by Jason Snell
Apple may work with Google on A.I. features
Quite a scoop from Mark Gurman at Bloomberg:
Apple Inc. is in talks to build Google’s Gemini artificial intelligence engine into the iPhone, according to people familiar with the situation, setting the stage for a blockbuster agreement that would shake up the AI industry.
The two companies are in active negotiations to let Apple license Gemini, Google’s set of generative AI models, to power some new features coming to the iPhone software this year, said the people, who asked not to be identified because the deliberations are private. Apple also recently held discussions with OpenAI and has considered using its model, according to the people.
As I said on Upgrade this week, these are the early days of large language models and there are a lot of competitors in the market. While Apple may continue to build its own LLM—it may one day become a major competitive advantage for Apple to control this technology, as it is for so many others—right now it’s probably far more reasonable for Apple to partner with one of the existing companies in the field. (As Ben Thompson noted on Monday, Apple probably couldn’t get the Nvidia GPUs it would need, regardless.)
On top of all that, a big LLM alone will not help the iPhone. Apple’s secret sauce will come from integrating A.I. assistant features with its voice assistant, with its own apps, and with third-party apps on its platform (presumably via methods like App Intents, used by Shortcuts). If Apple can build a new version of Siri that is smart enough to use disparate sources to serve the user, and one of those sources is a Google Gemini LLM, that’s all good.
Put another way, I’d much rather have Apple technology mediating my relationship with various APIs and A.I. models. If Apple can build a great intermediary, it won’t matter so much to users about what technologies are powering that experience.