NAll news

All news

Monday537 items

  1. Alibaba Cloud Developer Community AI
    AI score38

    GEO Engineering Practice Analysis: How to Build a Content System That AI Can Understand, Trust, and Cite

    Generative Engine Optimization (GEO) adjusts the semantic structure, information hierarchy, and authority signals of content to make it easier for large language models to crawl, understand, and cite. Papers from Princeton University and other institutions (arXiv:2311.09735, KDD 2024) show that adding credible external source citations to content can increase relative visibility by up to approximately 115.1%, and adding specific statistical data and numbers can increase it by up to approximately 40%. The article also presents six major strategies, including extractability, structured markup, and multi-source cross-validation, and points out that GEO and SEO are complementary rather than substitutive.

    #Tutorials / practice#RAG#Search

  2. The Beijing News
    AI score37

    RoboScience completes several hundred million yuan Series A2 financing to accelerate iteration and commercialization of embodied large model Visics

    RoboScience announced on November 2 that it has completed a Series A2 financing round of several hundred million yuan, jointly led by Amber Capital, Hongruida Investment, and Shanghai Semiconductor Industry Investment. The funds will be used for continuous iteration of its embodied large model and for POC validation and large-scale deployment in scenarios such as logistics, retail, and industry. Founded in December 2024, the company has launched the physics simulation engine RoboMirage, the general-purpose embodied large model Visics, and the wheeled humanoid robot REX G1.

    Same story · 3 reports

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

  3. Beijing News
    AI score40

    Meta and Fourier reportedly jointly develop humanoid robot Max; Fourier declines to comment for now

    Regarding reports that Meta and the Chinese robotics company Fourier are jointly developing a humanoid robot, Fourier told the Beijing News Shell Finance that it would not comment for now. According to The Information, citing people familiar with the matter, the two sides jointly developed a humanoid robot prototype codenamed Max and have already begun development of a second-generation product, Max 2. The collaboration is mainly for internal research and experimentation, and there are currently no plans to sell the related prototypes.

    #Industry news#Embodied AI#Meta

  4. Yicai
    AI score32

    How to Avoid Uneven Benefits and Risks in the AI Era: Boston University Scholars on Protecting Jobs and People-Centered Governance

    Scholars Wang Yan and Wu Yijun of Boston University's Global Development Policy Center propose that AI governance must incorporate both technological safety and worker protection into its scope, avoiding "K-shaped growth" in which benefits and risks are asymmetric. Taking as their entry point the fact that the "Joint Commitment on Frontier Responsibility" signed by six technology companies on September 29, 2026 did not address employment and income distribution, the article advocates prohibiting unverified autonomous AI execution in critical areas and explores policy ideas such as an "agent social security tax."

    #English text is being prepared.#English text is being prepared.#Safety / alignment

  5. Yicai
    AI score27

    Anji Explores "AI + Manufacturing" Through Public Platforms and Enterprise Scenarios

    Anji County plans to drive more than 100 enterprises in the green home furnishing industry to apply artificial intelligence by 2027, build more than 8 industrial intelligent agents, and raise the share of cognitive decision-making application scenarios to over 30%. Relying on self-developed large models, screen semantic understanding, and RPA execution capabilities, the locality enables AI to retrieve data and fill out forms across systems without modifying existing systems or opening interfaces, forming an empowerment network of "public platforms + specialized platforms + enterprise scenarios." Anji has currently released 18 artificial intelligence application scenarios, and the digital transformation 2.0 coverage rate of above-scale enterprises has reached 87%.

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

  6. Ayi AI NotesProduct release
    AI score76

    DeepSeek open-sources the dsh-superagen multi-agent development project

    DeepSeek has open-sourced a new project, dsh-superagen, which acts as a bot residing in GitHub repositories and Feishu groups, automatically capturing requirements from Issues and group chats, generating tickets, and then writing code, running tests, and submitting pull requests on its own. Under the hood, the project uses a multi-agent architecture in which a main router dispatches tickets and multiple parallel worker nodes execute them; each node is physically isolated via its own Git worktree. In testing, it cut the time for complex tasks from 29.5 minutes to 15.7 minutes. Review comments on pull requests are automatically subscribed to and routed back to the original worker node, with no need for repeated manual summoning.

    Quoted postAYi · @AYi_AInotes

    English text is being prepared.

    Same story · 2 reports

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

  7. indigo
    AI score29

    indigo: Profit margins of non-Mag 7 S&P 500 companies have not risen in a decade

    indigo pointed out that the profit margins of the 493 companies in the S&P 500 excluding the Mag 7 have barely risen since 2015, with growth being captured by the leading companies. He believes that the leading companies monopolize computing power, talent, and capital, and reinvest their profits into computing power and R&D, forming an accelerating cycle; after new companies such as Anthropic go public, the Mag 7 may expand into a Mag 8 or Mag 10. He judges that companies unrelated to AI will find it hard to rise further unless their corporate structures are transformed by AI.

    #English text is being prepared.#Industry news

  9. Ethan Mollick
    AI score48

    Mollick reviews SemiAnalysis's new benchmark GDPx

    SemiAnalysis released the GDPx benchmark to measure agents' capabilities in high-value domains, and the highest-scoring model scored only 12%. Ethan Mollick considers it a worthy successor to GDPval, but notes that it lacks a human baseline like GDPval has; he also mentioned that GPT-6 Pro found some errors in the scoring, and believes that the 0% scores obscure economically meaningful differences in success rates.

    Quoted postSemiAnalysis · @SemiAnalysis_

    Today we’re releasing GDPx, a benchmark for measuring agentic capabilities across the most valuable domains. GDPx is a true frontier-level benchmark: even the best model scores just 12% (1/7)🧵 https://t.co/RNnkcuUkih

    #Evaluations / benchmarks#OpenAI

  10. East Money
    AI score44

    The National Data Administration symposium requires clarifying "data usage needs"; the physical AI data collection segment may lead the way

    On October 28, the National Data Administration held a symposium on the work of chain leader units in building high-quality industry datasets, requiring chain leader units to clarify the data needs for artificial intelligence and to adhere to "determining supply by demand, promoting construction through use, and improving quality through evaluation." Sinolink Securities stated that the data required for a mature embodied large model is at least on the order of tens of millions of hours, while current high-quality training data is only on the order of hundreds of thousands to one million hours, leaving a supply-demand gap of about 90%. Zheshang Securities estimates that the midstream data collection market will reach 174.1 billion yuan by 2030, with a CAGR of 203% from 2026 to 2030.

    Same story · 2 reports

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

  11. Phoenix Technology
    AI score78

    Anthropic says an unreleased Claude model found a Linux kernel vulnerability that lay dormant for 11 years, gaining Root in 53 minutes

    Anthropic disclosed that its internal unreleased Claude model, during systematic testing, discovered a serious Linux kernel vulnerability that had lain dormant for 11 years, CVE-2026-98432, with a CVSS score of 9.8, at the remote code execution level. The model also generated a working two-stage exploit: the first stage took about 48 minutes to defeat kernel address space randomization, and the second stage took 5 minutes to execute arbitrary commands with Root privileges, for a total of about 53 minutes. The upstream fix was completed on October 15, but as of November 1, Ubuntu 22.04 LTS, 24.04 LTS, 26.04 LTS, and 26.10 are still marked as affected, and the fix is still in development.

    Same story · 3 reports

    #Safety / alignment#Anthropic#English text is being prepared.

  12. QbitAIModel releases
    AI score50

    GPT-6.2 Astra Begins Gray Testing, Claude Mythos 5.5 Exposed on the Same Day

    GPT-6.2-Astra appeared in some users' model lists, while Claude Mythos 5.5 briefly appeared in the official website's front-end code and was removed several hours later. Reports say Mythos 5.5 is a version that includes biological research and cybersecurity, and another repository claiming to contain leaked Mythos weights was published on HuggingFace and has already been downloaded by more than 5,000 people.

    #Model releases#OpenAI#Anthropic

  13. Sina Technology
    AI score24

    First CMG World Robot Mount Tai Climbing Competition: Beijing Lingyi Technology Team Wins Footed Racing Championship in 43 Minutes 41 Seconds

    The first CMG World Robot Mount Tai Climbing Competition was held in Tai'an, Shandong. The Beijing Lingyi Technology Team won the footed robot racing championship with a time of 43 minutes and 41 seconds. The course is approximately 3.9 kilometers long with more than 3,100 steps, and nearly 90 teams participated, taking on the challenge of climbing to the Zhongtian Gate of Mount Tai. The list of winners for the humanoid robot racing competition, the footed robot load-bearing competition, the exoskeleton robot assistance competition, and the all-terrain mobile pathfinding challenge will be announced later.

    Same story · 3 reports

    #Industry news#Embodied AI

  14. Sina Tech
    AI score44

    Envision's world-first 800V SST high-voltage DC system enters commercial operation at AI data center

    Envision Technology Group's integrated 800V high-voltage DC power supply system has entered commercial operation at the Ulanqab Xinghe Base, becoming the world's first system of its kind to operate commercially in a large-scale AI data center. The system consists of the Tianshu controller, solid-state transformers, DC energy storage, and DC/DC; the Xinghe Base was put into operation in August this year, with a planned capacity of 3GW and a direct green power connection ratio set to exceed 80%.

    Same story · 2 reports

    #Industry news#English text is being prepared.

  15. Sina Tech
    AI score22

    Yixin's Auto Finance AI Smart All-in-One Machine Wins 2026 World Internet Conference "New Light" Product Award

    The first domestic auto finance AI smart all-in-one machine launched by Yixin (02858.HK) has won the "New Light" Product Award at the 2026 World Internet Conference "Light of the Internet" Expo, making Yixin the only company in the auto finance sector selected in this year's evaluation. Guided by an AI smart digital human, the device allows users to independently complete pre-approval, information collection, card binding and repayment, contract signing and other steps, compressing the traditional processing time of about 90 minutes to 15 to 20 minutes, and supports 7×24-hour service. The all-in-one machine is currently in operation at some new energy vehicle stores, with a self-service completion rate exceeding 95%.

    Same story · 2 reports

    #Industry news#Agents

  16. Sina Tech
    AI score31

    360 Showcases Nano Work and Tulongfeng at the 2026 World Internet Conference, Accelerating Its "AI Applications + Security" Layout

    At the 2026 World Internet Conference "Light of the Internet" Expo, 360 exhibited its AI office platform Nano Work and its AI security offense and defense system Tulongfeng, advancing implementation around the dual main lines of "AI + Security." Nano Work supports document editing in Word, Excel, PPT, PDF and other formats, and provides a Super Canvas and Super Video Editor for content creation. Tulongfeng adopts the approach of "security-domain large models + specialized agent swarms," has cumulatively discovered more than 15,000 vulnerabilities, and has been adopted by over 500 enterprises for specialized security testing.

    #Industry news#Agents#Safety / alignment

  17. cnBeta
    AI score47

    Kioxia's stock price plunged 17%, and the outlook for the AI storage super-cycle has changed

    Kioxia announced on Friday last week that its revenue for the second quarter of fiscal year 2026 was 2.48 trillion yen, a year-on-year increase of 453.20%, but it expects operating profit of 2 trillion yen for the December quarter, below analysts' expectations of 2.35 trillion yen, and its stock price plunged 17% on Monday. The company attributed the weak growth to shipment timing rather than deteriorating demand, and said that its planned production capacity for next year has sold out and more than half of its capacity for the year after has been allocated. The quarter-on-quarter increase in the average selling price of NAND plunged from about 70% in the second quarter to about 20% in the third quarter, which the market sees as a signal of a possible turning point in the upward cycle.

    Same story · 2 reports

    #Industry news

  18. cnBetaOpinion
    AI score58

    Microsoft AI CEO warns: AI should remain unconscious, we have no right to create a digital species

    The Microsoft AI CEO responded to the controversy sparked by his article on model welfare, emphasizing that current AI systems are not conscious and that this state must be maintained, because there is currently no evidence that AI possesses consciousness or should be granted welfare. He believes that once machines are deemed to have consciousness, it would amount to creating an entirely new digital species, which would overturn the moral order and the foundations of human civilization; therefore, he advocates designing only AI that is not conscious and does not deserve rights, with the sole goal of advancing human welfare. He also affirmed Anthropic's transparency regarding AI constitutions and public discussion, while stating that the two sides hold different views.

    Same story · 3 reports

    #English text is being prepared.#Safety / alignment#Anthropic

  19. cnBeta
    AI score38

    OpenAI GPT-6.2 Astra Exposed: Some Users' Requests Already Routed to This Model

    In some users' account usage details, "gpt-6.2-astra" has appeared separately, with about 19.8% of requests routed to this model, while the model selector still shows GPT-6 Astra. Multiple users have observed the same situation, which may be a gray-scale test. Token Gremlin claims that GPT-6.2 Astra may deliver a major performance improvement over GPT-6 Astra, even surpassing Opus 5.5 in several aspects, and may be released soon.

    #English text is being prepared.#OpenAI

  20. cnBeta
    AI score33

    Mercedes-Benz's new long-wheelbase E-Class launches: first to feature an AI cockpit developed by a Chinese team

    The 2026 long-wheelbase E-Class has launched, becoming the first to feature an AI smart cockpit developed by a Chinese team. The virtual assistant wakes up in as little as 0.25 seconds, and multi-zone interaction recognition and positioning accuracy reaches 98%. The new cockpit supports voiceprint recognition (up to 6 user voiceprints stored per account), a scenario assistant with 11 types of trigger conditions and 15 types of execution actions, and newly adds Huawei HiCar connectivity; the in-car Amap navigation frame rate has been increased to 60 fps. The new car comes in three models: the E 300 L Huiyin Edition at 423,800 yuan, the Liujin Edition at 466,800 yuan, and the Xingbo Edition at 509,800 yuan. The cockpit supports quarterly major-version OTA upgrades.

    Same story · 2 reports

    #Product updates#Speech#English text is being prepared.