This article will take you to learn about this week (5.12-5.18)AI hot news
Tsinghua University establishes the world’s first AI hospital
China established the world’s first AI hospital composed of 42 artificial intelligence doctors, developed by Tsinghua University’s Intelligent Industry Research Institute, integrating 21 cutting-edge technologies in medical departments.
In the future, we will build an ecological closed loop of “AI + Medical + Education + Scientific Research” to promote the efficient expansion and balanced layout of high-quality medical resources, and strive to allow more people to enjoy affordable and sustainable high-quality medical services.
The Central Cyberspace Affairs Office launched“Qinglang: Remediate the Abuse of AI Technology” Phase 2action
The Central Cyberspace Affairs Office launched the second phase of the “Qinglang and rectifying the abuse of AI technology”.
The second stage focuses on using AI technology to produce and publish prominent issues such as posting rumors, false information, pornographic and vulgar content, fake others, and engage in online water army activities, and concentrate on cleaning up related illegal and bad information, dealing with and punishing illegal accounts,MCN Agencyand website platform.
The U.S. Department of Commerce abolished Biden’s AI Proliferation Rules and required that Huawei AI chips be prohibited from being used worldwide
The US Department of Commerce announced the abolition of the AI Proliferation Rules signed by Biden, plans to launch more simplified rules, and add a new ban on Huawei’s global use of AI chips.
The original rule hierarchy system was complex, which aroused opposition from technology giants. The new rule replaced hierarchy with bilateral agreements to strengthen supply chain control.
This policy adjustment is beneficial to US technology companies, the Middle East market has gained new opportunities for chip imports, while China continues to be under pressure.
Google releasesAlphaEvolve，Can invent new algorithms, improve chip design, and solve difficult mathematical problems
Google DeepMind team launches based onGeminiDriven by general scientific artificial intelligence AlphaEvolve.
Using Gemini Flash and Gemini Pro models, the most promising algorithms are continuously improved based on the evolutionary framework.
AlphaEvolve has achieved remarkable results in data center scheduling, hardware design, AI training and complex mathematical problem solving:
- In data center scheduling, Google’s computing resources can be recycled by 0.7%;
- In terms of hardware design, the TPU circuit is optimized to make the chip more energy-saving;
- AlphaEvolve improved the method of splitting matrix multiplication into subproblems, shortening Geimini’s total training time by 1%, and increasing the running speed by 23%.
- Of the 50+ public problem tests in the field of mathematics, AlphaEvolve rediscovered the best solution to date in 75%. It includes the “kissing number problem” that has plagued the mathematics community for 300 years, and also subverts the Strassen algorithm myth 56 years ago.
From then on, AI is no longer just about solving existing problems, but also discovering new frontier fields.
Tencent Hunyuan Image 2.0 is released to experience “millisecond” AI image generation
Tencent releases the industry’s first real-time biographical model with millisecond response—Hunyuan Image 2.0realize the image generation while inputting text, and the quality of the raw images is significantly improved.
The world’s first professional design agent—Windward，Complete the full-link design and execution in one sentence
Lovart is the world’s first professional design agent created by LiblibAI for designers.
Lovart can think and perform design tasks like professional designers, providing high-level design solutions.
The entire visual process from creative disassembly to professional delivery can be completed in just one sentence. It can execute hundreds of steps in a single time, and up to 40 finished design pictures are produced; images, videos, and music are freely scheduled, and a variety of cutting-edge AI models are integrated.
OpenAI launches Codex programming intelligent body
OpenAI launches cloud-based AI intelligent bodyCodexfocusing on automated software development, supporting tasks such as coding and error repair in parallel.
Codex is based on the codex-1 model and is currently in the form of research previewChatGPTThe platform is online, and Pro, Enterprise and Team users can take the lead in experience.
Other news
QwenChatThe web page development function is officially launched. Users only need to enter a natural language command to generate website page codes with clear structure and beautiful style, which supports direct preview.
LikeAchieve the latest cooperation with Xiaohongshu, and users can enter Kimi through Xiaohongshu official account @Kimi Smart AssistantaiConversation and generate notes with one click.
Kuaishou launches AI mapping toolsPoofyincluding AI model fitting, background change, one-click cutout, partial repainting and other functions, empowering creativity and transformation in the e-commerce field.
OpenAI upgradeChatGPT AI Chatbot，GPT-4.1Programming expert model appears. GPT-4.1 can accurately follow instructions and avoid verbose output. It is more efficient, more stable output, and has better overall performance than GPT-4o.
ManusAnnounced full open registration, and users can perform one task for free every day. The image generation agent function was then launched, which can call a variety of tools to work together.
MiniMaxquietly launching manus products, universal AI intelligent platformMiniMax AgentAbility to efficiently and conveniently handle a variety of tasks, such as audio generation, market research, document verification, travel planning, social media analysis, patent identification, investment strategy analysis, and creative content creation.
MiniMax releaseTTS system MiniMax Speech 02Her Moment of AI voice, can easily control vocals in 32 languages, different accents, and different emotions.
Clear wordsThe intelligent community is launched, and users can directly talk to creators and post to communicate. The app supports one-click import of WeChat files.
ChatGPTIn-depth researchThe export as PDF function is added, which can retain the report format. In addition, a GitHub connector is added to further expand ChatGPT’s application capabilities in code management and writing.
NotionLaunch AI Conference NotesAI Meeting Notesable to transcribe the meeting content in real time, and automatically generate a complete summary containing key summary, user notes comparison and action items after the meeting.
Tongyi Qian QuestionsLaunch new Deep Research! Based on the Qwen model, it integrates multiple capabilities to plan complex research tasks, combine massive Internet information, search, analysis and summary through multiple steps, and generate detailed reports in more than ten minutes, with clear citation sources.
Ask XiaobaiAcademic SearchHere, the function is equipped with DeepSeek R1/V3, Qwen3 and other models, which can retrieve a large number of authoritative documents.
WindsurfingLaunches a series of AI models designed for software engineering:SWE-1SWE-1-lite and SWE-1-mini. It covers development links such as code generation, testing, debugging, and data search to improve development efficiency and user experience.
Xindi DigitalRelease three products, Xinditiangong CAD 2025, Xinditiangong Cloud CAD 2025, and Xindi 3D Lightweight Engine 2025, deeply integrating AI technology into three-dimensional design software.
INAIR releases new smart products—INAIR AI Space Computerplanned to be officially released in June, equipped with a system-level AI assistant, supporting multi-windows same-station display and cross-device access.
Alibaba open source Tongyi Wanxiang Wanxiang Wan2.1 – VACE model, one model solves multiple video editing tasks
Alibaba officially opens the source of the “Tianyi”Wan2.1-Vace”, supports the complete series of basic generation and editing capabilities such as Wensheng video, image reference video generation, video repainting, video partial editing, video background extension, and video duration extension.
Kunlun Wanwei officially open source Matrix-Game: Building a controllable interactive world based on images
Kunlun Wanwei Open SourceMatrix-GameThe big model is the first open source 10B+ space intelligent big model in the industry. It is an interactive world basic model specially modeled for the game world.
ByteDance’s open source 8B code model Seed-Coder helps efficient programming
ByteDance Open Source 8B Code ModelSeed-Coderwhich includes three versions: Base, Instruct and Reasoning, which are suitable for code completion, directive compliance and complex reasoning tasks respectively.
Multimodal new, Step Star Open Source 3D Large Model
Step Yuexingchen officially released and open source 3D model-Step1X-3Dthe total parameter volume reaches 4.8B, consisting of geometric module 1.3B and texture module 3.5B, which can generate high-fidelity and controllable 3D content.
Meituan invests exclusively in the A round, and the independent variable robot “Embroidered Intelligence” company has raised more than 1 billion yuan
The independent variable robot recently completed a round A financing of hundreds of millions of yuan led by Meituan and followed by Meituan Dragon Ball. Less than a year and a half since its establishment, it has completed seven rounds of financing, with a cumulative financing amount exceeding 1 billion yuan. This financing will be used to accelerate the iteration of models and robots and the implementation of intelligent solutions.
Embodied intelligent newcomer “Lingyu Intelligent” completed a seed round of financing of tens of millions, led by Inno Angel Fund
The domestic embodied intelligent enterprise “Lingyu Intelligent” announced that it has completed a seed round of financing of tens of millions. This round of financing was led by Inno Angel Fund and followed by Shuimu Alumni Seed Fund and Telescope Venture Capital. Lingyu Intelligence was founded by the team of the Department of Automation at Tsinghua University, promoting the widespread application of robots in the fields of industrial manufacturing and home services.
“Aoyi Technology” has completed a B++ round of financing of nearly 100 million yuan, and the second generation of dexterity hands will be launched soon
“Aoyi Technology” has recently completed a B++ round of financing of nearly 100 million yuan. This round of financing is jointly invested by Infiniti Capital, Zhejiang Development Asset Management Co., Ltd., a subsidiary of Zhejiang State-owned Capital Operation Co., Ltd., and Womeda Capital, and Shengkr Capital serves as long-term financial advisors. This round of funds will be used for the research and development of smart hands technology, the launch of new products, capacity construction and market expansion.
Zhongke Zidong Taichu completes its first round of financing of hundreds of millions of yuan
Google establishes AI venture capital fund
Google established an AI venture capital fund – “AI Future Fund” to invest in startups. Selected startups can receive investment, early access to Google’s AI mockups and professional team guidance.
AI AI ETF received net financing purchase
AI Artificial Intelligence ETF (512930.SH) has net financing purchases of 7.7839 million yuan, ranking first in the market. It has received net financing purchases in 3 days in the past five trading days.
AI inference model has the potential to generate knowledge independently
Jakub Pachocki, senior model development leader at OpenAI, believes that AI inference models are showing the potential of autonomous knowledge generation, and substantial progress will be made in autonomous AI research at the end of this century, and nearly autonomous software development systems may appear this year.
AI investment will drive structural financing
Steven Moffett and John Greenwood of Goldman Sachs Group believe that companies seeking a large amount of funds for AI-related projects has created financing demand in the structured financing market, and more companies will benefit from this financing tool in the future.
Wikipedia announces new AI strategy: Utilizing AI to improve volunteer time utilization efficiency, rather than replacing human
Original link:AI Weekly丨AI industry trends that cannot be missed this week (5.12-5.18)
© Copyright Statement
