Apple execs explain Apple's position in the AI race & how it isn't necessarily 'behind'
Marketing head Greg Jozwiak and software chief Craig Federighi share some familiar arguments about Apple Intelligence, Siri's place in it, and how they aren't technically in the same AI race.
Apple's running a different AI race
Apple's executives rarely speak in interviews, let alone share details about how they think about products beyond them being the best in class. However, a refreshingly honest snippet of an upcoming wider interview showed exactly where Apple sees itself in the bigger picture.
Wall Street Journal's Joanna Stern sat down with Apple SVP of Software Engineering Craig Federighi and SVP of Worldwide Marketing Greg Joswiak, and she's shared a seven-minute snippet focused on AI before the larger interview is released. Much of what Federighi discussed on Apple Intelligence echoed what was seen in an earlier conversation with iJustine, but new details emerged once Joz weighed in.
In both interviews, Federighi explained that Apple had working versions of the contextual Siri powered by app intents, and in fact, what was shown was actually running. Joz even scoffed at the idea of it being "demo ware," in what seemed to be a pointed comment at Daring Fireball's John Gruber.
Stern pressed the executives, asking about the usefulness of Apple Intelligence and if Apple could "keep up" with the competition. She also shared that she didn't use Apple Intelligence much herself and instead relied on competitors' products.
Here's Joz's response in full:
"Again, it's important to realize our strategy is a little bit different than some other people, right? Our idea of Apple intelligence is using generative AI to be an enabling technology for features across our operating system -- so much so that sometimes you're doing things you don't even realize that you're using Apple intelligence, or, you know, AI, to do them, and that's our goal.
Integrate it. There's no destination, there's no app called Apple Intelligence, which is different than a chat bot, which, again, what I think some people have kind of conflated a bit, like, 'Where's your chat bot?' We didn't do that.
What we decided was that we would give you access to one through ChatGPT, because, you know, we think that was the best one, but our idea is to integrate across the operating system, make it features that, you know, I certainly use every day."
His response echoed a lot of the sentiment in my piece I wrote about Apple's position in the AI race. So much pits Apple as a direct competitor to ChatGPT, and that just seems wrong.
Federighi takes it a step further, explaining that Apple doesn't need to deliver every technology on Earth. No one asked why Apple wasn't a shopping destination like Amazon, or why it didn't build a YouTube competitor, so it seems odd that everyone is clamoring for Apple to supply a chatbot.
You can watch the clip below.
These aren't new arguments
Instead of a destination, the executives explain that Apple Intelligence is a background framework that enhances what users do every day. They technically shouldn't even be thinking of the fact that they're using AI, let alone Apple Intelligence.
Image Playground made this sloppy image when prompted with 'WWDC'
The on-device, private, personal Apple Intelligence is only just starting to spread out across the operating systems outside of their early feature sets. Developers can start using the on-device model to achieve results they'd have had to pay ChatGPT for and siphon off user data to do so.
There are rumors that Siri will get an LLM backend, and even with those contextual updates via app intents, it doesn't seem like Apple will release an AI app or chatbot. Instead, the Apple ecosystem will act as a backend for personalized, contextual, and proactive interactions that occur across devices and apps.
As I've shared before, it may seem like Apple is behind because Image Playground makes terrible images and Siri still doesn't know how calendars work, but I wouldn't bet against them. While Apple pulls back to bide its time and ensure its features are well and truly ready for prime time, the rest of the world will pursue increasingly powerful slop generators running on heavily polluting power sources.
It's not that Apple is behind in the AI race, or even that it's waiting to leap ahead at the right moment, as I suggested in a previous piece, it's that they're running a totally different race. One that brings powerful apps and systems to iPhone while still giving users access to the tools they need via Visual Intelligence and integrations with ChatGPT.
For now, I'll continue to use Apple Intelligence every day for my work. I'm excited to see what developers will be able to do with the Foundation Models framework later in the fall.
Read on AppleInsider
Comments
Now we know why Gruber wasn't granted an interview with Apple execs this year.
What we don't know is A.) if they were using a demo version of the LLM that had no guardrails, B.) what the actual query was, or C.) whether or not the LLM itself has changed (this latter point is is extremely likely).
We also don't know the reliability of the answer, whether it would behave similar for ten thousand, one hundred thousand, one million, one billion, one trillion user requests. Apple senior management knows that their LLM is unreliable. It worked for that one specific instance that video recorded in 2024 but it wouldn't be the same for Joe Consumer using the production LLM model. Which is why there is no production Siri LLM model.
We have gone over this before.
Currently Apple is behind even their own plans while stating that everything is fine.
To give up on privacy and use OpenAI - this only happened as Apple was and is behind. To have still only 1 partner - another.
At this point the execs should be honest about it and tell how Apple will get back with a solid plan.
I use Claude, Gemini, ChatGPT and Grok (plus local models with the outstanding Apple MLX framework running on the beautiful M series Apple Silicon) along with a series of other models and services like Veo3, Midjourney etc.
All models have pros and cons and different "personalities" if one can call it that.
These services solve real problems for me and has transformed my productivity. When reasoning models trickled through another jolt of uplift of outcomes and frankly my main constraint is that my use case ideation process is too slow to fully use the vast capabilities.
Google IO was a mic drop of awesome. Apple is behaving like AI is an add-on when AI has already become integrated into a large set of use cases for daily life and business processes. The MLX crew are continuously knocking it out of the park but feels like an engineer led sideshow rather than strategy.
Apple's narrative that AI Siri is not good enough for Apple yet is just nonsense when their operational Siri stinks. I think most customers would take a half baked AI Siri over the current Siri. Perfect getting in the way of good.
There is no other race. Nor is there the notion of Apple racing with itself.
Apple was behind and remains behind.
With generative AI and Ai in general it literally doesn't matter if it is system-wide or in an app (or multiple apps).
AI brings options to users. Exactly how that is done is irrelevant.
Gruber was absolutely correct with his comments and I can understand why Apple feels awkward about having an 'ally' be so blunt.
The reality is that Apple can simply no longer ignore AI (the 'ML' WWDC) or promise major features (the 'AI' WWDC) and underdeliver them.
If Apple weren't behind, they would have not had to mention any of this in interviews. A competitive product would already be shipping and it would have shipped around 2022/23.
In the same vein, Apple would have its own AI training and inference hardware rolled out across the board by now.
In terms of systemwide AI (in the ML sense), that has been seeping into systems since 2017 with the first NPUs.
Of course, it's all got better and faster since then and will continue to do so.
Apple got very wrongfooted by all of this and we have watched things play out accordingly.
So maybe better to have the real journalists do the interviews and have Gruber do his thing (which I like a fair bit).
Where is Swift Assist? Why isn’t Siri better? Why is image playground so limited?
Apple isn’t losing IoT, because it does not sell IoT home goods. Nor do I believe it wants to. I am skeptical they want to sell deadbolts and outlets, which are niche markets. It’s just not their bag. The rumors of a HomePod-iPad mashup seem very unlikely to me.
As for not having Jobs, reality doesn’t match your claim. Apple has achieved more growth and more success under the “normal” CEO Cook than it did Jobs. Most CEOs are not also product managers and spokesmen, Jobs was the exception there and wound up delegating typical CEO duties to…Cook. Most CEOs of most brands are relatively nameless people who work on the business. That’s the job. That’s what Cook does. Product development is for product managers. The expectation that Apple must crank out new product category hit after hit is not how any company operates. Having the Mac, iPod, iPhone, iPad and Watch, all hits, is extremely rare. Expecting new ones and the lack of them as evidence that Apple lost its way or whatever, is contrary to normal reality.