According to the latest edition of Mark Gurman's Power On newsletter, some Apple employees believe the company is about two years behind in AI development. We also get some insight into several internal studies and a look ahead at Apple's strategy.
Apple Intelligence Recap
Apple unveiled Apple Intelligence back at WWDC24 in June, marking the company’s first foray into the world’s current AI craze. Apple has built a number of interesting features, including AI notification summaries, intelligent breakthroughs for important notifications, a brand new Siri with personalized context, Image Playground, Genmoji, and more.
However, one interesting part of Apple’s AI strategy that has remained under wraps until relatively late in the development cycle is that they don’t do it all themselves.
Apple Intelligence relies heavily on models that can run on-device, which also means that the requirements for Apple Intelligence to work are quite high. You need an A17 or M1 chipset or later with at least 8GB of memory. However, the fact that they run on the device also inherently limits the amount of information they can contain.
And for this reason, Apple also announced a partnership with OpenAI to integrate ChatGPT throughout the system. You don't have to use it, but if you want to gain additional insights, it's available.
ChatGPT vs Siri
OpenAI is building some of the world's greatest AI models, and Apple announced that it would support GPT-4o in iOS, iPadOS, and macOS. It will be integrated into Siri, as well as Writing Tools. The integration of ChatGPT was expected to close a knowledge gap, and now we know exactly how big that knowledge gap is.
According to Gurman, Apple's internal research shows that ChatGPT is about 25% more accurate than Siri and can answer about 30% more questions. He also later states that “some at Apple believe that its generative AI technology is — at least for now — more than two years behind the industry leaders.”
Apple's Future Strategy
Historically, Apple has proven its ability to catch up in areas where it seems to be lagging, such as Apple Maps. Gurman believes that Apple will catch up whether they do it themselves, hire people to do it, or acquire the companies needed to do it.
Best comment by Augusto Galindo
Honestly, any AI is at least 1 year behind ChatGPT, their model works in most situations, and all the tools they have like memory between chats or the ability to customize your own chat with specific data are amazing. For me, it’s become an everyday tool that Claude, Gemini, or even Copilot, which is ChatGPT in theory, won’t replace.
View all comments
Also, Gurman says that Apple Intelligence will be running on every device with a screen by 2026, with the iPhone SE getting the A18 chip in March as we expected, and the entry-level iPad “likely” getting an update later in 2025.
Apple obviously has the advantage of having a lot of devices with a high capacity to run AI models, so as they iterate on them, we’ll all benefit from it quickly. It’s just a question of how Apple evolves things in the future.
What do you think of Apple’s AI strategy? What kind of support would you like to see from Apple Intelligence in the future? Let us know in the comments.
Follow Michael: X/Twitter, Threads, Instagram
