NAll news

All news

Friday160 items

  1. 36Kr · AI
    AI score74

    SemiAnalysis's hands-on testing claims Anthropic is running an A/B gray test on Claude subscription quotas

    SemiAnalysis ran an Opus 5.5 agent workload on two identical $200 Max 20x accounts and found that one account's 5-hour window quota was 2.076 billion tokens, 19% higher than the other account's 1.74 billion tokens, but the weekly total quota fell from 6.579 billion tokens to 5.266 billion tokens, a 20% shrinkage. Claude Code technical staff member Lydia Hallie said the quotas were not adjusted and suggested users run the feedback command to submit their ID for investigation; core developer Boris Cherny said the team is urgently investigating. The article also mentions that OpenAI previously halved the quota of its $200 Pro plan and introduced a $500 tier, as well as DeepSeek's open-source inference framework DeepInfer and its deployment practice of 155 trillion tokens per day of throughput on a single cluster.

    Same story · 2 reports

    #English text is being prepared.#Industry news#Anthropic

  2. 36Kr · AI
    AI score62

    Kimi K3.1 suspected to appear on Arena, official video reveals model name

    Kimi officially released a 24-second video, in which the model field of Kimi Code shows Kimi K3.1; previously an anonymous model Kibitz appeared on Arena and was suspected by netizens to be K3.1. Kibitz performed well in code tests such as 3D courtyard and pelican SVG, and a model-tracking account judged through Tokenizer features that it belongs to the Moonshot AI Kimi model family. Code commits in the SGLang repository show that K3.1 uses the same Tokenizer and 16 special Tokens as K3, and may add a new medium reasoning intensity tier.

    #English text is being prepared.#Agents#English text is being prepared.

  3. The Beijing News
    AI score58

    MIIT: Full-year humanoid robot complete-unit output expected to exceed 100,000 units

    Wei Wei, Director of the Department of Science and Technology of the Ministry of Industry and Information Technology, said at a State Council Information Office press conference that in the first half of this year China's humanoid robot complete-unit products exceeded 400 models with output surpassing 40,000 units, and full-year output is expected to exceed 100,000 units. Related products have been validated and deployed in more than 300 key scenarios including warehousing and logistics, catering and retail, and home services. Next steps will advance key technology breakthroughs, scenario opening, and release the "National Humanoid Robot Industry Standards System Construction Guide (2026 Edition)".

    #Industry news#Embodied AI#English text is being prepared.

  4. TMTPost · AGI
    AI score53

    Kimi K3.1 suspected to have appeared on Arena; official video reveals model name

    Kimi officially released a 24-second video, in which the model field of Kimi Code shows Kimi K3.1; previously an anonymous model Kibitz appeared on Arena, which tracking accounts judged to have characteristics of the Moonshot AI model family. Commits in the SGLang repository show that K3.1 uses the same Tokenizer and 16 special Tokens as K3, and adds a medium reasoning intensity tier. K3 was previously priced at 3 USD per million input Tokens and 15 USD per million output Tokens, while Claude Opus 5.5 has dropped to 4 USD and 20 USD.

    #English text is being prepared.#Agents#English text is being prepared.

  5. Semiconductor Industry Observation
    AI score55

    The Central Cyberspace Affairs Commission issued the "15th Five-Year Plan for National Informatization," proposing to improve advanced-process chip manufacturing capabilities

    The Central Cyberspace Affairs Commission issued the "15th Five-Year Plan for National Informatization," proposing that by 2030 China's informatization will achieve major progress, and explicitly stating that mature-process chips should be refined and advanced-process manufacturing capabilities improved, and that breakthroughs and applications of technologies such as compute-in-memory integration, 3D integration, and optoelectronic fusion should be advanced. The plan also proposes promoting the integrated development of operating systems, software development toolchains, and other areas with artificial intelligence, promoting research on artificial intelligence theories and algorithms such as foundation models, brain-inspired intelligence, and swarm intelligence, and cultivating internationally leading technology products in fields such as integrated circuits, new displays, artificial intelligence, and blockchain.

    Same story · 7 reports

    #English text is being prepared.#English text is being prepared.

  6. Caixin
    AI score68

    PHP exposed two high-severity vulnerabilities enabling unauthenticated remote code execution, affecting about 70% of websites worldwide

    On October 21, the official PHP website stated that two critical vulnerabilities had been discovered that enable unauthenticated remote code execution, codenamed PHP2Shell in the industry, and it is estimated that about 70% of websites worldwide are built with PHP. OpenAI said in a post the same day that the vulnerabilities were discovered by AI security company Calif with the help of GPT-5.6 Sol; frontier AI discovered this batch of vulnerabilities before attackers, buying a defensive window for maintainers and security service providers. The China National Vulnerability Database (CNVD) recommends that affected users immediately upgrade to the latest version.

    #Safety / alignment#OpenAI#Industry news

  7. gihyo.jp
    AI score62

    AAIF takes over four projects including Anthropic's Agent Skills

    The Agentic AI Foundation (AAIF) under the Linux Foundation announced on October 22, 2026, that it has taken on four projects: Anthropic's Agent Skills, OpenAI's Open Responses, Hugging Face's OpenEnv, and Alibaba's OpenSandbox. Agent Skills is a general-purpose format for enabling different agents to reuse business knowledge and workflows; Anthropic made it public as an open standard in December 2025, and according to the company it has already been adopted by more than 70 platforms including ChatGPT, Cursor, GitHub Copilot, and VS Code. The donation includes the Agent Skills name, logo, GitHub organization account agentskills, and the agentskills.io domain; going forward, a technical operations committee composed of core maintainers from multiple organizations will decide the development direction, with Anthropic's David Soria Parra and Mahesh Murag serving as lead maintainers. The number of projects under AAIF has grown from three at its founding in December 2025—MCP, goose, and AGENTS.md—to ten.

    Same story · 6 reports

    #Industry news#Agents#Open-source ecosystem

  8. Yicai
    AI score58

    Unitree Technology holds partner conference, launches $10 million Physical AI competition and ecosystem incubation program

    Unitree Technology held its first partner conference in Hangzhou, announcing annual support of $10 million for the World Physical AI Competition and launching the "Yuchuang Program" for startup teams, with an estimated collaborative scale of several billion yuan over the next five years and plans to incubate more than 200 ecosystem companies. Wang Xingxing said Unitree currently produces tens of thousands of robots per year, and with large-scale adoption annual output could reach at least ten million units; the company is also developing its own models and cooperating with domestic and international large-model companies. The day after the conference, Unitree Technology's stock hit the 20% daily limit, closing at 624 yuan per share, with total market value returning to 250 billion yuan.

    Same story · 8 reports

    #Industry news#Embodied AI#Agents

  9. Rohan Paul
    AI score71

    DeepSeek and MiniMax continue to open-source models after completing registration under China's new AI export rules

    DeepSeek and MiniMax completed registration under China's new AI export rules and then continued to open-source models. Since October 16, China has required AI companies to register with the government before sending regulated AI technology overseas, and publicly releasing model weights or code for global download (i.e., open-sourcing) is also regarded as such export, so open-source releases now also require prior filing.

    Same story · 4 reports

    #English text is being prepared.#Open-source ecosystem#DeepSeek

  10. Cailian Press
    AI score56

    Anthropic report projects AI could push US unemployment rate close to 12%, industry disputes the logic

    A report released by Anthropic in September this year projected an extreme scenario: with large-scale AI adoption, by 2030 US GDP would jump sharply, the share of capital income in national income would rise from about 40% to nearly 55%, and the unemployment rate would rise to close to 12%. Macro economist Tom Orlik and others question the logic of this projection, arguing that simultaneous surges in growth and unemployment are unrealistic in economic history. Another economic research team, measuring by the share of tasks in each occupation where AI can play a substantive role, estimates that about 26% of the US labor force (about 45 million people) faces impact, with white-collar positions such as financial practitioners hit first.

    #Industry news#Anthropic#English text is being prepared.

  11. Yicai
    AI score62

    State Council General Office issues opinions, deploying 18 tasks across 6 areas to address the employment impact of artificial intelligence

    The General Office of the State Council issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology to Help Build an Employment-Friendly Development Model," deploying 18 major tasks across 6 areas. The opinions propose expanding new space for AI employment in emerging fields and accelerating the development of new occupations in the AI field, while requiring a steady and prudent grasp of the scope, intensity, and pace of AI application in labor-intensive industries, and establishing a mechanism for assessing the employment risk of AI applications.

    Same story · 15 reports

    #English text is being prepared.#Industry news

  12. Wallstreetcn
    AI score58

    The Central Cyberspace Affairs Commission issued the "National Informatization 15th Five-Year Plan," proposing the development of high-performance AI chips and the timely launch of 6G commercialization

    The Central Cybersecurity and Informatization Commission issued the "National Informatization 15th Five-Year Plan," proposing that by 2030 a new-generation communications network be fully built and 6G commercialization be launched in due course. The plan explicitly calls for developing high-performance artificial intelligence chips and a highly available basic software stack, accelerating the construction of a nationally integrated computing power network, promoting the coordinated development of computing power and power networks, and laying out embodied intelligence technologies such as humanoid robots. The plan puts forward a total of 12 main indicators, 25 key tasks in 10 areas, and 7 special columns.

    Same story · 7 reports

    #English text is being prepared.#English text is being prepared.#English text is being prepared.

  13. East Money
    AI score62

    The General Office of the State Council issues opinions, deploying AI development and an employment-friendly development approach

    The General Office of the State Council issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology to Help Build an Employment-Friendly Development Approach," making arrangements for related work. The document proposes deepening implementation of the "AI+" initiative, encouraging the eastern region to lay out core AI industries, the central and western regions to take on upstream and downstream industries such as data annotation and intelligent terminal manufacturing, and leveraging the role of the National Venture Capital Guidance Fund, the National Artificial Intelligence Industry Investment Fund, and others to support AI enterprises. The document also requires promoting the human-machine collaboration model, steadily and prudently grasping the scope, intensity, and pace of AI application in labor-intensive industries, and encouraging enterprises to explore establishing an unemployment buffer period system.

    Same story · 15 reports

    #English text is being prepared.#Industry news

  14. Rohan Paul
    AI score78

    The U.S. Treasury Department proposes banning Americans from investing in Chinese companies that develop any AI system

    The U.S. Treasury Department proposes banning Americans from investing in Chinese companies that develop any AI system. Current rules only restrict investment in the most powerful Chinese AI models; smaller AI projects can still receive funding after reporting to the Treasury Department. The new proposal would comprehensively ban such AI investment and add high-bandwidth memory, photonic chips and radiation-hardened chips to the banned chip categories. The proposal also adds two prohibited investment areas, hypersonic weapons and biotechnology (including brain-computer interfaces and automated laboratory software), and expands the restricted countries from China to Russia, Iran, North Korea and Cuba; purchases of public stocks, index funds or small fund shares of no more than $2 million are still allowed, so it mainly affects venture capital and private equity deals, and the Treasury Department will not publish a banned list, so investors must check each transaction themselves.

    Same story · 2 reports

    #English text is being prepared.#English text is being prepared.#OpenAI

  15. Pangpai News
    AI score62

    Meta's AI spending surge triggers another selloff in US tech stocks, Wall Street questions whether the market can absorb the massive investment

    Pangpai News reporter Hu Hanyan reports that on October 22 local time, after Meta released its earnings report, US tech stocks suffered another sharp selloff: Meta shares fell 6.98% to close at $648.85 per share, Nvidia fell 4.4%, and the Philadelphia Semiconductor Index closed down 6.72%. Meta's third-quarter revenue beat expectations, but the lower bound of its full-year capital expenditure forecast was raised again, free cash flow turned negative for the first time, and its total future spending commitments rose from about $700 billion in July to $825 billion; its 5-year credit default swap quote hit a record high. Citadel founder Ken Griffin said the current AI boom has almost certainly formed a capital bubble, while Blackstone President Jon Gray believes demand is growing almost parabolically and there is no need to worry.

    #Industry news#OpenAI

  16. Shanghai Observer
    AI score68

    The General Office of the State Council issued opinions, deploying an employment-friendly development model that adapts to AI development

    The General Office of the State Council issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology and Helping Build an Employment-Friendly Development Model," deploying 18 major tasks in 6 areas. The tasks include strengthening AI's job-creation effect, building an employment-friendly path for AI upgrading, improving workers' digital literacy and employability, improving worker rights protection and the social security system, and improving the AI application employment monitoring, early warning and risk response system. The document requires human resources and social security departments and development and reform departments at all levels to strengthen overall coordination and promote policy implementation.

    Same story · 15 reports

    #English text is being prepared.

  17. Cailian Press
    AI score65

    State Council General Office issues opinions deploying 18 tasks in 6 areas to build an employment-friendly AI development approach

    The General Office of the State Council issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology and Helping Build an Employment-Friendly Development Approach," deploying 18 major tasks in 6 areas. The document proposes expanding new space for AI employment in emerging fields, accelerating the development of new occupations in the AI field, and establishing an employment risk assessment mechanism for AI applications, conducting employment risk assessments in tandem during review for application scenario projects that may cause a relatively large employment shock. The document also requires steadily and prudently grasping the scope, intensity, and pace of AI application in labor-intensive industries, and encourages enterprises to explore establishing an unemployment buffer period system.

    Same story · 15 reports

    #English text is being prepared.#Industry news

  18. Latent Space
    AI score83

    OpenAI says Patch the Planet used GPT-5.6 Sol to discover the PHP2Shell vulnerability

    OpenAI says researchers on its Patch the Planet project used GPT-5.6 Sol to discover the PHP core memory corruption vulnerability PHP2Shell (CVE-2026-109618 and CVE-2026-109619); Wiz independently discovered the same set of vulnerabilities, and the two parties disclosed it publicly only after about two months of coordinated remediation. The researchers demonstrated unauthenticated remote code execution against WordPress, Drupal, and Joomla; Cloudflare deployed free WAF mitigation rules before the patch release and called them a temporary measure. The original text also notes that key information such as the affected and fixed version matrix, leaked details, and the actual scale of exploitation still cannot be verified in the available materials.

    Same story · 6 reports

    #Safety / alignment#OpenAI#Google

  19. Yicai
    AI score55

    State Council Office document encourages eastern regions to lay out core AI industry

    The General Office of the State Council issued the "Opinions on Adapting to the Development of Artificial Intelligence Technology and Helping to Build an Employment-Friendly Development Model," proposing to encourage eastern regions to lay out the core AI industry and guide central and western regions to undertake upstream and downstream industries such as data annotation and intelligent terminal manufacturing in light of local conditions. The document also requires accelerating the cultivation of AI-native enterprises and specialized, refined, distinctive and innovative "little giant" enterprises, and including eligible enterprises within the scope of support from special loans for financial assistance to enterprises to stabilize and expand employment as required.

    Same story · 15 reports

    #English text is being prepared.#Industry news

  20. CNN Business Tech
    AI score66

    AI finds a 22-year-old PHP vulnerability within two days; OpenAI and security firms jointly disclose it and release a patch

    AI security firm Calif and OpenAI's Patch the Planet initiative used a large language model to find an exploit for a 22-year-old PHP vulnerability within two days, and the related patch was released on Wednesday. The vulnerability can cause residual data to be read after server memory is exhausted or allow malicious code to be executed, and Google-backed Wiz also discovered a new PHP exploitation method. Cloudflare said the vulnerability was suspected to have leaked before the patch was released, with thousands of exploitation attempts appearing within hours, and CrowdStrike warned that a large number of old PHP versions are still running.

    Same story · 6 reports

    #Safety / alignment#Agents#OpenAI