2026年WWDC上的Siri AI
英文摘要
Apple announced new Siri AI features at WWDC 2026, licensing a custom Gemini-derived model for Private Cloud Compute. They are leveraging vision LLMs to extract information from the user's screen, avoiding the need for app-specific integrations. The new Core AI library allows developers to run PyTorch models on Apple hardware. The Private Cloud Compute Gemini models will run on Google Cloud using NVIDIA GPUs, with security and privacy protections.
中文摘要
苹果在2026年WWDC上宣布了新的Siri AI功能，授权了定制的Gemini模型用于Private Cloud Compute。他们利用视觉LLM从用户屏幕提取信息，避免了应用特定集成的需求。新的Core AI库允许开发者在苹果硬件上运行PyTorch模型。Private Cloud Compute Gemini模型将在Google Cloud上使用NVIDIA GPU运行，并具有安全隐私保护。
关键要点
Apple licenses a custom Gemini-derived model for Siri AI and Private Cloud Compute.
苹果授权了定制的Gemini模型用于Siri AI和Private Cloud Compute。
Vision LLMs extract information from the user's screen, eliminating need for app-specific integrations.
视觉LLM从用户屏幕提取信息，消除了应用特定集成的需求。
Core AI library bridges PyTorch and Apple hardware, enabling developers to run their own models.
Core AI库连接PyTorch和苹果硬件，使开发者能够运行自己的模型。
Private Cloud Compute on Google Cloud uses NVIDIA GPUs with strong security measures.
Google Cloud上的Private Cloud Compute使用NVIDIA GPU，具有强大的安全措施。
iOS 27 Developer Beta includes new Siri AI features, but access is via a waitlist.
iOS 27开发者测试版包含新的Siri AI功能，但需要通过等待列表获取访问权限。