With AI, is Apple Getting the Vapors?
Mark Gurman
While the main gist of Gurman's latest newsletter – that many of the "Apple Intelligence" features won't be coming until later this year or next – is hardly a surprise as Apple would casually sprinkle the "coming later this year" or "coming over the next year" throughout the presentation, a couple things are worth spelling out.
First, while I do think this is the right approach for many of the reasons Gurman notes (resource deployment, cloud infrastructure build out, smoothing usage bottlenecks, etc), it also makes me continue to worry more than a little bit that we can't trust in all the upgrades we heard about with Siri, in particular. Because nearly everything demoed with the voice assistant in the keynote won't be coming until next year, says Gurman. So does Apple even know and trust in how well it works right now? Is this one of those, let's draw a line in the sand and hope that the continued improvements in AI over time get us there?
Second, this entire roll-out – alongside several other iOS 18 features that won't be coming until later – continues Apple down the iterative software development path. I think this is a good thing and long overdue in many ways, but it's interesting from a marketing/optics perspective. Apple is still announcing all these new features at once in a big keynote at WWDC 2024 even though some of the features will launch a lot closer to WWDC 2025.
A lot of companies have long done these types of announcements. And it has led to the derisive use of the term "vaporware" within the Apple community over the years. Again, Apple has been slowly but surely going down this path, but this is the first year where a lot of the features and functionality are coming beyond the initial release. By pre-announcing, they run the risk that not all of them make it in time – or ever, see also: AirPower.
Should they be thinking different about the software releases now?
One more thing:
Like the big Siri improvements, Apple Intelligence’s support for additional languages and regions won’t arrive until 2025. The much-discussed integration with ChatGPT also may miss the initial release, though it should be ready later in 2024, based on language in Apple’s marketing materials.
It's surprising that ChatGPT is one of the feature that may not come until later. It feels like it should be pretty straightforward given that OpenAI is going to be doing most of the heavy-lifting there. But it may point to Apple's AI "routing" system not being ready yet – i.e. the way they determine if their own AI should handle a request on device or in-cloud, or if they should suggest ChatGPT for the query. And, in fact, if Apple's own AI isn't ready for prime time at the launch of the new OSes, would they really want to launch with just ChatGPT? Seems risky...
On the flipside, if they really wait until the end of the year, might deals be done with other AI-providers beyond ChatGPT by then? Is there some window of exclusivity baked into OpenAI's contract with Apple?
It would seem a bit wild if Apple launches with multiple LLM partners after touting ChatGPT on stage. But I certainly wouldn't put it past Apple!