Just over a week ago Google released a joint statement with Apple on their blog, The Keyword, revealing Siri will be powered by Gemini in the near future. I’ll include the entire statement because it couldn’t be shorter.
Apple and Google have entered into a multi-year collaboration under which the next generation of Apple Foundation Models will be based on Google’s Gemini models and cloud technology. These models will help power future Apple Intelligence features, including a more personalized Siri coming this year.
After careful evaluation, Apple determined that Google’s Al technology provides the most capable foundation for Apple Foundation Models and is excited about the innovative new experiences it will unlock for Apple users. Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple’s industry-leading privacy standards.
Four sentences. Apple’s own Newsroom blog has nothing to say about it, which says so much.
Okay, snark box checked.
This is a Good Thing™ for Apple’s users. Gemini is what I recommend to people who don’t know anything about LLMs but want to use one. It’s broadly capable, multi-modal, and, since Google is a part of many internet users’ daily lives, immediately useful in ways other LLM products are not. On top of that, Google has been around for a long time and doesn’t seem to be going anywhere. The future for OpenAI, Anthropic, and others is less certain.
The deal, along with Siri’s current ability to offload queries to ChatGPT, signals something else about Apple’s strategy with LLM-based products. They’re not joining the model arms race. Instead, they’ve resigned themselves to being a front-end for whichever seems most appropriate for their end user features.
And that’s the best position for Apple to be in right now. A position they blundered into through failure. By partnering, they can focus on delivering a suite of features that make Siri useful in ways its never been and in ways competitors haven’t thought of. That requires Apple to deliver in a space they have pedigree in—user experience.
At this point, not offering LLM features weakens current products and jeopardizes the ones on the horizon. Smart glasses will be a non-starter if your personal digital assistant can’t communicate with you via voice and understand the world you’re looking at. Pulling out a phone while wearing them will be a failure state companies will scramble to optimize away. Similar story for screenless smart devices made for the home.
The end game is too big. Taking an L now doesn’t preclude a W later. In hindsight these details may look fuzzy enough to seem like good strategy.