private cloud. Older devices would need to offload pretty much 100% of the AI workload to the cloud because those old devices don't have enough power. That's expensive.
Apple is not charging the user for this. The user has to pay for it with their device purchase.
The M-series have less performance, but I think they're compensated by being in iPads where they can use more power (battery) without impacting people's lives too much.
[0] https://en.wikipedia.org/wiki/Apple_A17
But it might be significant too. Still it's not an arbitrary limitation, but an actual lack of processing power on the older devices.
For Apple to build out a dedicated server farm to accommodate even 50% of all their current users would be extremely challenging. I think they need users to run an LLM locally just simply because they wouldn’t be able to handle the amount of compute it would require to handle every request server-side.
I’m very interesting to see how local AI will impact the amount of RAM that comes standard with their Macs and iPhones. The 12 year old Macbook I still have in a cupboard somewhere has 16 GB of RAM. That was max at the time, but today 128 GB is max while they still offer 8 GB. If they had done that 12 years ago they would have been offering 1, 2, 4, 8 and 16 GB of memory.
fyi there have been rumors that apple is looking to scale out Mac Pro production for their own server farms to support this. Since ultra is a super niche product normally this obviously is going to be like, several orders of magnitude of increase here.
Obviously there is the conventional wisdom that apple doesn’t do server and the server market doesn’t do apple, and they may continue to prefer commodity/off-the-shelf server hardware for the bulk of their iCloud infra, but things may be different with the bottleneck on AI hardware at the moment.
A huge set of media companies have shifted to using AV1 and these older devices are gonna get hammered for battery.
I was planning on upgrading anyway just to get hardware AV1 decode given youtube music/youtube are one of my most frequent apps.
Surprised there wasn't more furore when Google mandated that shift.
Such as? Most YouTube videos I am watching are still VP9 at 1080p/1440p, and there's no reason to watch 4K on phones (you still can, but lower battery life is your own choice in that case).
Prices are above 1200 EUR in Germany. Err, no. I'll stick to ChatGPT and Claude.
---
Perplexity.ai:
The iPhone 15 Pro in Germany costs:
For the 6.1-inch model:
- 128GB storage: €1,299[1][2][3]
- 256GB storage: €1,399[1][2][3]
For the 6.7-inch Pro Max model:
- 256GB storage: €1,499[2][4]
These prices are for an unlocked, SIM-free iPhone purchased directly from retailers like Coolblue, MediaMarkt, and the Apple online store in Germany. The prices include VAT but no carrier contract.[1][2][3][4]
Citations: [1] https://www.coolblue.de/en/mobile-phones/smartphones/apple/a... [2] https://www.mediamarkt.de/de/brand/apple/iphone/iphone-15-pr... [3] https://www.apple.com/de/shop/buy-iphone/iphone-15-pro [4] https://www.apple.com/de/shop/buy-iphone/iphone-15-pro/6,7%2... [5] https://www.apple.com/de/shop/buy-iphone/iphone-15-pro/6,1%2...
I’m in Hungary not Germany and there is no financing if you buy it from carrier. Price / 22 months (or 24) and that’s it. There is no APR or such. In fact I get it cheaper because I also get data and talk.
There are some plans here like yours, however. I'm on Google Fi and it's the same. Like your setup, by paying in monthly installments (with no finance fees) I also received device insurance at no cost. I would have purchased the phone outright up front, but it made no financial sense to do so.
I don't understand any of it, except consumers here seem only to think in terms of how much they can afford per month, rather than how much they can afford (overall). Decades of cheap credit, I guess.
Customers could get a "$1000" phone for effectively free on contract (the discount contingent on completing the 2 or 3 year contract, but with an option to "upgrade" the phone and renew the contract every year). To exit the plan, the "discount" was revoked, and the remaining value of the phone, which was often 80% of its MSRP, had to be paid off in full.
These phones were also required to be "carrier locked" (unable to be moved to another network, of which there are many in the US), installed with a carrier firmware that limited features, forced to run carrier apps and software, and were often stuck on the major OS version they were purchased with. Even high-end phones became effectively useless before their functional lives ended.
I think the EU didn't put up with this, but the US only much more relatively recently established similar rights and protections for cellular customers.
In addition such plans, when they are available, invariably lock you in for a set period committing you to regular payments. Either that or the “free phone” is paid back out of the term of the contract thus still requiring you to pay full price at the get go.
This is how it goes, at least here in the USA.
I’ve bought every iPhone I’ve owned without a contract both in Europe and the US and have come out ahead after 24 months every single time.
I don’t think Apple actually sells them to carriers at a meaningful discount vs. retail. If they do, that makes it an even worse deal.
As an iPhone 14 Pro owner that's been stuck in a carrier agreement for the past couple of years, there hasn't been any "incentive" for me to upgrade, and I'm naturally at my upgrade point now. I'll likely be going to a 16 Pro no matter what.
People are looking to find fault with apple and are going to construct a reason to be upset no matter what. It’s literally every single thread that mentions apple at all, just wildly unfair attacks and special-pleading and arguments from perfectionism that no other vendor is ever held to.
We really, really need to come to terms with these problems around apple discourse, it is so bad-faith and so noxious every single time. I don’t mean mod action but as long as people keep tolerating it and encouraging it and especially the whole “apple sheeple” “people who buy apple are doing it because they’re morons chasing blue bubbles” etc needs to not be on a site like this, and that is going to require people in the communities to stop tolerating it.
Since this service is launching as beta later in the year and is limited to US English, it seems like just a beginning. Anybody buying an iPhone 16 or iPhone 15 Pro just for this may be in for a rude awakening when they’re left out next year with iPhone 17 getting better features on AI.
I agree with Mark Gurman when he says — in this piece [1] titled “ Apple’s Push to Infuse Devices With AI Will Take Years to Pay Off” — that the current plan of release isn’t going to fuel an iPhone sales supercycle for the iPhone 16 series.
[1]: https://archive.ph/srPHa