Last week, Tim Cook indicated Apple would introduce an AI product “later this year,” and analysts celebrated. Deepwater Asset Management’s Gene Munster said Cook uttered “the magic letters for the first time,” adding “I’d officially welcome Apple to the generative AI freight train.” Even after mixed earnings and light guidance, Apple’s stock price has jumped 5% since.
But Apple’s generative AI rollout may be more complicated than flipping a switch and enjoying NVIDIA-like prosperity. The company will have incumbency tradeoffs to consider, including how much to change iOS to make room for AI. And its culture isn’t built to develop sweeping AI products, as evidenced by its failures with the Homepod. Whatever the route, it won’t be simple.
That Apple missed the generative AI boom is not entirely surprising — and it definitely missed. “I can tell you in no uncertain terms that Apple executives were caught off guard by the industry’s sudden AI fever and have been scrambling since late last year to make up for lost time,” Mark Gurman wrote recently in Bloomberg. The company had been shipping iPhones and Macs at a record pace through the pandemic. And with business booming, its modus operandi was to refine its products, not reimagine them with radical innovation, so large language models didn’t merit much attention.
But as the “generative AI freight train” picked up steam, it became evident Apple would need to get involved. An AI device called the Rabbit R-1 became the hit of CES — and sold out immediately — after a demo of its conversational layer on top of standard apps wowed observers. Ex-Apple employees at Humane also released an AI pin with similar functionality. OpenAI, meanwhile, began working to enable AI agents. None of these products are an immediate threat to the iPhone and iOS, but all could eventually displace some operating system functionality. It would be foolish for Apple to ignore them. And it isn’t.
The right way to add generative AI to the iPhone isn’t entirely straightforward though. Apple can’t make the iPhone into the Rabbit R-1. It would simply be too big of a change. So, its options for a big product overhaul may be mostly limited to the chronically disappointing Siri. And indeed, a better Siri may roll out later this year, according to Gurman, along with incremental improvements like AI-assisted writing in Pages and AI playlists in Apple Music. Similar to Google, Apple may be reticent to risk its core product to introduce AI functionality, and hence realize limited benefits from it.
Even if Apple were to figure out the perfect AI product solution, its culture is not set up to build it. When reporting my book, Always Day One, I came across frustrated ex-Apple employees who said the company’s silos held them back meaningfully. Machine learning engineers working on FaceID, for instance, couldn’t speak with counterparts solving similar computer vision problems in its self-driving car division. As a result, everything moved slower.
Apple does, at times, concentrate resources toward critical initiatives when the situation calls for it, and that may happen for AI as well. “When those projects get the attention at the highest levels of the company, then they can draw from the whole company,” legendary Apple watcher John Gruber told me on Big Technology Podcastthis week. Apple, for instance, might bring its best camera engineers to a product like the Vision Pro when it’s a high enough priority. The HomePod might not have merited that push, but with hundreds of billions of dollars in market cap now at stake, Apple could develop a super team to move faster.
There are already signs that Apple is willing to step outside its comfort zone to build AI products. The notoriously secretive company open-sourced an image editing AI model this week, signaling it’s willing to part with some of its longstanding norms to recruit AI talent and push the cutting edge forward. It’ll take more moves like this — and perhaps some more radical — for Apple to capitalize on the moment.
The post Apple’s AI Moment Is Coming. It May Not Be Smooth. appeared first on TheWrap.