ByteDance's Doubao large model team recently announced a breakthrough in addressing key bottlenecks in Mixture-of-Experts (MoE) architecture, open-sourcing a significant optimization technology called COMET. This technology dramatically improves large model training efficiency, achieving a remarkable 1.7x speedup and a 40% reduction in training costs. Image Note: Image generated by AI, image licensing provider Midjourney. COMET has been deployed in ByteDance's multi-thousand-GPU cluster training, resulting in millions of GPU hours saved.

Recently, in Hangzhou, Zhejiang, the rise of the AI tool DeepSeek has seen increasing numbers of entrepreneurs leveraging its powerful capabilities for profit. News of a Gen Z streamer generating 330 million Yuan in sales in a single day using DeepSeek for a live stream has attracted significant attention. Reports indicate that many businesses and individuals have experimented with new sales methods with the help of DeepSeek. For example, a business owner from Yiwu successfully sold out their entire inventory using DeepSeek.

Reports suggest that ServiceNow, a US software company, is in acquisition talks with AI-powered employee assistant provider Moveworks. This deal could be ServiceNow's largest acquisition ever, potentially reaching $3 billion. While discussions are reportedly in advanced stages, sources indicate the possibility of delays or the deal falling apart.

Recently, the tech world and developer community have been captivated by an open-source project called Browser Use! This tool is like giving AI wings, allowing it to control browsers as naturally as humans. Using natural language, users can direct AI to automatically complete various web tasks. Its powerful automation capabilities and flexible deployment have ignited the passion of tech enthusiasts globally, creating a wave of excitement on X (formerly Twitter). Browser Use is rapidly advancing the field of AI-powered browser automation.

Recently, the Modelers community officially launched Step-Video and Step-Audio, two open-source multimodal large models developed by Step-Star. These models are designed for video generation and voice interaction, respectively, aiming to provide developers and enterprise users with more powerful AI tools. Step-Video, formally known as Step-Video-T2V, is a 30-billion parameter model, making it the world's largest open-source video generation model. This model can directly generate 20...

Sony Music has announced a large-scale takedown of AI-generated counterfeit music. Reports indicate over 75,000 tracks mimicking artists like Harry Styles have been removed. Sony executives suggest this is only the tip of the iceberg, with the current takedowns representing a fraction of the AI-generated music circulating. Current identification and removal methods rely heavily on manual screening, meaning the actual number of AI-generated songs in the market likely far surpasses those already deleted.

Douyin's Safety Center recently announced a crackdown on stock investment content on its platform, aiming to foster a healthy online environment. While encouraging financial and securities professionals to share reliable information, Douyin is targeting accounts using AI tools to spread misinformation and even engage in fraudulent activities without proper qualifications. These accounts often promote supposedly high-efficiency AI stock picking tools to attract users.

On March 10th, Chengpian, an application under Baidu Wenku, officially integrated the full version of DeepSeek-R1, providing users with more efficient and powerful AI assistance. Based on Baidu Wenku's massive content accumulation and AI capabilities, Chengpian, after this upgrade, can better meet users' diverse needs in learning, life, and creation.

California Senator Jerry McNerney introduced the "No Robo Bosses Act" on Thursday. This bill, Senate Bill 7, seeks to ensure human oversight of artificial intelligence (AI) decision-making in the workplace, making it a first-of-its-kind proposal nationwide. The bill would prohibit California employers from solely relying on AI or automated decision systems (ADS) for hiring, promotion, discipline, or termination.

Reports indicate that the National Supercomputing Internet Platform has integrated Alibaba's Qwen large language model, officially providing the QwQ-32B API service. Users can access up to 1 million tokens free of charge, offering a valuable opportunity for developers and researchers. QwQ-32B is the latest open-source inference model from Alibaba's Tongyi team, demonstrating excellent performance. According to multiple authoritative benchmark tests, QwQ-32B's capabilities rival those of a fully-fledged 671B model.

Recently, Luo Yonghao's AR startup, Thin Red Line, announced its official launch of spring recruitment for 2025, attracting considerable attention. The company's currently open full-time positions are all for Product Managers, including Senior Software Product Manager, AI Software Product Manager, IM Software Product Manager, BI Data Product Manager, and Commercialization Product Manager. Locations are in Shanghai and Beijing, though specific salary ranges haven't been publicly disclosed. According to industry media outlet 36Kr, Luo Yonghao's latest...

The Department of Justice, according to latest court filings, is still seeking to force Google to divest its Chrome web browser. The proposal, initially suggested last year under President Biden, has persisted through the second Trump administration. However, the DOJ is no longer seeking divestiture of all AI investments, including Google's multi-billion dollar investment in Anthropic. The DOJ stated in court filings that Google's illegal conduct has created an economic behemoth that has severely impacted the market, ensuring that Google remains...

According to Qichacha APP, Huawei Technologies Co., Ltd. recently published a patent for "A Method and Apparatus for Generating AI Models." The patent abstract indicates that this technology, related to the field of artificial intelligence, can improve the efficiency of generating AI models and reduce resource consumption. The method involves: the first processor, based on the execution time of each operator in the multiple operators of the AI model, the memory occupancy of the second memory of the second processor during the operation of each operator, and the memory offloading time, constructs a target optimization with the operation instruction information as the optimization variable.

The AI revolution in image recognition is rapidly evolving, moving beyond simple cat-dog classification. Current trends focus on much finer distinctions, such as identifying the year and model of a specific car or even subtle differences like the thickness of a bird's eyebrow. However, while neural networks excel at classification, explaining their reasoning ('Why did you classify this as X?') often proves challenging. Traditional Class Activation Maps (CAM)

A recent study by DeskTime reveals ChatGPT as the most widely used AI tool in global offices, projecting a 75.9% adoption rate by the end of 2024. This survey encompassed 97 companies and 2385 employees, spanning from January 2023 to December 2024. With advancements in AI technology, ChatGPT's applications in the workplace continue to expand, its convenience and efficiency garnering widespread employee approval. Image source not specified.

According to the latest data from Similarweb, ChatGPT reached a record 390.5 million visits in February 2025, a slight 1.44% increase from January. While month-over-month growth has slowed, year-over-year growth compared to February 2024 remains strong at 137%. This continued growth has propelled ChatGPT further up the global website rankings; it currently ranks fifth among global desktop websites and seventh when combining desktop and mobile traffic. Data shows ChatGPT's significant hold on global internet traffic.

In a recent podcast, Microsoft CEO Satya Nadella stated that AI models are increasingly becoming standardized and commoditized as language models mature. He noted this shift is driving a strategic change in enterprise AI development, moving the focus from model research to system integration and product development. Nadella emphasized that models alone are insufficient; companies need to focus on complete system architecture and successful product creation. He believes the current AI industry is experiencing a "wave-like activity" since November 2022.