Lo! The AiPhone Nears...
Michael Acton
On one hand, all of this feels fairly obvious. Yes, Apple is planning to have AI be a part of the iPhone going forward – just like every company currently has plans to have AI as a part of their products going forward.
Apple’s goal appears to be operating generative AI through mobile devices, which would allow AI chatbots and apps to run on the phone’s own hardware and software rather than be powered by cloud services in data centres.
Again, yes, of course this is the plan. Not only does Apple have billions of devices out there which they can use to differentiate any offering, those devices now mostly have their own custom silicon. But this is an interesting high level way of thinking about what Apple might be able to do which is operationally different:
Apple researchers published a paper in December announcing that they had made a breakthrough in running LLMs on-device by using Flash memory, meaning queries can be processed faster, even offline.
In October, it released an open-source LLM in partnership with Columbia University. “Ferret” is at present limited to research purposes and in effect acts as a second pair of eyes, telling the user what they are looking at, including specific objects within the image.
“One of the problems of an LLM is that the only way of experiencing the world is through text,” said Amanda Stent, director of the Davis Institute for AI at Colby College. “That’s what makes Ferret so exciting: you can start to literally connect the language to the real world.” At this stage, however, the cost of running a single “inference” query of this kind would be huge, Stent said.
I mean, plenty of these LLMs now have capabilities to, say, upload images or videos. But what Apple could do here is potentially different – again, because they have direct access to billions of devices, the majority of which have insanely great cameras built into them. This is part of what excited people about the Rabbit "R1" device that seemed to steal the show at CES this year. It wasn't just another software-focused AI tool, it was a tangible device for using AI in the real world.1
But unlike a newly created and shipping soon product (just like the forthcoming Humane "Ai Pin"), imagine if you had a device already in a billion pockets around the world. A device that is massively more powerful than anything a startup could possibly produce. Not to mention the planet's best channels to move new devices...
Apple is getting dinged for being "late" to the AI revolution. But given how fast things are moving and how much has yet to be determined here, if anything, anything they do with AI soon may still be too early. Things should get decidedly more interesting when they can ferret out what's being promised with "Ferret" above. But that will undoubtedly take time (and research), which is why Apple needs to start shipping AI stuff now. The war is for talent more so than products at the moment. That's the real risk to Apple.
1 Also undoubtedly part of the reason OpenAI is looking into hardware as well. With that old knight of Apple, no less.