Yep - bout all I use Siri for these days. Cooking timers, alarms for work meetings, etc. lol
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple in Talks With Google to Power Next-Gen Siri With Gemini AI
- Thread starter MacRumors
- Start date
- Sort by reaction score
That’s Apple and Ai in a nutchell.
But I’’m sure Siri have a ’grea’t answer to that.
It will be a great marriage then.
Siri rebuilt for Declined iIntelligence
I’ve been doing some research recently and what some may find surprising. LLM’s require some device to run on! Every LLM made by every company that’s distributing them requires some device, NOT made by those companies, in order to be available to their customers! So, folks will buy some device they prefer, either from Google or their partners or from Apple, to use those LLM’s.
Meaning, even with no LLM strategy, they will still sell 100 million plus iPhones in the next year. And, very likely, 100 million plus iPhones in the year after that. That’s a small number compared to how many Android devices will sell, but not sure if I’d define that as ‘dim’.
That's a big "if". What this tells me is that Apple really doesn't believe the privacy is a human right.
They did though, and it's in iOS 26. It just isn't integrated into Siri yet. But you can create a shortcut to tap directly into their language model and ask questions.
Please don't talk about things you know nothing about.
1. Why is Apple working with Google in this space? Because Apple trains its (MULTIPLE) models on TPUs.
2. Apple has at least two known "sexy" models, along with multiple more specialized models. The two sexy models are called AFM (Apple Foundation Model). They are both *combined* language and vision models.
The on-device model is ~3B parameters, quantized to 2b per parameter.
The server model runs on Apple HW (the GPU of what are presumably M2 or M3 Ultras) and has about 200B models, quantized to about 3.5b per parameter.
Both are competitive (not obviously worse or obviously better) than counterparts *at this size* (which is definitely not the largest size, guesses are eg chatGPT5 is ~8 times larger).
Unusual (or at least non-obvious) aspects of these models include
- combined text and vision rather than two separate models
- multilingual, handling about 15 languages
- somewhat curated training rather than a raw flood of internet text (unclear if this curation helps or hinders performance, but it is there)
- emphasis on "IQ 100" level training not advanced training or reasoning. Apple wants the model to answer sensibly if you want to split tips, but does not [for now...] care how it responds if you give it your calculus homework
3. BY FAR the most important difference of these models, compared, to other models, is that they have been trained to handle a very specific task: a developer can, to simplify immensely, using natural and simple Swift, construct a query to be given to the LLM, and get back a response that is defined in terms of the structs and APIs of the calling app. This is not the same thing as "ask the LLM a question and get back text", and it's not the same thing as " ask the LLM a question and it gives you code that, hopefully, you can compile to do what you want".
The idea is that in a random app, like Uber Eats, I can say "I'd like to order that food we had two weeks ago, it was Asia, Thai I think, but I can't remember the name" and this will result in Uber Eats throwing up a useful order that you can click on. Look at what's involved here: the query has to go into the LLM (so that it can be "understood", Uber Eats also has to provide the database of recent orders (so that the LLM has a clue what was ordered over the relevant time period) and the response can't just be a text string like "looks like you ordered Satay Chicken, and by the way that's Indonesian not Thai", it has to be some sort of structure that plugs into Uber Eats API's to allow the construction of a genuine order.
No-one else has anything like this. THIS is what I mean when I say that Apple is constructing an AI UI, not an AI model.
4. So the position Apple finds itself in is that
- it's trying to figure out how to utilize LLMs AS API.
- this is a research task, so it makes no sense to try to do this research at the same time as you're constantly modifying an LLM that takes $100M per training run! Instead you fiddle around with tiny models, until you think you have the basic concepts required for the APIs (and their pieces within the OS runtime and Swift) all working.
- then you scale this up to a mid-sized model and validate that it still works.
- are they scaling it up to a much larger model? Maybe? Or maybe there is no point in doing that until we get a year or so of experience with this machinery to see what needs to be improved, changed, or discarded?
Apple is not playing the same game as Gemini, OpenAI etc. And it doesn't need to, just like Apple is not trying to compete with Google as a search engine. As long as the leading edge LLMs continue to provide as good an experience on Apple as they do anywhere else, then no-one feels any need to stop buying Apple HW just to get "optimal" ChatGPT.
This is all described in
Aspects of the Swift API were demo'd and examples given at multiple talks at WWDC earlier this year.
Apple did something very stupid last year with the announcement of Apple Intelligence before it was ready and before all this infrastructure was in place. That tells us that Apple Marketing were stupid in this case, and should have followed their normal rule of **** until you're ready to ship. But Apple Marketing is not Apple Engineering, and Apple Engineering have a plan in place grander than anything you can imagine.
Apple misses the boat and now scrambles to buy it in 😂 Tim needs to go quickly
Seeing a lot of misplaced doubt in the first few pages of comments.
Reminder, the first iTunes capable phone was made by Motorola (and from what I understand wasn't very good). The first iPhone defaulted to Google for maps. Apple doesn't have a search engine, or a social media platform (although they tried a few times). They also didn't invent Siri, they bought it. They didn't used to make their own CPU (remember Motorola, IBM, and Intel?). Anyone remember the IE as the default browser back in the day?
IF they choose to use a 3rd party LLM that doesn't mean anything. They can switch it out the moment they think they (or anyone else) have something just as good or better. Heck, they could make it pluggable, pick your preferred LLM that works best for you.
Also, for all those saying they wouldn't trust a Google LLM, why? If it is running on Apple's private platform and doesn't send any data to Google, why does it matter what LLM it is running?
Reminder, the first iTunes capable phone was made by Motorola (and from what I understand wasn't very good). The first iPhone defaulted to Google for maps. Apple doesn't have a search engine, or a social media platform (although they tried a few times). They also didn't invent Siri, they bought it. They didn't used to make their own CPU (remember Motorola, IBM, and Intel?). Anyone remember the IE as the default browser back in the day?
IF they choose to use a 3rd party LLM that doesn't mean anything. They can switch it out the moment they think they (or anyone else) have something just as good or better. Heck, they could make it pluggable, pick your preferred LLM that works best for you.
Also, for all those saying they wouldn't trust a Google LLM, why? If it is running on Apple's private platform and doesn't send any data to Google, why does it matter what LLM it is running?
Several things wrong here....
Firstly this is a clear admittance by Apple that they have utterly failed in the Ai market, totally, when you ignore the market for sooo long that your competitors not overtake you, but jump so utterly far ahead it is impossible for you to catch them in a time frame that is worth it.
Secondly bang goes any and all security and the realisation by Apple that no, you cannot have a phone capable of running complex Ai models locally on it... you HAVE to use the cloud!
IMO this alone should see Cook and several others kick out, it is ALL on them. Also should make some think why bother with Apple when you can go direct to Google for far less cost....
First they utterly failed wasting billions and endless months on a car project that went no where, and as such they chose to all completely ignore THE hottest biggest and most tech important market on the planet...
Firstly this is a clear admittance by Apple that they have utterly failed in the Ai market, totally, when you ignore the market for sooo long that your competitors not overtake you, but jump so utterly far ahead it is impossible for you to catch them in a time frame that is worth it.
Secondly bang goes any and all security and the realisation by Apple that no, you cannot have a phone capable of running complex Ai models locally on it... you HAVE to use the cloud!
IMO this alone should see Cook and several others kick out, it is ALL on them. Also should make some think why bother with Apple when you can go direct to Google for far less cost....
First they utterly failed wasting billions and endless months on a car project that went no where, and as such they chose to all completely ignore THE hottest biggest and most tech important market on the planet...
It has been quite obvious for a while now that Cook only concerned himself with stock price, not leading any markets. Not what Jobs did when he was alive. Jobs released so many new innovative products, Cook has released a few that were Jobs idea anyway and wasted time and money on everything else, but boosted that share price and stock value...
Yes. But this part of your comment is key:
"If they can come up with a secure and private way to use Gemini."
Google scrapes our personal data. It is who they are.
Apple announcing the next gen Siri powered by Google would be hilarious.
Rite!! All that Money it's that should be Poaching the Best LLM Minds on the Planet instead of watching evreyone else Zoom by at Warp Speed. They just lost another one to Meta that now makes Six (6) unbelievable what The Hell is going on at Park?
The vast majority of people buying iPhones dont care about AI. Apple is printing money from hardware. They are printing money from their services. AI is not driving, nor is it hurting Apple's main businesses. The only people who are focused on AI are us nerds here on the forums.
This is all contingent on app developers, who primarily make multi platform apps, adopting specialized proprietary app intents frameworks.
That’s the biggest downfall and short-sighted decision right now, Apple should be creating a standard not a system for lock-in because third party developers won’t adopt it quickly if at all.
There definitely is a long-term plan that will probably materialize in the first half of the 2030s with specialized hardware but they keep kneecapping themselves for the immediate term and are losing mindshare.
They seem willing to take the risk but time will tell if it’s wise or not.
My speculation is Apple’s development of their own LLM (as well as Siri) is handicapped by their adherence to privacy and possibly other positions they take, so just like they did with Google search and ChatGPT, they’ll offer more functional tools by way of built-in third party, as an option for those willing to agree to those third party terms. It gives users an option and keeps Apple’s hands clean as long as they make the change in terms clear.