Apple Embraces Google Gemini Technology! A New Era of AI Architecture with “Apple Intelligence” Takes Off

📰 News Overview

• Full Integration of Google Gemini Technology: Apple co-developed the foundational model for Apple Intelligence with Google. By integrating Gemini technology, understanding and reasoning capabilities have been significantly boosted.
• Hybrid Execution Environment: The new architecture is designed to operate optimally in both on-device processing and Apple’s exclusive “Private Cloud Compute” servers.
• Enhanced Multimodal Capabilities: The system now offers advanced multimodal functionalities, including realistic image generation, sophisticated photo editing, and visual question answering (Visual QA).

💡 Key Points

• Introduction of a New “System Orchestrator”: Positioned at the core of the architecture, it personalizes responses based on running apps and user tasks, seamlessly coordinating between platforms.
• High-Output Models for Specific Devices: Some devices will feature high-performance models that enable enhanced voice generation, improved dictation accuracy, and more powerful natural language understanding.
• Robust Privacy Protections: External experts will ensure a verifiable system where user data is guaranteed not to be shared with Apple or third parties.

🦈 Shark’s Eye (Curator’s Perspective)

The fact that Apple has partially opened its doors to integrate Google’s Gemini technology is a historic turning point, indeed! What’s noteworthy here is not just swapping out models, but the “System Orchestrator” that governs the whole system. This allows AI to understand “what app is currently running and what it’s doing” as context, making the system behave as a single intelligence. This sets it apart from competitors rushing to create “single-function AIs” and builds a robust user experience that’s very concrete! With privacy as a shield, integrating the best engines from competitors into their own chassis is a clever strategy typical of Apple!

🚀 What’s Next?

With Apple Intelligence obtaining what is essentially the industry standard for inference accuracy, the collaboration of third-party apps with AI is set to skyrocket! Furthermore, the lines between Android 17’s “Gemini Intelligence” and Apple’s offerings may blur, leading to a more interoperable AI ecosystem that transcends OS boundaries!

💬 Haru-Shark’s Take

Peeling back the layers of this tech apple reveals the sharp fangs of Gemini inside… what a shocker! A powerhouse duo racing through the AI ocean! 🦈🔥

📚 Terminology Explained

• System Orchestrator: A command center technology that determines which AI functionalities to use based on the user’s current tasks and open apps, controlling consistent operations across the system.

• Private Cloud Compute: A cloud computing framework proposed by Apple. It uses cloud resources while ensuring user privacy remains protected at levels comparable to on-device processing.

• Multimodal: Technology that can simultaneously understand and process multiple types of data, including text, images, audio, and video.

• Source: Apple reveals new AI architecture built around Google Gemini models