AI Daily: Tencent Launches AI Assistant Marvis; Zhipu AI Releases AutoClaw Mobile App; Photoshop 27.7 Major Update

Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers to help you understand technical trends and innovative AI product applications.

Discover new AI products click here:https://app.aibase.com/zh

1. Tencent Launches an AI Assistant at the Operating System Level - Marvis, Supporting Cross-Device Control and Local Privacy Mode

Tencent's Marvis AI assistant is deeply integrated at the operating system level, offering intelligent system management, security and privacy protection, and cross-device collaboration features, lowering the user's entry barrier and improving efficiency.

AiBase Highlights:

📱 Marvis, as an AI assistant at the operating system level, enables intelligent computer operations and task automation.

🔒 Offers a privacy mode and security mechanisms to ensure data stays on the device and sensitive operations require user confirmation.

🔄 Supports cross-device visual control, enabling real-time interaction and remote control between mobile phones and computers.

2. Zhipu AI Officially Launches AutoClaw Mobile APP, Supporting Cloud and Local Dual-Mode Execution

Zhipu AI launched the AutoClaw mobile app, supporting cloud and local dual-mode execution, further expanding the usage scenarios of AI agents, lowering the user access threshold, and laying the foundation for future feature expansion and ecosystem improvement.

AiBase Highlights:

📱 The AutoClaw mobile app is now available, broadening the usage scenarios of AI agents.

⚙️ Supports running in both local and cloud modes, enhancing the flexibility of task execution.

🔄 Future functions will be gradually expanded, completing the multi-terminal collaborative ecosystem loop.

3. Photoshop 27.7 Major Update: Removes Tool Supports Local AI, Generative Credit Benefits Doubled

The Photoshop 27.7 update introduces local support for the removal tool, enhancing privacy and security, and optimizing generative AI usage benefits and workflow.

AiBase Highlights:

🖼️ The removal tool introduces edge-side AI models, enhancing privacy and reducing reliance on the cloud.

💡 Generative AI benefits are upgraded, with credits increasing from 25 to 100, improving the workflow experience.

⚙️ Workflow becomes more user-friendly, interface is refreshed, performance is improved, and key bugs are fixed.

4. Amazon SageMaker AI Launches Real-Time Inference Endpoints Compatible with OpenAI API

Amazon SageMaker AI has launched real-time inference endpoints compatible with the OpenAI API, allowing users to easily call models on SageMaker without additional client customization or code rewriting. This feature supports multi-model hosting and simplifies the authentication process, providing users with a more efficient and flexible inference solution.

AiBase Highlights:

🌟 New OpenAI Compatible API: SageMaker AI real-time inference endpoints now support OpenAI API, requiring only a URL change to call the model.

🛠️ Multi-model Hosting: Users can host multiple models on the same endpoint and access them using the same OpenAI SDK.

🔑 Simplified Authentication Process: Supports Bearer Token authentication, making it easy for users to securely access SageMaker AI endpoints.

5. SpaceX's Prospectus Reveals Its Financial Status: Satellite Internet Drives Profitability, Aerospace and AI Still in Loss

SpaceX's prospectus reveals its financial status as the company transitions from a rocket launch enterprise to a satellite internet and AI infrastructure giant. Satellite internet is currently the main revenue source, while aerospace and AI businesses still face loss pressure.

AiBase Highlights:

🚀 Satellite internet business has become the main revenue source for SpaceX, generating $3.26 billion in revenue in Q1 2026.

📉 Aerospace and AI businesses are still in loss, with losses of $662 million and $2.469 billion respectively.

🌐 SpaceX plans to integrate rockets, satellite internet, and AI computing power, aiming to deploy orbital data centers by 2028.

6. Alibaba Tengzhuo Zhenwu M890 Makes a Big Debut: Performance Surges Three Times, Entering the Full-Stack Agent Era of "Chip-Cloud-Model-Inference"

Alibaba Cloud announced at a summit on May 20, 2026, the completion of a full-stack technology system upgrade aimed at the "Agentic (Agent) Era," marking a shift from "serving human users" to building an "AI factory" that supports massive agents.

AiBase Highlights:

🧠 Core Foundation: The Tengzhuo Zhenwu M890 chip boosts performance three times, suitable for high-precision training and low-precision inference in agent scenarios.

💻 Core Access Point: Reconstructing "Qianwen Cloud" to achieve AI-native interaction, quickly calling cloud platform capabilities through standardized Skills and CLI tools.

🚀 Technical Strategy: Full-stack integration of "Chip-Cloud-Model-Inference" solves challenges such as irregular elasticity and short lifecycle of agent loads, improving task goal alignment efficiency.

7. Combating Fake Images! OpenAI Introduces Google SynthID Image Watermark

OpenAI and Google have introduced a dual protection mechanism, enhancing AI image authenticity verification through metadata and invisible watermarks, aiming to establish industry standards.

AiBase Highlights:

🧠 A dual insurance mechanism of metadata and invisible watermark ensures image authenticity verification.

🛡️ C2PA standard provides transparent generation records, while SynthID watermark resists common processing.

🌐 Industry collaboration promotes development of AI image detection technology.

8. Alibaba Cloud BaiLian Upgrades Heavily: Full-Stack Open Access, Building a "Super Market" for Models

Alibaba Cloud BaiLian platform upgrades, integrating several third-party top models, creating an AI model super market, providing enterprises with a one-stop model selection and integration solution, promoting the development of China's AI ecosystem.

AiBase Highlights:

🧠 Integrates multiple industry-leading third-party top models, covering various dimensions including text, image, video, and multimodal generation.

🛠️ Provides an extremely simple integration solution, calling mainstream models through a unified API interface, reducing R&D and maintenance costs.

🌐 Builds an AI model super market, promoting the evolution of China's AI model ecosystem toward a more efficient and collaborative direction.