The Apple Intelligence functions will not reach all Apple devices because not all of them are prepared internally to give the boost that the execution of the large language models needs, although it will reach the iPhone 15 due to its chip and RAM.
Apple Intelligence, Apple’s artificial intelligence system, which will be integrated with iOS 18, iPadOS 18 and macOS Sequoia to reach iPhone, iPad and Mac under a differential approach: it combines generative AI with the personal context of each user to offer an experience useful.
Although it is expected to arrive in the technology firm’s next devices, it will also do so in some already marketed models, as is the case with the iPhone 15 Pro series, but not the iPhone 15.
This is due to the combination of the processor and RAM used by the most advanced iPhones, given the computing demands required by large language models, as transmitted by the company’s executives after WWDC, and now reported by MacRumors.
Specifically, the A17 Pro chip has a neural engine or neural processing unit, with 16 cores with the capacity to execute 35 trillion operations per second. And all supported devices must have at least 8GB RAM.
“This is the hardware it requires. It’s a pretty extraordinary thing to run models of this power on an iPhone,” said Head of Software Engineering Craig Federighi.
“These models, when you run them at runtimes, are called inference, and inference of large language models is incredibly computationally expensive,” said Head of AI and Machine Learning John Giannandrea.
This manager also explained that “it’s a combination of bandwidth on the device, it’s the size of Apple’s neural engine, it’s the drive on the device to speed up these models enough to be useful.”
