虎嗅

"Siri AI is finally here: everything you need to know from Apple's WWDC26 in one article."

原文:Siri AI 终于来了,一文看完苹果WWDC26所有干货

Summary of Key Highlights

The core focus of Apple's WWDC26 event was the full integration of AI: the long-awaited Siri AI, which has been in development for two years, is finally available (using Google's Gemini technology while ensuring user privacy). Apple also introduced the Apple Intelligence intelligent agent that can automatically complete tasks. There were significant improvements in system design and performance. Additionally, it was announced that Tim Cook will step down as CEO in September, with John Ternus taking over as the new hardware lead. This event marks Apple's official entry into the AI competition and signals a shift towards a technology-driven leadership approach.

1. Siri AI Finally “Becomes Smart”: From “Artificial Fool” to “True Assistant”

The previous version of Siri was like a person with amnesia—forgetting context after asking for concert information followed by ticket purchase requests. With the new Siri AI, powered by Google’s Gemini, it can now think coherently:

  • Smooth Multi-Session Conversations: During the demo, users asked about concert times, how to buy tickets, set up a lottery reminder, and requested a new song from the artist. Siri understood each request without repetition.
  • Understanding Screen Content: When opening a beach photo on the phone and asking, “Siri, where is this?”, it identified it as Santa Cruz Beach. It could also find out the address of a friend named Jeff (stored in chat records) and plan a route to visit him while going to the beach.
  • App-Specific Integration: In a World Cup party scenario, Siri managed tasks like checking the schedule, recommending recipes, generating a menu, and sending group invitations without the need to switch apps, integrating Calendar, Safari, WeChat, and other tools.
  • Independent App and Multi-Device Support: Siri now has its own app and can create text/images or analyze files. It can be used on iPhones, iPads, and Macs, with various activation methods (such as swiping the screen or using the Memoji button).

However, there are some drawbacks: Siri currently only supports English, and EU users on iPhones and iPads cannot use it due to the DMA legislation; domestic users will need to wait for regulatory changes.

2. Using Google’s “Brain” Without Selling Privacy: Apple’s Security Measures

Many people worry that using Google’s Gemini means their data might be shared with Google. Apple guarantees this won’t happen and has implemented three layers of protection:

  • Local Processing for Simple Tasks: Basic tasks like setting alarms or checking the weather are processed locally, keeping data on the device.
  • Apple Private Cloud for More Complex Tasks: For more complex tasks like route planning, data is stored in Apple’s cloud, out of reach of others.
  • Encrypted Data Transfer for Difficult Tasks: Only the most challenging computations (e.g., analyzing large files) are sent to Google’s cloud, using NVIDIA chips for secure processing. Your input, model parameters, and results are encrypted throughout the process, keeping them confidential.

Apple essentially “borrows someone else’s brain” while keeping its own privacy intact.

3. Apple Intelligence: AI Taking Action on Its Own

The new Apple Intelligence aims to make AI more proactive:

  • Automatic Password Changes: The password management app will work with Safari to automatically update insecure passwords on various websites without user intervention.
  • Openness to Third-Party Models: In addition to Google’s Gemini, Apple also supports Anthropic’s Claude, allowing users to choose the AI they prefer. This creates a competitive environment for different models.
  • Enhanced Creative Tools: Writing tools provide real-time suggestions and corrections, and image generation has improved. Dictation functions can now be used in third-party apps (e.g., to polish sentences during WeChat conversations).

This is good news for developers, as they can integrate their apps with Siri AI using Apple’s App Intents interface.

4. System Design and Performance: User Feedback Leads to Improvements

The event also brought several practical improvements:

  • More User-Friendly Design: The previously criticized liquid glass interface now includes a transparency slider; the Mac design has been revised with a unified toolbar and a cleaner sidebar, with icons featuring a liquid glass effect.
  • Significant Performance Boosts: App launches are 30% faster, photo loading in the album is 70% quicker, AirDrop is up to 80% faster, and file transfer between iPads and hard drives is five times faster. More importantly, the CPU scheduling logic has been improved, benefiting older models like the iPhone 11 and later versions (this is the most widely used iOS version, so users with older devices will see significant improvements).
  • Practical Updates: The search function has been revamped, and iCloud photo sharing supports full resolution, making it accessible on Android and Windows devices.

5. Cook’s Retirement: Apple Entering a “Post-Cook Era”

This was Cook’s last WWDC as CEO; he will step down on September 1st, with John Ternus taking over. Under Cook’s leadership from 2011, Apple transformed from a technology company into the world’s most profitable consumer electronics firm. However, for 2026, Apple needs a leader for the AI era. Ternus, with a background in hardware engineering, understands how to implement technology effectively, suggesting a focus on combining AI and hardware.

In summary, WWDC26 marked Apple’s “AI coming-of-age” moment: finally bringing Siri AI to life after years of development, while establishing a strong foundation through privacy protection and an open ecosystem. The success of these features will depend on how well they meet user expectations in real-world use.

(The translation maintains the original Markdown structure, using natural language suitable for financial journalism, and adapts expressions to fit the target audience.)