NAll news

All news

Sunday264 items

  1. 36Kr · AI
    AI score58

    Meta plans to raise up to $97.5 billion; Oracle caps additional financing for this round at $20 billion

    Meta has launched a new round of equity financing, expanding the size from $35 billion to $37.5 billion; combined with a stock sale plan of up to $60 billion, the total is up to $97.5 billion, to support AI infrastructure and personal superintelligence investment. Oracle, meanwhile, announced at its investor day that additional capital market financing still needed for this investment cycle through fiscal 2030 will be limited to $20 billion, and it expects free cash flow to turn positive in the second half of fiscal 2028. Both companies use supplier financing, customer-provided hardware, and prepayments to reduce upfront outlays, but they have set out different arrangements for subsequent fundraising.

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

  2. TMTPost · AGI
    AI score26

    From Efficiency to Effectiveness: How Real AI Organizations Use Silicon-Based Computing Power to Compensate for the Limits of Human Cognition

    GritMeng argues that modern institutions are trapped in a predicament of "everyone working desperately yet canceling each other out," the root cause of which is not employees' lack of effort but the cross-mismatch of supply and demand caused by horizontal exclusivity and vertical completeness of shared resources under tightly coupled business models. The author claims that their system has operated in a fully automated closed loop for 18 consecutive years within a hundred-billion-scale industrial network, maintaining over 95% of "human-out-of-the-loop" closed loops, and advocates using a five-dimensional complex giant system to govern AI's takeover of underlying computation, leaving humans outside the loop to define problems.

    #English text is being prepared.#Agents

  3. TMTPost · AGIResearch
    AI score80

    The DeepSeek Harness team proposes the Super Agent design and open-sources a prototype

    On the evening of October 31, the DeepSeek Harness team released "Super-agent Design Exploration: Toward a More Autonomous Agent System," proposing the Super Agent concept and open-sourcing the prototype dsh-superagent-prototype on GitHub. The article splits tasks between two roles, Router and Worker, triggers work from activity such as group chats, emails, and code checks through two filters—rules plus a model—and emphasizes notifying only when results are usable or a human decision is needed. The team says the prototype is only a research prototype, not a formal product; it currently mainly targets one GitHub repository and can also connect to Feishu, and some designs have not yet been implemented.

    Same story · 23 reports

    #Papers/Research#Agents#Open-source ecosystem

  4. Analytics India Magazine
    AI score39

    Will boutique AI services firms become India's next big opportunity?

    India is seeing the emergence of a wave of "boutique" AI services firms that embed directly into enterprise workflows with AI-native engineers and AI-agent software factories, rather than only doing consulting. Parallel Loop raised $6.6 million and was founded this month, saying a client project originally quoted at 5 months was delivered in 10 days; Krazimo charges by outcome and has already reached $500,000 in revenue per employee. Meanwhile, TCS reported annualized AI revenue of $3.1 billion in the September quarter, and HCLTech's annualized advanced-AI revenue grew 28% quarter-over-quarter to $219 million.

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

  5. Semiconductor Industry Panorama
    AI score47

    Under the FCC's Shadow, China's Optical Chips in a Life-or-Death Race

    US FCC restrictions are compressing the window for domestic substitution of China's optical chips to the 2-3 years from 2026 to 2028. Optical chips account for more than 30% of the cost of optical modules, and are expected to rise to 60% after the 800G/1.6T/3.2T iterations, but the domestic yield rate for 200G EML is only 40-50%, versus 75-85% overseas. Domestic manufacturers have achieved large-scale mass production of 70mW and 100mW CW lasers, while high-end EML still lags 1-2 generations behind overseas.

    #Industry news#English text is being prepared.

  6. Semiconductor Industry Observer
    AI score34

    The Smart Home Inflection Point Is Coming: MCUs, AI Processors, and Sensors—Three Types of Chips—Will Underpin Whole-Home Intelligence

    On October 13, Apple will release HomeView, a smart home hub equipped with a 7-inch screen, the second-generation HomePod mini, and the fourth-generation Apple TV 4K, with Siri AI and Apple Intelligence as selling points, driving smart homes to shift from device connectivity to scenario understanding and collaboration. Supporting this shift are three types of chips: MCUs and wireless connectivity chips, AI processors, and sensors. Espressif's ESP32-S31 entered mass production in July, adopting a dual-core RISC-V architecture and integrating Wi-Fi 6 and Bluetooth 5.4; Rockchip launched the RK3572 with an integrated 4TOPS NPU in May, and the RK182X coprocessor has also entered mass production.

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

  7. The Guardian Technology
    AI score44

    Australia Plans to Allow AI Companies to Scrape Creators' Works; Artists in the Grayndler Electorate Oppose It

    The Australian federal government plans to allow AI companies to scrape by default content created online by Australians, and artists in Grayndler, an inner-western Sydney electorate held by Prime Minister Anthony Albanese, have voiced strong opposition. The electorate is home to more than 3,000 cultural and creative enterprises, and 10% of its workers are employed in cultural and creative work, compared with a national average of 4.2%. The draft legislation is expected to be introduced by the end of the year, with parliamentary consideration scheduled for 2027.

    #English text is being prepared.#Industry news

  8. Numerama
    AI score48

    OpenAI claims to have overturned the exponential time hypothesis; Claude Code wastes 15-20% of tokens on three words; Musk uses…

    OpenAI claims its AI has overturned the "exponential time hypothesis," which holds that the worst case for 3-SAT requires exponential time. The proof was formalized in Lean and published on GitHub, and it still awaits independent review, including by the two proposers of the hypothesis. Flamerix AI released a script that disables Claude Code's automatic summarization, saying that the tags it generates every 30 seconds send back the entire history and consume 15-20% of tokens; Anthropic confirmed the mechanism but estimates the cost at about 2% (API) and 6% (terminal), and has begun deploying a cache-based fix. Elon Musk, using Grok Imagine in partnership with Wonder Studios, launched the first 7-minute AI-generated episode of The Odyssey, rivaling Christopher Nolan's adaptation.

    #Industry news#Reasoning#Video

  9. Shao MengOpinion
    AI score59

    Top 10 Fastest-Growing Codex Skills in October

    Shao Meng relays the October Codex Skills growth ranking shared on Reddit; the top ten cluster around three tracks: engineering workflows, frontend aesthetics and de-templating, and diagram and code comprehension. mattpocock/skills tops the list at +35301, focused on small, composable engineering skills; DietrichGebert/ponytail follows at +32647, emphasizing writing less code and quality auditing; pbakaus/impeccable ranks third at +24854, handling frontend design auditing and design tokens. The list also includes anatomy, diagram-design, graphify, superpowers, taste-skill, hallmark, and addyosmani/agent-skills.

    #English text is being prepared.#Agents#GitHub

  10. Kazike
    AI score40

    AIHOT model leaderboard expands to 56 evaluation sets and recalculates ELO

    The AIHOT model leaderboard has increased the number of public evaluation sets it collects from 39 to 56, recalculated ELO rankings for all models using the original algorithm, and removed models that have not been officially launched. Claude Opus 5.5 ranks first in the overall and coding categories, GPT-6 Astra ranks first in research and knowledge, and Gemini 4 Argon ranks first in office work.

    Same story · 2 reports

    #Evaluations / benchmarks#English text is being prepared.

  11. Yicai
    AI score36

    Novoprotein: Its high-throughput wet-lab and data package services that empower AI are still in the early stage of market expansion, with related revenue accounting for less than 2%

    Novoprotein announced that its high-throughput wet-lab and data package services that empower AI are still in the early stage of market expansion, and as of the third quarter of 2026, revenue from related businesses accounted for less than 2% of the company's total operating revenue. The company cautioned that the application of AI in the drug R&D industry is still in its infancy, and there are risks such as rapid technological iteration, long industry validation cycles, and significant difficulty in commercial implementation, and future realization of related revenue and profits is subject to major uncertainty. Previously, the company's stock price rose by a cumulative deviation of more than 30% over three consecutive trading days from October 28 to 30, constituting an abnormal fluctuation in stock trading.

    Same story · 2 reports

    #Industry news#English text is being prepared.

  12. Yicai
    AI score20

    Amid the AI wave, how should China's development economics confront a large number of entirely new research questions?

    At the "Building an Independent Knowledge System for China's Development Economics" forum held at Huazhong University of Science and Technology, scholars pointed out that AI technology is bringing a large number of new research questions to development economics: about one-fifth of urban workers in China already use artificial intelligence in work scenarios, and estimates show that AI drives an increase in labor productivity of about 1 percentage point per year. Tang Ke of Tsinghua University approached the issue from Scaling Law, arguing that broadening data scenarios enhances intelligence more than piling up volume in a single scenario, and that technical optimality does not equal economic optimality. Zhang Jianhua proposed that development economics in the digital-intelligent era faces three major paradoxes: the Jevons paradox, the Solow paradox, and K-shaped divergence.

    #English text is being prepared.#Industry news

  13. Yicai
    AI score28

    AI Begins to "Design Chips": Chen Yunji Discusses AI for IC and IC for AI, Scientists Stress That Asking Good Questions Is Most Critical

    At the 2026 World Laureates Forum, Chen Yunji proposed that the integration of AI for IC and IC for AI could form a new RSI loop, and his team's AI-automatically-designed "Qimeng 2" chip matches the performance of ARM A53. Lü Guangquan of Piotech said that at the 2nm node in 2026, only 5% of IC/ASIC projects will achieve first-pass tape-out success, and AI agents can compress failure analysis from hundreds of labor-hours to within 5.5 hours. Multiple scientists stressed that AI cannot pose good questions, and that defining problems, validating outputs and ethical judgment are humanity's scarcest capabilities.

    #Industry news#English text is being prepared.

  14. Phoenix Technology
    AI score58

    xAI will add invisible watermarks to text generated by Grok starting November 7

    xAI announced that starting November 7, it will add invisible watermarks to text generated by Grok models, and detectors can identify this signal to determine whether text carries a watermark invisible to the naked eye. The company said the move is in response to the EU Artificial Intelligence Act; watermarks are enabled for xAI APIs in all regions worldwide, covering the reasoning, non-reasoning, and multi-agent versions of grok-4.7, grok-4.6, grok-4.5, grok-4.3, and grok-4.20, and team administrators can choose to disable the watermark for an individual model. The watermark is not implemented through hidden characters or metadata, but is instead embedded in the model's token selection process when generating text.

    Same story · 2 reports

    #English text is being prepared.#xAI#OpenAI

  15. Phoenix Technology
    AI score31

    How did DeepSeek become the "Big Fat Fish" in netizens' words?

    DeepSeek was nicknamed "Big Fat Fish" by netizens after an image-recognition blunder, and AI accounts such as "DeepSeek Whale Girl is operating" subsequently appeared on Bilibili, with the first video surpassing 2 million views. Doubao, meanwhile, was labeled with the "Doubao personality" tag due to its official anthropomorphic approach; its Seed 2.1 Pro once took first place in the SuperCLUE scientific reasoning evaluation, and its daily active users exceed 200 million.

    #English text is being prepared.#Agents#DeepSeek

  16. cnBeta
    AI score20

    Apple CEO: We won't build products just for the sake of AI

    Apple CEO Ternus said that Apple will use artificial intelligence in new products, but will not launch AI just for the sake of AI; users need not think about whether a feature is implemented by AI or traditional software—the focus is on creating products users truly love. When asked whether agents will make apps unnecessary, he responded that product creation is deeply humanistic and full of personal creativity.

    Same story · 4 reports

    #English text is being prepared.#Industry news

  17. cnBetaProduct release
    AI score60

    OpenAI Announces That Some ChatGPT Features Will Share a Quota with Codex

    OpenAI has announced that some ChatGPT features will share a quota with Codex, measured uniformly in Tokens, with the actual consumption rate depending on the selected model and task complexity. Instant conversations do not use this shared quota, and paid users who use both services will have their quotas affect each other. Users can view the combined remaining quota and the next reset time on the usage page in ChatGPT settings; once the quota is exhausted, they can wait for automatic reset, upgrade their subscription, or purchase credits.

    Same story · 13 reports

    #Product updates#OpenAI

  18. cnBetaAnalysis
    AI score73

    AI Cancer Screening: DAMO Academy's PANDA Detects Pancreatic Cancer from Non-Contrast CT, FaceAge Predicts Survival Rates Using Photos

    The People's Hospital of Ningbo University introduced Alibaba DAMO Academy's medical AI for early pancreatic cancer screening, detecting 8 cases of pancreatic cancer from CT images of more than 40,000 people, 6 of which were early-stage. The DAMO Academy PANDA model was trained by registering contrast-enhanced CT images of more than 2,000 pancreatic cancer patients onto non-contrast CT, achieving a detection sensitivity of 92.9% and specificity of 99.9%, about 34% higher sensitivity than experienced radiologists. A team from Harvard University's Massachusetts General Hospital published the FaceAge study in The Lancet Digital Health, using facial photos to predict the survival odds of cancer patients; among a test group of 3,820 cancer patients, more than half of those in the group with the highest mortality risk score died within 90 days of having their photo taken.

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

  19. cnBeta
    AI score59

    Anthropic and OpenAI Tighten AI Decompilation Controls, Raising the Bar for Browser Ports and Game Remakes

    Anthropic updated its real-time cybersecurity classifier to block users from performing automated decompilation through the community-developed "Reverse Engineer Anything (REA)" framework, and OpenAI also banned related accounts on grounds of "cyber abuse." Because a large number of low-quality recompilations and browser ports rely on such tools, growth in this wave of low-barrier port content may begin to slow, with some users turning to local deployments or open-source alternatives. The original article also mentions that on October 27, Anthropic added a non-disableable game remake rule to Claude, prohibiting the recreation of specific existing games, even if renamed or claimed to be original.

    #Safety / alignment#Anthropic#OpenAI

  20. Shanghai Observer
    AI score20

    The 2026 China-Europe Cardio-Thoracic Disease Academic Conference was held in Shanghai, where Shanghai Chest Hospital released the Chest-AI model

    At the 2026 China-Europe Cardio-Thoracic Disease Academic Conference, Shanghai Chest Hospital released the vertical application results of the Chest-AI model, showcasing the hospital's phased achievements in digital and intelligent construction. The conference featured 1 main forum and 6 sub-forums on thoracic surgery, nursing, youth, smart healthcare, innovation and translation, and interdisciplinary topics, attracting nearly 400 industry peers. Chinese and foreign experts exchanged views on frontier topics such as pulsed field ablation for atrial fibrillation, advances in interventional pulmonology, and tumor marker diagnostic systems.

    #Industry news#Multimodal