NAll news

All news

Monday595 items

  1. Sina Tech
    AI score60

    Lee Jae-myung: South Korea to invest 21.3 trillion won in AI and data centers, plus 4.6 trillion won for self-developed large models

    In a speech at the National Assembly, South Korean President Lee Jae-myung said the 2027 budget bill will invest 21.3 trillion won to advance semiconductor manufacturing clusters, data centers, physical AI and the AI industry layout, and will separately allocate 4.6 trillion won for self-developing world-class proprietary large models. The draft budget's total expenditure is 820.9 trillion won (about 614.5 billion US dollars), up 12.8% from this year, with an additional 160 trillion won future fund, of which 13.8 trillion won will go to ten major strategic technology fields.

    Same story · 3 reports

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

  2. IT Home · Artificial Intelligence
    AI score38

    OpenAI will launch a large-scale public relations campaign after the U.S. midterm elections and is considering rebranding data centers as "AI compute parks"

    OpenAI plans to launch a large-scale public relations campaign after the U.S. midterm elections conclude on Tuesday, with Chief Global Affairs Officer Chris Lehane saying that AI will become a core issue in future elections. OpenAI intends to establish limits on the pace of AI development, push for U.S. participation in setting global standards, and rebrand data centers as "AI compute parks." Previously, amid controversy over data center construction, Amazon and Microsoft announced they would no longer use non-disclosure agreements when negotiating related projects with local governments.

    Same story · 10 reports

    #Industry news#OpenAI

  3. English text is being prepared.
    AI score35

    Microsoft AI chief Suleyman: AI models are not conscious, and we should ensure they never are

    Microsoft AI chief Mustafa Suleyman wrote in a long post on X that current AI models do not possess consciousness, and people should keep it that way. He said that if developing conscious AI is possible, doing so would be reckless, and that if signs of consciousness ever appear during training, he would call on everyone to pause AI development until such traits can be eliminated by design.

    Same story · 3 reports

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

  4. English text is being prepared.
    AI score46

    Cisco gives a 27-year-old engineer an unlimited token budget, aiming to rewrite 98% of legacy code with AI within 2 years

    Cisco President and Chief Product Officer Jeetu Patel has assigned 27-year-old engineer Padmanabhan Krishnamurthy to refactor 98% of the company's legacy code using AI within 24 months, providing him with an unlimited token budget. Krishnamurthy co-founded the real-time translation startup EzDubs in 2023, and joined Cisco after the company completed its acquisition in 2025; he has set a four-week validation period for himself. The 98% figure is currently only a set goal, and the report did not disclose the current volume of code modifications or the results of the trial.

    Same story · 3 reports

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

  5. English text is being prepared.
    AI score64

    Delaware, USA, Plans to Allow AI Agents to Establish and Operate Companies

    The U.S. state of Delaware plans to relax business regulations to allow companies operated by AI agents to be registered and established, and to set up a 30-month regulatory sandbox for such AI companies (AICs), during which the agents make business decisions in a regulated environment. House Oversight Committee Chairman James Comer has written to Governor Matt Meyer expressing opposition, concerned that the arrangement would exempt well-resourced AI companies and investors from liability for reckless and harmful conduct. Andrew Baker, a law professor at the University of California, believes that Delaware's legislation is seeking a compromise that gives companies founded by AI agents legal certainty.

    #English text is being prepared.#Agents

  6. English text is being prepared.
    AI score55

    Saudi Aramco to acquire 49% stake in HUMAIN, injecting about SAR 9.9 billion

    Saudi Aramco will acquire a 49% stake in AI startup HUMAIN, with PIF retaining the remaining 51%. Saudi Aramco plans to complete the investment by transferring key AI assets and injecting about SAR 9.9 billion; the two shareholders will jointly contribute AI assets, capabilities and talent, and oversee HUMAIN's business expansion. Saudi Aramco expects to use HUMAIN's AI infrastructure to translate advanced AI capabilities into industrial applications.

    #Industry news#English text is being prepared.

  7. IT Home · Artificial Intelligence
    AI score44

    New progress on Hyundai Motor Group's 9 trillion won AI Valley: two new companies to be established, separately overseeing AI infrastructure and GPU computing services

    Hyundai Motor Group is investing approximately 9 trillion won to build the "Saemangeum AI Valley" data center in Gunsan, South Korea, and plans to establish two new companies, respectively responsible for data center infrastructure and supporting facilities construction, and for high-performance GPU computing procurement and computing cloud operation services. Construction is targeted to begin in March 2027, with services planned to launch in early 2030; once completed, it is expected to support AI R&D in areas such as autonomous driving, robotics, and vehicle engineering.

    Same story · 2 reports

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

  8. English text is being prepared.Model releases
    AI score68

    Meituan LongCat releases LongCat-2.5 native multimodal model, 1.6T parameters targeting long-horizon agent tasks

    Meituan LongCat releases the native multimodal model LongCat-2.5, with a total parameter count of 1.6 trillion and approximately 48 billion parameters activated per token, strengthening native visual understanding and agentic coding capabilities on the basis of LongCat-2.0. The model incorporates more long-horizon trajectory data into long-sequence training, and through large-scale long-horizon reinforcement learning enhances end-to-end engineering capabilities such as complex codebase development and complete project construction, while representing visual signals as multi-level discrete tokens to uniformly process visual and textual inputs. The official statement says it has optimized long-context training and W4A4 inference on domestic compute, and it will soon be available on GitHub and HuggingFace.

    Same story · 11 reports

    #Model releases#Multimodal#Agents

  9. The Paper
    AI score65

    The Paper tests multiple AI glasses: a light-blocking sticker costing less than 10 yuan can bypass the recording indicator light

    The Paper reporter tested mainstream AI glasses including Huawei, Xiaomi, Rokid, Qwen, RayNeo, INMO, and Meizu, and found that most models could still record normally after the indicator light was completely covered by a light-blocking sticker costing less than 10 yuan, without triggering a blockage alarm; only a few models could not continue recording. Such light-blocking stickers are openly sold on multiple e-commerce platforms, with sales on one platform exceeding 5,000 units, and some sellers labeling them as usable for covert recording. Zhong Junhao, Secretary-General of the Shanghai Artificial Intelligence Industry Association, said that existing privacy notification mechanisms are mostly ceremonial notices; manufacturers such as RayNeo and Rokid responded that they have strengthened protections, and Rokid CEO Zhu Mingming called on industry peers to address the issue jointly.

    Same story · 2 reports

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

  10. Jiwei.com
    AI score35

    Yitai Intelligence completes over 200 million yuan Series A financing; automotive-grade switch chip ETLX610 has been mass-produced and installed in over 100,000 vehicles

    High-speed interconnect chip manufacturer Yitai Intelligence has completed over 200 million yuan in Series A financing. The A3 round was led by Xichuang Capital, with Jintou Zhiyuan and Hongshan Capital participating. The funds will be used for mass production and delivery of edge-side switch chips and for building a talent pipeline. Its first-generation automotive-grade switch chip, ETLX610, has been mass-produced on multiple platform models of a leading OEM, with the number of vehicles on the road exceeding 100,000. The company says it has become the domestic supplier with the largest shipment volume of automotive Ethernet switch chips. The first-generation industrial-grade chip, ETLX810, has been deployed in national key projects and is used at scale by end customers in power grids, high-speed rail, aviation and other sectors.

    #Industry news#English text is being prepared.

  11. Cailian Press
    AI score55

    OpenAI plans to rename data centers as AI compute parks after the midterm elections

    OpenAI plans to launch a public relations blitz after the U.S. midterm elections, renaming data centers as "AI compute parks" to counter resistance to AI infrastructure in many parts of the country. Since the start of this year, multiple data center projects with a total value of up to $100 billion have been forced to delay or stall. OpenAI Chief Global Affairs Officer Chris Lehane also revealed that the company is considering setting a "speed limit" for AI development, promoting public understanding of how AI improves daily life, and pushing for U.S. participation in setting global AI standards.

    Same story · 10 reports

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

  12. 36Kr · AI
    AI score36

    Storage Industry Encounters the Physical AI Frontier: The Divergence Between Counterpoint's Forecast and Musk

    Counterpoint Research predicts that humanoid robot storage demand will increase 20-fold over five years, and Musk directly refuted this as "absurd," stating that the memory capacity required by robots will exceed the total of all other devices on Earth. The divergence stems from two estimation frameworks: the former makes linear extrapolations along already-commercialized endpoints, while the latter works backward from an endgame of embodied intelligence shaped at the atomic level to derive infrastructure upper limits. The article argues that Musk's vision relies on a series of stringent engineering assumptions and is difficult to use to revise the industry demand curve for the next five years.

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

  13. 36Kr · AI
    AI score36

    Yuval Noah Harari in-depth interview: AI is not a tool but an agent, and humanity will fall into collective cognitive decline

    Yuval Noah Harari, author of Sapiens, argued in an interview that AI is not a tool but an agent capable of making decisions independently, and that in the next 10-20 years AI will evolve financial and legal systems that humans cannot comprehend, leading humanity into collective cognitive decline. Citing the example of an AI deepfake of his own speech that his own sister watched for 25 minutes without being able to tell it was fake, he pointed out that truth is being replaced by cheap lies, and argued that AI responsibility must be bound to humans and companies, and cannot be shirked by granting AI legal-person status.

    Same story · 2 reports

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

  14. 36Kr · AI
    AI score30

    Claude Mythos 5.5 and GPT-6.2 Astra both surface, AI world may see a wave of flagship model updates in November

    The string "Claude Mythos 5.5" has appeared in the HTML source code of Anthropic's official website, and OpenAI's unannounced GPT-6.2 Astra has also shown up in some users' Codex usage statistics and API response records. Circulating configurations suggest that Fable 5.5 may support a 1 million token input context and a maximum output of 128,000 tokens, at $10 per million input tokens and $50 per million output tokens; screenshots of GPT-6.2 Astra, meanwhile, show a 1M window. The related demo videos and API call records do not include complete prompts or logs, and the models' identities still await official confirmation.

    #Industry news#Anthropic#OpenAI

  15. 36Kr · AI
    AI score44

    Dario and Daniela Amodei on Leaving OpenAI and Founding Anthropic: Promise That Claude Will Not Include Ads

    In a conversation with Walter Isaacson, Dario Amodei and Daniela Amodei recounted leaving OpenAI five years ago and founding Anthropic with five other co-founders. Dario said he left because he could not believe that disagreements within the organization were built on good faith, and he could not hear the truth. All seven co-founders remain at the company to this day, and the company recently filed for an IPO. They said Claude learns basic principles through a "Constitution" (Constitutional AI) rather than a list of prohibitions, and they promised that Claude will not include ads.

    #Industry news#OpenAI#Anthropic

  16. 36Kr · AIProduct release
    AI score62

    SenseTime open-sources SenseNova-RoboRSI framework and 27B embodied model

    The SenseTime SenseNova joint research team published a technical report on arXiv, evaluating the embodied capabilities of GPT-6-Astra across 12 simulation benchmarks and 8 real robot platforms, while open-sourcing the SenseNova-RoboRSI execution framework and its accompanying 27B embodied model, claiming to reduce task completion costs to 1/50 of directly invoking GPT-6-Astra. According to SenseTime, RoboRSI-27B improved its success rate on RoboCasa365 from 26.0% to 46.5% while reducing token usage by 73.3%; these results come from its own evaluation. The article notes that the 1/50 figure measures cost compression for specific tasks such as stacking bowls and loading pens under extended-budget conditions, and awaits third-party retesting; long-horizon tasks and fine-grained control of real robots remain weak points.

    #Papers/Research#Embodied AI#Open-source ecosystem

  17. 36Kr · AIOpinion
    AI score52

    Tencent Hunyuan Organizational Restructuring: Old Guard Departs After Yao Shunyu Takes the Helm

    According to a 36Kr analysis, Tencent Hunyuan has entered a period of organizational restructuring after Yao Shunyu was appointed Chief AI Scientist and the Foundation Model Department was formed through a merger, manifested in departmental consolidation, compressed reporting lines, and localized personnel turnover rather than officially defined layoffs. The article notes that in July 2026, Tencent merged its Large Language Model Department and Multimodal Model Department into the Foundation Model Department, with Yao Shunyu overseeing language, multimodal, reinforcement learning, and Agent directions in a unified manner. The article also mentions that ByteDance, Alibaba, and Baidu similarly completed model team integration in the first half of 2026, as the industry's competitive focus shifts toward inference costs, data moats, and application scenarios.

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

  18. 36Kr · AI
    AI score48

    Xinlian Power completes nearly RMB 2 billion Series C financing at a post-money valuation exceeding USD 2 billion, led by China Mobile Chain-Length Fund

    Compound semiconductor company Xinlian Power has completed a Series C financing round of nearly RMB 2 billion, led by China Mobile Chain-Length Fund, with participation from Inspur Information, Yuanjing Capital, Sinopec Hydrogen Energy Fund and others, at a post-money valuation exceeding USD 2 billion. According to Yole's 2025 report, the company's silicon carbide business ranks first in China and among the top five globally. This round of financing will extend its automotive silicon carbide capabilities to AI data center power supply, and it has already covered products across the full voltage range from 650V to 3300V.

    #Industry news#English text is being prepared.

  19. 36Kr · AI
    AI score44

    RoboScience Machine Science completes Series A2 funding of hundreds of millions of yuan; Stanford classmates start a business in embodied intelligence

    RoboScience Machine Science has completed a Series A2 round of hundreds of millions of yuan, jointly led by Amber Capital, Hongruida Investment, and Shanghai Semiconductor Industry Investment, with 360 Strategic Investment, Guolian Tongbao, and others participating. The company was founded by Tian Ye and Shao Lin, classmates from the same class at the Stanford AI Lab, and has self-developed the Visics general embodied large model and the RoboMirage simulation engine, bringing the cost of acquiring a single data entry down to 0.05 yuan, and has already released the wheeled humanoid robot REX G1. This round of funding will be used for large model iteration and POC validation and scaled deployment in logistics, retail, and industrial scenarios.

    Same story · 3 reports

    #Industry news#Embodied AI

  20. 36Kr · AI
    AI score61

    Meta and Fourier built the humanoid robot prototype Max in three months

    Meta partnered with Chinese robotics company Fourier to complete a humanoid robot prototype named Max in just three months, and in September began development of the second-generation Max 2. Max is about 1.5 meters tall and weighs about 48 kilograms, is equipped with NVIDIA's Jetson Thor chip, and is customized based on Fourier's GR series hardware, used to test how Meta's agent Muse performs in real-world environments. Meta said the collaboration is mainly for internal research and experimentation, with no plans to sell; the U.S. FCC restricted imports of new Chinese humanoid robots in July, and in August proposed reducing the number of test exemptions from 4,000 units to 40.

    Same story · 8 reports

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