NAll news

All news

Sunday264 items

  2. t3n
    AI score52

    Over a Thousand Mathematicians Form the AHM Association to Boycott OpenAI and AI Research

    More than 1,000 mathematicians have formed the Association for Human Mathematics (AHM), pledging not to use AI in their research and calling on peers to stop collaborating with OpenAI. The association criticized OpenAI for releasing 722 papers proving or disproving mathematical problems all at once earlier this month, arguing that this was a display of power rather than scientific rigor, and noted that mathematical progress depends on human expertise formed through writing and thinking. Those who join must pledge not to publish AI-generated mathematical texts and not to use AI models for research; members include scholars from institutions such as MIT, Caltech, and the Technical University of Munich.

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

  3. Lingxi Li
    AI score24

    Lingxi Li marvels at the Stageshow skill

    Benji Taylor introduces Stageshow: a skill that helps AI agents generate better faux UI stills, animations, and bento grids, installable via `npx skills add benjitaylor/skills --skill stageshow`. Lingxi Li reposted it and exclaimed "wowwwww".

    Quoted postBenji Taylor · @benjitaylor

    Stageshow: a skill that helps your agent make better faux UI stills, animations, bento grids, and more. npx skills add benjitaylor/skills --skill stageshow https://t.co/Qve3C2cECg

    #English text is being prepared.

  5. QbitAIResearch
    AI score67

    DeepSeek Unveils Super-agent Architecture: Two-Tier Router and Worker Design

    DeepSeek has published an engineering-oriented architecture article titled "Super-agent: Toward More Autonomous Agent Systems," proposing that an upper-layer Router is responsible for receiving messages, assigning tasks, and maintaining global context, while lower-layer Workers execute specific tasks in parallel. In a simulated environment with approximately 300 messages and 20 work requests, repeated tests 5 times showed that the Super-agent's median task completion time was 12.4 minutes and its 95th percentile was 15.7 minutes, whereas the traditional single-Session approach had a 95th percentile of 29.5 minutes under neutral prompts. The article also introduces two-level event filtering, an asynchronous ReAct Loop (already implemented in DeepSeek Harness but not enabled by default), separation of computation and state from the execution environment, and a Watchdog supervision mechanism.

    Same story · 23 reports

    #English text is being prepared.#Agents#DeepSeek

  7. Yicai
    AI score55

    Five tech giants' earnings released: Amazon, Alphabet, and Meta post negative free cash flow for the quarter

    Five of the Magnificent Seven have released their latest quarterly earnings. Amazon, Alphabet, and Meta posted quarterly free cash flow net outflows of approximately $33.5 billion, $9.1 billion, and $2.1 billion, respectively, while Microsoft and Tesla posted positive values of approximately $18.6 billion and $1.5 billion. Revenue for the three major cloud businesses grew 39%, 48%, and 103% year over year, respectively, but capital expenditure rose in tandem, and Alphabet's stock fell 12.6% in a single day after its earnings report. The report cited analysts as saying that AI investment is shifting from competing on the scale of capital expenditure to testing commercial returns, with order visibility and capital discipline becoming key indicators.

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

  8. Browser UseFirst-partyProduct release
    AI score58

    Browser Use launches BU Ultrafast browser agent: median task time 22 seconds, cost $0.0064

    Browser Use has opened its fastest browser agent, BU Ultrafast, from the waitlist to all Cloud plans in the US, including the free tier. Official data shows a median task time of 22 seconds and a cost of $0.0064 per task, compared with 79 seconds and $0.0098 for the default agent on GPT-6 Luna. To use it, simply set model to bu-ultrafast; if you need more step-by-step reasoning, you can switch to bu-fast from the same agent, which is billed at standard rates.

    BU Ultrafast: a white craft racing over red canyons, median task 22s, $0.006

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

  10. Sina Tech
    AI score71

    AI Spending Erodes Free Cash Flow, US Tech Giants' Earnings Trigger Selloff

    After tech giants reported earnings over the past two weeks, investors sold off shares as rising data center construction costs eroded free cash flow, and the Nasdaq 100's October decline widened to 6.7%. Alphabet's stock plunged 13% on Thursday, its largest single-day drop in history, as it posted negative free cash flow for a second consecutive quarter; Amazon and Meta were also negative this quarter, while Microsoft was the only one of the four hyperscalers with positive free cash flow, at $18.6 billion. Alphabet's future spending obligations jumped to well over $1 trillion, and Meta's future spending commitments rose to $825 billion.

    #Industry news#OpenAI#Anthropic

  11. cnBeta
    AI score58

    Anthropic Tightens Claude Usage Boundaries, Prohibits Replicating Existing Games

    Anthropic has updated Claude's usage rules, explicitly disallowing the construction of a replica of a specific existing game, even if it is renamed or called original. Some users report that Claude has therefore refused to replicate work the developers themselves completed earlier, and the scope of impact extends beyond games, including refusing to invoke the "Reverse Engineer Anything" capability. The clause is believed to be intended to avoid copyright and legal risks, but it has already sparked a large number of new issues on Claude's GitHub page.

    Same story · 14 reports

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

  12. Zhidongxi
    AI score66

    Anthropic Adjusts Subscription Plans: Max Renamed, Ultra's 5-Hour Limit Removed

    Anthropic has announced adjustments to its subscription plans: Max 5x is renamed Max, and Max 20x is renamed Ultra, with prices remaining at $100 and $200 per month. It has been verified that Ultra accounts have the 5-hour usage duration limit removed, while the Pro and Max plans still retain this limit. Tibo, the head of OpenAI Codex and ChatGPT, responded in the comments that it would be best to announce a 28-day reset period.

    Same story · 12 reports

    #Industry news#Anthropic#OpenAI

  13. Zhidx
    AI score35

    Kuaishou proposes the "Skunk Works" model: How AI moves from individual productivity gains to organizational transformation

    At a tech salon, Kuaishou, together with its enterprise-grade AI productivity platform iWork, shared the four stages of AI moving from assisted programming to team collaboration and organization-level efficiency gains, and proposed the "Skunk Works" organizational model. After organizational and engineering adjustments, the risk control team of Kuaishou's Magnetic Engine increased requirement throughput by 133%, shortened the delivery cycle by 79%, and increased the number of full-stack delivery requirements by 800%. OPPO, in turn, introduced its AI Native evolution path and its agent collaborative delivery platform CoMind.

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

  14. Zhidongxi
    AI score41

    Baidu AI Cloud launches a dual-layer AI Infra and Agent Infra foundation to support the implementation of industrial agents

    On October 29 at the Wuhan Super Agent Conference, Baidu AI Cloud launched a dual-layer infrastructure consisting of AI Infra and Agent Infra, covering capabilities such as sandboxing, computing, networking, storage, observability, and security, enabling agents to run stably in enterprises over the long term. Baidu Dazi has reached tens of millions of monthly active users since its launch in March this year, growing more than 10-fold month over month; after Songxianxian adopted it, data aggregation was shortened from two days to 3 minutes. IDC predicts that by 2030, there will be 2.216 billion active agents worldwide, with a compound annual growth rate of 139%.

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

  15. ZhidxResearch
    AI score70

    DeepSeek Harness unveils Super-agent architecture, using five major modules to take over session scheduling

    The DeepSeek Harness team has unveiled the underlying architecture of Super-agent, an autonomous Agent exploration project, bringing classic computer systems ideas such as scheduling, caching, compute-storage separation, and Watchdog into Agent design. The architecture consists of five modules: a dual-layer architecture, event-driven design, asynchronous multi-channel interaction, compute-storage separation, and an external Watchdog. Its goal is to enable the Agent to stay continuously online, advance multiple tasks in parallel, and proactively respond to environmental events. In preliminary experiments, Super-agent's p50 and p95 task completion times were 12.4 minutes and 15.7 minutes, respectively. After removing subscription passthrough, the number of Router decisions increased by about 30%, and read-cache tokens rose from 23.23 million to 31.77 million. The team said the experiments were based on synthetic workloads and a simplified environment, and that this plugin is only a prototype for validating the technical approach, not an official DeepSeek product.

    Same story · 23 reports

    #English text is being prepared.#Agents#DeepSeek

  16. Hacker News PopularResearch
    AI score71

    Jepsen tests ParticleDB 1.1.4: numerous consistency and data loss issues found on a single node

    Jepsen published a consistency test report on Orkes' ParticleDB 1.1.4 through 1.1.18.11.rc-22, covering all transaction modes and three WAL synchronization modes, with testing limited to a single node. The report documents ignored NOT NULL and UNIQUE constraints, incorrect WHERE predicate matching, loss of committed writes, anomalies such as G-Single and Stale Read, and data loss under WAL bit-flip and file block misalignment faults. Orkes says most issues have been fixed in the 1.1.19-rc series, but Jepsen has not yet tested these versions.

    #Evaluations / benchmarks#English text is being prepared.#English text is being prepared.

  17. Hacker News TrendingResearch
    AI score64

    Cisco Talos analyzes VOLTA infostealer using LLM to rank victim assets

    Cisco Talos disclosed a self-deleting Windows infostealer named VOLTA-AUTODOXER that, after stealing credentials, sessions, and crypto wallets, packages the victim's structured information and sends it to an LLM, asking it to rank assets by monetization value and assess their security habits. The trojan by default calls anthropic/claude-opus-4.6 on OpenRouter with temperature set to 0.2, and when a model refuses, it falls back in sequence to moonshotai/kimi-k2 and google/gemini-2.5-flash. Talos also provided mitigation recommendations, including restricting processes that launch Chrome with --headless and preserving the APPDATA%\volta_debug.txt log after compromise.

    Same story · 4 reports

    #Safety / alignment#Agents#Anthropic

  18. heise online
    AI score34

    Vishay receives €215 million subsidy to build a 300mm wafer fab in Itzehoe

    The American semiconductor manufacturer Vishay has received a subsidy of approximately €215 million from the German federal and state governments and will build a new factory in Itzehoe, Schleswig-Holstein, to produce 300mm wafers for manufacturing Power-Mosfets for applications including automotive, industrial, and AI high-performance computing. Compared with the existing 200mm production line, 300mm wafers can provide 125% more theoretically usable chip area per run, and the company plans to add up to 250 jobs. Vishay has not disclosed the total investment amount; CDU state parliament member Andreas Hein said it is approximately €1 billion.

    #Industry news#English text is being prepared.

  19. Unite.AIResearch
    AI score62

    Cisco Talos analyzes VOLTA infostealer using LLM to rank victims' assets

    Cisco Talos researcher Ryan Fetterman published an analysis of VOLTA-AUTODOXER on October 31; this Windows infostealer, after stealing credentials, calls a large language model to rank the victim's assets into low, medium, and high tiers by monetization value. The malware packages the collected identity information and POSTs it to OpenRouter, using anthropic/claude-opus-4.6 by default, with a fallback chain of moonshotai/kimi-k2 and google/gemini-2.5-flash, a temperature of 0.2, and checks for refusal phrases such as "I can't" before switching models level by level. Talos believes the AI layer brings a shift in workload rather than gains in speed or scale, and recommends restricting Chrome from launching with the headless flag, retaining volta_debug.txt logs, and revoking potentially leaked tokens.

    Same story · 4 reports

    #Safety / alignment#Agents#OpenAI

  20. SiliconANGLE
    AI score41

    Can enterprise AI pay for this construction boom?

    Episode 332 of SiliconANGLE's Breaking Analysis discusses whether enterprise AI investment can achieve economic self-sufficiency; guest Crawford Del Prete believes early productivity gains have emerged but are not yet reflected in the data. The program cites Anthropic's S-1: revenue grew 14-fold to $16 billion in the first half of 2026, it holds $132 billion in liquid assets, has raised approximately $140 billion in cumulative funding and burned $10 billion, and carries $537 billion in multi-year compute commitments. The research team previously identified 2029 as the year with the highest bubble risk.

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