Manus AI, a general-purpose AI agent developed by Chinese startup Monica, has resumed normal operations after a brief social media crisis. Positioned as the world's first autonomous AI agent, the product can handle complex real-world tasks such as creating websites, analyzing stock market trends, and planning travel itineraries. Its performance in early 2025 was lauded by the media as a pride of Chinese AI, ranking alongside models like DeepSeek. On March 7th, the official Manus AI X account was suddenly suspended; the initial reason points to mentions from third parties.

Today at noon, Manus, the world's first general-purpose AI agent, officially joined Weibo, sparking enthusiastic attention and discussion among netizens. According to Weibo's AI official account, Manus is not just a smart assistant, but a "hands-on" super colleague capable of handling various tasks with a single instruction. From resume screening and travel planning to data analysis, PPT creation, and even coding and deploying web games, Manus is poised to become an efficiency powerhouse for office workers, students, and entrepreneurs alike.

Tongyi Lingma has announced the launch of its latest inference model, Qwen2.5-Max, offering developers powerful programming and mathematical capabilities. Trained on over 20 trillion tokens and incorporating a meticulously designed post-training scheme, Qwen2.5-Max demonstrates exceptional performance.

On February 21, 2025, Beijing time, the DeepSeek AI team, focused on exploring Artificial General Intelligence (AGI), made an exciting announcement on social media, officially kicking off their participation in the #OpenSourceWeek event. This small yet ambitious tech team announced that starting next week, they will open source five code repositories to share their research progress transparently with the global developer community.

Artificial intelligence is sweeping across the globe at an unprecedented speed, but a significant issue is becoming increasingly prominent: while general-purpose AI is proficient in many skills, it often struggles with specific regional cultures and languages. This is particularly true in culturally and linguistically diverse areas such as the Middle East and South Asia, where users urgently require AI models that understand the 'local language'. Although large general models can handle multiple languages, they often reveal a lack of cultural depth and in-depth regional knowledge when dealing with scenarios rich in regional context.

Recently, Tencent Technology (Shenzhen) Co., Ltd. published a patent regarding a training method and related equipment for large language models on the Tianyancha app. The patent is titled 'Training Method, Device, Computer Equipment, and Storage Medium for Large Language Models' and aims to enhance the learning capacity and accuracy of large language models through innovative training methods. In the training process of large language models, traditional methods often rely on a single text summary, which may lead to model overfitting and negatively impact the accuracy and diversity of generated content. However, Tencent's new...