Welcome to the "AI Daily" column! 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, helping you understand technical trends and learn about innovative AI product applications.
Fresh AI products click to learn more:https://app.aibase.com/en
1. Google Gemini 3 Pro Preview model launched, developers can flexibly adjust parameters in AI Studio
Google announced its latest large language model, Gemini 3, and launched the Gemini 3 Pro Preview model on the AI Studio platform for developers and researchers to experience in advance. The model allows users to flexibly adjust key parameters such as context length and temperature to optimize application performance. Additionally, the Gemini 3 Pro version has also appeared in Vertex AI, offering users more development possibilities.

AiBase Highlights:
🌟 The Gemini 3 Pro Preview model has been launched on Google AI Studio, allowing developers to experience it in advance.
⚙️ AI Studio allows flexible adjustment of model parameters, improving development efficiency.
🚀 Gemini 3 will be gradually opened to the public, showcasing more powerful AI features.
2. Google suddenly released a big move! Antigravity AI IDE powered by Gemini 3 has arrived, directly beating Cursor and Claude?
Google released its new flagship large model Gemini 3 and launched a new AI-native integrated development environment (IDE) — Google Antigravity. The platform is centered around an 'agent-first' design philosophy, upgrading AI from a code assistant to an 'active partner', enhancing developer efficiency and solving trust issues.

AiBase Highlights:
🧠 **Agent-First Design**: Developers only need to describe high-level tasks, and the agent automatically creates plans, proposes architectural suggestions, and runs multiple tasks asynchronously.
🔍 **Verifiable Output Mechanism**: The agent generates task lists, screenshots, and recordings to ensure the credibility of code quality.
🖌️ **Visual Collaboration Feedback**: Supports annotating modification suggestions on web screenshots, providing a smooth human-machine collaboration experience similar to Figma.
Details link: https://antigravity.google/download
3. Cloudflare's automatic configuration file anomaly caused global downtime, restored within 6 hours
Cloudflare suffered a system crash due to an unexpected number of automatically generated configuration files for threat traffic, causing interruptions to multiple well-known services. During the peak of the failure, over 2.1 million abnormal reports were received, and the company restored core services within 6 hours.

AiBase Highlights:
🚨 Cloudflare's configuration file anomaly caused a global network crash, affecting multiple important services.
⏱️ Over 2.1 million abnormal reports were received during the peak of the failure, and the repair took about 6 hours.
🌐 The company emphasized that the incident was unrelated to a cyber attack, and a detailed technical analysis will be released later.
4. Microsoft tests "AI File Connector" in Windows 11: Claude can directly request local files
Microsoft tested the MCP (Model Context Protocol) connector in the latest Windows 11 preview version, allowing third-party AIs like Claude to request access to local files through the File Explorer without uploading them to the cloud to read content and complete tasks. This technology uses JSON-RPC 2.0 messaging mechanism, supporting both local stdio and remote HTTP transmission methods. The AI side declares the required file types through a standardized interface, and Windows returns the file handle and read-only stream, ensuring data remains on the local machine.

AiBase Highlights:
🧠 The MCP connector allows third-party AIs like Claude to directly request access to local files without uploading them to the cloud.
🌐 The technology is based on JSON-RPC 2.0, supporting local and remote transmission, ensuring data security.
🔒 Although the feature is powerful, the media reminds that AI might access sensitive files due to hallucinations, raising privacy concerns.
5. Baidu's Q3 AI revenue reached 9.6 billion yuan, up 50% year-on-year, Luobo Kuaipai travel volume doubles
Baidu's AI-related business performed well in the third quarter, with the three major segments of intelligent cloud, AI applications, and AI-native marketing totaling 9.6 billion yuan in revenue, an increase of over 50% year-on-year. At the same time, the autonomous driving ride-hailing platform 'Luobo Kuaipai' saw a significant increase in orders, showing Baidu's continuous investment and achievements in the AI field.

AiBase Highlights:
🧠 Baidu's AI business revenue increased significantly, indicating strong development momentum in the field of artificial intelligence.
🚗 The order volume of 'Luobo Kuaipai' doubled, highlighting Baidu's breakthroughs in autonomous driving technology.
💼 Baidu is accelerating the transformation to non-advertising businesses, aiming for AI revenue to account for more than half within three years.
6. Xiaomi AI Glasses Firmware Update: Live Streaming, Voice Control New Features Added
The firmware update for Xiaomi AI glasses brings several new features, including support for Douyin live streaming, controlling cars with Xiaoi, and English speaking practice, further enhancing the device's intelligence and user experience.

AiBase Highlights:
🗣️ New voice control function for Xiaomi cars, enhancing the smart experience for car owners.
🌐 Supports English speaking practice, helping users improve their language skills.
📱 Adds Douyin live streaming function, making it easier for users to share online.
7. TikTok Launches New Feature: Users Can Adjust the Amount of AI Content They See
TikTok launched a new feature that allows users to choose how many AI-generated content (AIGC) they want to see on the 'For You' page. This feature aims to enhance the user experience by letting users adjust the frequency of AI content display according to their personal preferences.

AiBase Highlights:
🔥 TikTok allows users to adjust the proportion of AI-generated content they receive.
💡 Users can select topics they are interested in through the 'Topic Management' tool and adjust the proportion of AI content they receive.
🌐 TikTok states this adjustment will not completely remove or replace a certain type of content but will allow users to enjoy recommendations that better match their preferences.
8. Microsoft Copilot AI Fully Integrated into Office: Enhancing Office Efficiency
At the Ignite 2025 event, Microsoft announced a series of updates for Microsoft 365 Copilot, aiming to push the office environment toward an intelligent era. These updates include fully applying the agent mode to the three core software of Office, such as Excel, Word, and PowerPoint, while upgrading Copilot's chat function and introducing a new business version.

AiBase Highlights:
📊 The new Excel agent can automatically analyze and visualize data, enhancing financial decision-making capabilities.
📝 The Word agent can quickly summarize documents and rewrite content, helping with efficient creation and understanding.
🌟 90% of Fortune 500 companies have started using Microsoft 365 Copilot, driving the intelligent transformation of office work.
9. Google Launches Generative UI: AI Will Generate Interactive Interfaces in Real-Time
Google Research released the Generative UI technology, enabling AI to generate interactive interfaces in real-time, enhancing users' understanding and operation experience of complex information.

AiBase Highlights:
🌟 Generative UI is a new technology launched by Google, allowing AI to generate interactive interfaces in real-time.
📊 Users can use simple instructions, and AI will generate dynamic visual experiences to improve understanding efficiency.
🔧 This technology relies on multiple core mechanisms, promoting the development of human-computer interaction toward a more intelligent direction.
Details link: https://generativeui.github.io/
10. Too tough! Manus turns your browser into a "Private AI Agent", crawl CRM, Crunchbase, PitchBook at will, no longer afraid of CAPTCHA and banning!
Manus introduced the Browser Operator feature, transforming the browser into a full-featured AI agent browser, using real IP and login status to achieve efficient automation, solving the pain points of traditional cloud tools.

AiBase Highlights:
🌍 Local-first: AI operations run entirely in the user's browser, using real IP and logged-in sessions, avoiding CAPTCHA and anti-scraping mechanisms.
⚙️ Efficient automation: Supports scenarios such as paid databases, CRMs, and internal networks, improving work efficiency.
🔒 Secure control: Users always maintain control, and all actions are completed locally, ensuring data security.
Details link: https://manus.im/en-us/features/manus-browser-operator
