The New Siri is... Google (Explained)
Quick Overview
Apple officially partnered with Google to power its AI-enhanced Siri using Google's Gemini models, with the update expected in 2026, a move that seems to acknowledge Apple's lagging position in the generative AI race despite their own announced 'Apple Intelligence' features.
Key Points: Apple selected Google's Gemini models to power a new iteration of AI-powered Siri, expected to launch in 2026. The partnership involves Apple paying Google roughly $1 billion annually to use a custom 1.2 trillion parameter Gemini model. This announcement follows earlier rumors from August 2025 and confirms Apple's decision to utilize an external model for complex AI tasks. The speaker highlights that while Apple's 'Apple Intelligence' features are being announced, complex world knowledge requests will likely be offloaded to ChatGPT or Google's models, as demonstrated by an on-screen prompt asking Siri to use ChatGPT for a complex trip planning request. The video contrasts this with Google's Circle to Search feature, demonstrating its immediate utility in image search and information retrieval on Android, suggesting Apple is playing catch-up. The host notes that the interface controlling which AI model is used (like in Raycast AI) is crucial, as users can select various models including Gemini, Claude, and others, making the user-facing integration important. The deal secures Google's position as the default search engine in Safari, a long-standing partnership worth billions annually, now extended into the generative AI domain.
Context: The video discusses recent news, confirmed by Mark Gurman, that Apple will integrate Google's Gemini large language models (LLMs) into Siri, likely arriving in 2026. This development comes after Apple announced its own AI initiative, 'Apple Intelligence,' but the partnership suggests Apple is leveraging Google's advanced models for complex generative AI tasks, a sector where Apple has been perceived as trailing competitors like Google and OpenAI.