Latest AI News

Jiemeng AI launched its first collaborative narrative creation tool, Xiaozhangyu "Octo, introducing the VibeCreate (Atmosphere Creation) mode, aiming to change the way AI content is generated, shifting from one-way instructions to a partner model of on-screen co-creation. Currently, the tool is only open for web-based beta testing, supporting conversation and multimodal interaction, marking a significant evolution in the paradigm of AI creation interaction.

Moxt is an AI-native online workspace designed to transform AI from a tool into an autonomous AI employee team, helping users cope with information anxiety and focus on high-value content. Its core philosophy holds that 95% of user folders are noise rather than assets, and by using AI agents to process massive amounts of information, it helps users escape from information overload.

Tencent Cloud announced that starting from May 9, 2026, it will increase the prices of core AI services, including AI computing power, container service TKE-Original Node, and elastic MapReduce product, with a uniform increase of 5%. Users who have already purchased services will not be affected by the new pricing on their current orders.

Google introduces the notebooks feature, making Gemini a personal knowledge assistant. This feature enables AI to have long-term memory by centrally managing content on specific topics. Users can consolidate scattered files, historical conversation records, and custom instructions into specific notebooks, achieving high integration and reusability of information, thereby improving the accuracy of context.

WeChat's new rules ban AI-generated content to protect originality, targeting automated creation and publishing.....

To maintain its leading position in the AI competition, Google DeepMind has integrated the company's computing power and talent, broken down internal barriers, and successfully transformed from a follower to a leader, with operational efficiency comparable to that of a startup. CEO Hassabis emphasized that computing power is the biggest bottleneck in AI research, and resource integration has reshaped its competitiveness.

ByteDance launches the native full-duplex speech large model Seeduplex, achieving 'simultaneous listening and speaking' processing, advancing AI speech interaction from 'turn-based' to 'real-time natural interaction'. The technology has been fully launched in the Douyin App, achieving large-scale deployment for hundreds of millions of users.

MiniMax launches the MMX-CLI command line tool, designed specifically for AI Agents. The tool simplifies interaction between Agents and large models, supporting native calling of MiniMax's multimodal models in mainstream development environments such as Claude Code, without the need for MCP Server adaptation, enhancing the capabilities of Agents in programming and creative environments.

Recently, the CEO of Amazon Web Services, Matt Garman, stated that despite Amazon's investments in two major AI competitors, OpenAI and Anthropic, this aligns with its business strategy. Garman believes that this investment relationship reflects healthy competition and will not lead to conflicts of interest, emphasizing that AWS focused on diverse collaboration from the early stages of its development.

A report from Boston Consulting Company states that within the next three years, artificial intelligence will reshape more than half of U.S. jobs, bringing about a productivity revolution. However, it will not completely replace human labor but rather change the way work is done.

A team from Stanford University published a study in "Nature Medicine," introducing the open-source AI model SleepFM. This model predicts health conditions and risk of death over the next six years by analyzing one night's sleep monitoring data (integrating brainwave, heart rate, breathing, and other deep physiological signals). The research was trained on 25 years of clinical data involving 65,000 people, surpassing the simple monitoring functions of traditional smartwatches.

AI applications provide personalized dietary advice for chronic disease patients, such as recommending porridge made with mixed grains instead of white porridge, steaming and boiling cooking methods, etc., to solve daily dietary problems, reflecting the trend of intelligent health consultation.

Anthropic launches the public beta version of Claude Managed Agents, offering a one-stop solution for building and deploying AI agents. The platform integrates optimized frameworks and production-grade infrastructure, providing comprehensive lifecycle management for AI agents, reducing development cycles to just a few days. Its highly integrated design allows developers to avoid building complex systems on their own.

Anthropic's new AI model Claude Mythos is priced at $125 per million output tokens, nearly 8 times higher than Claude Sonnet 4.6, causing a stir among developers. This pricing strategy signals a new phase of AI computing premium, with input and output costs far exceeding current flagship models.....

On April 9, Unitree's humanoid robot R1 debuted at Alibaba's Hangzhou campus, hinting at a collaboration. R1 mimicked human actions to operate a computer, displaying an 'invitation letter' on screen to promote the upcoming 'Shu Mai Tong' project launching in mid-April.....

Taobao's flash sale upgrade introduces an AI assistant for catering merchants, featuring voice search to execute complex tasks via commands, streamlining backend processes and boosting efficiency.....

Study reveals AI chatbots heavily rely on news sources, with a quarter of responses directly citing articles. Reuters and Forbes are top sources, highlighting reliance on industry media and journalists.....

Xiaomi Kaldi team open-sources the OmniVoice model, supporting over 600 languages. It achieves SOTA performance in multiple metrics on Chinese and multilingual TTS benchmark tests. The Chinese WER is as low as 0.84%, and the multilingual performance surpasses mainstream commercial models, achieving a new breakthrough in speech synthesis.

Google integrates the AI model Gemini into Gmail, improving email processing efficiency, and promises not to use user emails to train AI, ensuring data security isolation.

Dell's CEO warns that global memory demand for AI accelerators is experiencing explosive growth, and it is expected to surge 625 times by 2028 compared to 2023. This 'jump' in growth is primarily driven by both the capacity per machine and the scale of deployment, putting immense pressure on the semiconductor supply chain, with the imbalance between supply and demand unlikely to be resolved in the short term.