Apple’s AI strategy

This article is part of our series that explores the business of artificial intelligence

Like every year, Apple didn't fail to deliver at its annual Worldwide Developers Conference (WWDC). We were taken through a tour de force of new Mac computers, improved processors, operating system updates, application features, and of course, the Vision Pro headset. Apple continues to impress me with its feats of engineering, product development—and insane prices.

However, missing from the conference was the buzzword that has been dominating the developer conferences of other big tech companies: artificial intelligence.

Apple didn't mention AI in its WWDC presentations. But this doesn't mean that it is ignoring the technology and its developments altogether. In fact, AI was very much present in nearly every product presented at WWDC. The difference in the messaging comes down to how Apple views AI in comparison to Google, Microsoft, and Amazon.

Google I/O and Microsoft Build were mostly focused on highlighting the advances in the AI technologies that their respective companies are building. Basically, these companies are creating very advanced AI systems—and the current craze is around large language models and generative AI—and giving them to developers to use in various ways.

And this is a perfectly viable approach. OpenAI's ChatGPT seems to be a half-baked product. For every impressive example that you see with it, there are plenty of failures to show as well. But this reduces nothing from its usefulness. Developers and researchers are finding all sorts of techniques to overcome its flaws and put it to productive use. Microsoft has already built a $1 billion business on top of it and other LLMs and is integrating it into many of its products. Google has released its own rival, Bard, and is also using LLMs and generative AI to improve Search and other products.

This tech-focused approach is also why their conferences and product reveals are mostly focused on going into the technical details of the models. You hear a lot about benchmark performance and how the LLMs compare to other models.

Apple, on the other hand, bides its time and creates fully developed, well-engineered products. This is why the focus of their developer conference is not on describing their models and benchmark performances. Instead, they show you ready products and platforms that can solve specific problems. Developers then create applications for these platforms without being concerned about how the technology is working under the hood.

But Apple did present many AI-powered products at WWDC—though without necessarily mentioning the term. The Vision Pro headset, for example, is using AI for many of its features, including eye tracking, hand-motion tracking, sensor fusion, object detection, avatar generation, and more. Other products also used SOTA AI technology, including the AR and overlay functionalities in video conferencing tools and the new autocorrect feature in iOS 17 (the presenter used the term "transformer powered").

In some way, it makes perfect sense not to make AI the main highlight of the presentation. And that is because what makes Apple's products impressive is not how advanced the AI is but the way they combine cutting-edge technologies to create wonderful products. Just look at Vision Pro. Every piece of it deserves as much attention as the AI tech, including the advanced cameras, lidars, real-time processor, SoCs, AI algorithms, and the operating system that glues everything together.

The differences between the two strategies boil down to what you expect from the audience. The Microsoft/Google approach provides you with a technology that you can use in all kinds of ways, in different platforms and settings, etc. The Apple approach provides you with a fully baked platform, along with the entire hardware and software stack, on top of which you can create products.

The former leaves it up to your own skills and imagination to explore the space of possible products. The latter gives you the result of the combined work of some the brightest minds and expects you to explore possibilities within the confines of their vision.

But was Apple wrong to not play into the AI hype in WWDC? The company's stock took a hit after the developer conference, which could be because there was no mention of AI. And to be clear, Apple doesn't have a hyperscaler business to make sell language models APIs on top of it.

But in a way, Apple's presentation is a preview of what the future of AI can be. When a technology matures, it disappears and fades into the background. Just as companies don't boast about using transistors, microprocessors, touchscreens, UI, internet, cloud, WiFi, etc., there will come a day when neural networks, LLMs, generative models, and all technologies that fall under the umbrella of AI will become the norm. In this regard, Apple might already be ahead of its time.