This is such a dramatic turnaround from the way Apple has positioned itself in the last few years you have to really wonder about their confidence level for in-house AI efforts.
An alternative read is that AI in its current form can produce such unpredictable results they're content for now to use Google as a scapegoat for any oddities.
Apple's in-house map stumbled badly at the beginning. Apple has been using Google search since the very beginning. It's reasonable that if we can't run a gpt-3.5 level model on a macbook air for 20 hours without charging, it's much better to use existing infrastructure and existing models.
The most important question is what kind of product Apple is creating.
This strikes me as less likely to happen for a few reasons.
First, there's anti-trust. Are regulators going to let Google own the smart-assistant on both smartphone platforms? There's already so many eyes on the Google/Apple search deal.
Second, why would Apple cede the smart-assistant to Google now? It seems like it would put them in a bad position long-term like it did with maps. Apple let Google be the maps on iPhones and then they watched as Google Maps on iPhones quickly fell behind their Android counterparts - until Apple came out with their own maps that forced Google to upgrade their iOS maps app.
Third, why would Google want to offer this to Apple unless it was licensed for a good amount of money (or a discount off being the default search on the iPhone)? Companies are racing to create generative AI assistants, but is there money in it? I'm not saying there isn't money in generative AI. I'm not saying there isn't money in ChatGPT Plus/API for $20/mo. But I don't think there's money in the AI assistant itself for Google to want to give it to Apple for free. With search, Google is getting ad revenue. If Siri-Gemini starts providing answers without searching, that's not good for Google.
Fourth, there are other companies. It notes OpenAI, but there's also Anthropic's Claude 3 and I'm sure others I'm unaware of. Apple just bought DarwinAI within the past week.
Apple might be behind on generative AI, but it seems like a partnership with Google here could be a misstep (obviously it depends on what the deal would be). I know that Apple worries about being behind, but Gemini hasn't been the most impressive and Apple was able to turn an Apple Maps fiasco into a wonderful app that I think is nicer than Google Maps. Are people clamoring for generative AI? To an extent, but it seems hasty to execute a deal with Google just to hit a fall-2024 deadline that probably won't impact iPhone sales that much. It seems like Apple should be a bit more deliberate with its plans here if generative AI is going to be a major part of their future.
Siri has been partially on-device, partially in cloud for awhile. It would make sense for them to continue this split, especially if they want these new features to be supported by all ios18 phones, which will likely stretch back to the iPhone 11.
An alternative read is that AI in its current form can produce such unpredictable results they're content for now to use Google as a scapegoat for any oddities.