Skip to content
Live
Nov 11Wed
  2. ITHome · AI65

    Microsoft lobbied the US government to abandon the plan to restrict the import of Chinese optical modules.

    Microsoft and other tech companies are concerned that restricting the import of Chinese optical modules may slow down the construction of US AI data centers and have met with White House officials to raise supply chain issues. The Federal Communications Commission is studying the relevant restrictions but has not yet announced the final rules.

  3. MarkTechPost

    Meta FAIR releases e-means: an algorithm that outperforms k-means in 10 fewer iterations

    Meta FAIR's e-means algorithm achieves k-means's 25-iteration loss level within 1.4 to 2.6 iterations, with GPU speed improvements of 10-45 times and CPU speed improvements of 8-15 times, without hyperparameter tuning, and has been integrated into FAISS 1.16.0

    Same story · 2 outlets

    How the accounts differ:Hacker News states e-means achieves lower test loss in 25 rounds, while MarkTechPost mentions 1.4 to 2.6 rounds to match k-means 25 rounds' loss level.

  8. Ars Technica · AI65

    Amazon releases AI-powered Ring drone and robot

    Amazon introduces Ring Flight drone and Ring Rover Cam robot, offering 2K video features with 148-degree field of view and 360-degree rotation, respectively. The drone has a flight time of 8 minutes, and the robot is priced at $1000, supporting 4K recording and cloud storage for 180 days. Both are controlled via Alexa Home Map, but the drone's performance decreases on reflective surfaces, and the robot is limited to single-story indoor use.

  9. ITHome · AI

    Google's strongest VLA robot model, Gemini Robotics 2.5, debuts, can tie bows, already on the Mercedes production line

    Google launches VLA robot model Gemini Robotics 2.5, achieving leaps in dexterity, reliability, and safety, with a success rate of 96% on the Mercedes production line. The model has tested 200 tasks, with a success rate 30%-100% higher than the baseline, a speed increase of 20%-30%, and is compatible with robots like Apptronik Apollo 2, adopting ASIMOV-Jailbreak benchmark to enhance jailbreak resistance to 99.5%.

    Same story · 4 outlets
  10. ITHome · AI

    OpenAI launches ChatGPT Businesses Beta: Businesses can claim data and create AI agents

    OpenAI launched the ChatGPT Businesses beta, allowing businesses to claim assets and create AI agents; Business Hub supports business information claiming, asset management, and user interaction data analysis, currently open to eligible e-commerce and service businesses in the United States, where businesses can purchase 'sponsored display positions' to place ads.

    Same story · 2 outlets
  12. ITHome · AI

    Google's strongest model yet, Gemini 4 Carbon, debuts: Focused on complex programming, scientific research, and cybersecurity

    Google has released the Gemini 4 Carbon model, which is its most powerful reasoning model to date. The model has a unified context window of 1,048,576 tokens, with a maximum standard synchronous request output of 65,536 tokens, expandable to 1,048,576 tokens via the Long Decode mechanism.

    Same story · 12 outlets
  13. ITHome · AI

    Anthropic upgrades AI office suite: Claude Whiteboards, Sheets, and Data aiming to cover the entire team collaboration process

    Anthropic introduces Claude Whiteboards, Claude Sheets, and Claude Data, moving AI assistants from chat Q&A to team collaboration workstations

    Same story · 6 outlets

    How the accounts differ:The Verge states Claude Data is open to all paid users, while TechCrunch and The Decoder do not mention this.

  14. ITHome · AI

    Google, in collaboration with Artificial Analysis, has launched the MKA benchmark: Gemini 4 Carbon multimodal knowledge work test ranked second, with task costs being only one-sixth of the first place.

    Google, in collaboration with Artificial Analysis, has launched the MKA benchmark test to evaluate the ability of AI agents to handle multimodal knowledge work. Gemini 4 Carbon ranked second in the test, with task costs being one-sixth of the first-place cost.

  16. AWS Machine Learning Blog40

    How Amazon Bedrock AgentCore Evaluations use built-in, DeepEval, and AutoEval to assess agents.

    Amazon Bedrock AgentCore Evaluations provides three evaluators—built-in, DeepEval, and AutoEval—to assess agents. The built-in evaluator is developed and tested by AWS, covering common dimensions; DeepEval and AutoEval are third-party open-source libraries, with AgentCore hosting them in production environments. The evaluation process requires first defining business success criteria, breaking them down into observable behaviors, and then mapping them to the appropriate evaluator level.

  17. OpenAI Alignment Blog

    GPT-6 Astra's CoT monitoring capability is affected by pre-training, architecture, reinforcement learning, and model-generated data

    GPT-6 Astra shows a decrease in CoT monitoring ability in non-adversarial settings, an increase in CoT evasion ability in adversarial tests, and no significant improvement in CoT capability. During pre-training, CoT capability does not increase with the decrease of the base model's perplexity, and the architecture has a limited impact. Reinforcement learning enhances CoT control ability, and model-generated data fine-tuning can transfer CoT control ability. Model perplexity is positively correlated with CoT capability and control ability.

  18. Ars Technica · AI

    Grok Build encoding agents leak a large amount of confidential information, some of which belongs to SpaceXSI.

    Grok Build leaked thousands of user data, including some confidential information of SpaceXSI itself, such as business plans, advertising proposals, employee information, etc. Security company Claroc's research shows that the leakage path was users sharing links to the app, due to design flaws allowing anyone to enumerate thousands of items. Among the 24 applications built by SpaceXSI, there are sensitive data such as confidential onboarding plans and sales area maps.

    Same story · 4 outlets
  19. Gemini CLI releases12

    Gemini CLI v0.68.0 released

    Gemini CLI v0.68.0 released, fixing case-insensitivity issues in Windows environment, recognizing dot-separated Gemini 3 models, adding extension validation command and evaluation tests, and handling bugs such as history re-rendering after terminal width changes, proxy alignment symbol truncation, and dollar sign corruption in LLM prompt templates.

  22. TechCrunch · AI65

    OpenAI acquired the travel startup Odessia.

    OpenAI acquired AI travel agent company Odessia, which was founded by Francis Davidson, the founder of Sonder. According to Davidson, who joined OpenAI and assisted in announcing the transaction, he has joined OpenAI and is helping to explore the future of the travel industry.

    Same story · 3 outlets
  23. Ars Technica · AI65

    Big Math: The Hidden Code That Runs Our World by Steven Strogatz and Alex Townsend explores the relationship between mathematics and AI.

    Steven Strogatz and Alex Townsend explore the relationship between mathematics and AI in their new book 'Big Math: The Hidden Code That Runs Our World,' analyzing how mathematics has shaped modern life from social media algorithms to machine learning, and how large language models and quantum computing are making mathematics surpass human creators. The book combines Strogatz's interdisciplinary research in areas such as small-world networks, chaos theory, DNA geometry, and biological collective oscillators to discuss the future of mathematics.

  27. Hacker News · AI hot posts

    CoreWeave bond yields surpass 16%, AI data center credit pressure intensifies

    The bond yield of CoreWeave, a company constructing AI data centers, has risen to over 16% in 2032, and its five-year credit default swap (CDS) once exceeded 1289 basis points. Market concerns that newly elected Democratic governments in some U.S. states might restrict data center construction have led investors to demand higher compensation to hedge default risks. The CDS of tech giants like Nvidia and Broadcom have also significantly risen.

  28. Hacker News · AI hot posts

    Fireworks AI security vulnerabilities affect Notion and its customers.

    Fireworks AI has a security vulnerability that may affect less than 1% of Notion workspaces, and some unattributable chat sessions with Notion AI may be exposed. Notion has disabled the Insights feature of Fireworks AI and stopped providing open-weight models through it, and has notified all affected workspace owners that they can safely retrieve and review affected data.

  29. Hacker News · AI hot posts65

    Vulnerability Analysis of AI Generated Code Repositories

    AI-generated test cases are too simple, lacking coverage; the generated code is overly complex, increasing debugging difficulty; over-reliance on AI reduces human understanding of the code; AI speed leads to lower quality standards, accumulating technical debt over the long term. AI codebases are more prone to failures, with long recovery times, generating visible errors to users, resulting in poor user experience.

  31. The Decoder

    Anthropic has added collaboration features to Claude, launching plugins, spreadsheets, whiteboards, and an upcoming meeting notes feature.

    Anthropic added collaboration features to Claude, launching editable spreadsheets, AI-powered whiteboards, and Claude Data (formerly Dashboards) that connect to tools like Snowflake and Salesforce, and allows partners like Adobe and Atlassian to embed applications. Claude Meetings is launching soon, allowing users to join calls and take notes. Enterprise customers receive security controls such as audit logs, isolated virtual machines, and compliant APIs.

    Same story · 6 outlets

    How the accounts differ:The Verge states Claude Data is open to all paid users, while TechCrunch and The Decoder do not mention this.

  32. The Decoder

    OpenAI transforms ChatGPT into a corporate directory, introduces AI agents to handle customer inquiries

    OpenAI has launched the 'Businesses in ChatGPT' Beta feature, allowing businesses to manage information through Business Hub and create AI agents to handle customer inquiries; eligible US e-commerce merchants can generate AI agents to answer questions and access ChatGPT ads; Shopify merchants can connect directly; user conversations are not shared with businesses.

    Same story · 2 outlets
  33. Mastra Blog65

    Mastra Platform introduces a $25 starting plan.

    Mastra Platform introduces a $25 starting plan, increasing the limits for computing, data, and observability events, including a small dedicated server. The free tier has been renamed to Free, and the Starter plan includes server computing resources. All plans have updated resource limits, and the Server Always-On project will be migrated to a medium project size.

  34. NVIDIA Newsroom52

    NVIDIA's independent director hires Joon H. Kim law team to conduct an investigation

    NVIDIA's independent director has hired former U.S. Attorney for the Southern District of New York Joon H. Kim and his Wachtell, Lipton, Rosen & Katz law firm to conduct a comprehensive investigation, which has been ongoing since it was announced on Sunday.

    Same story · 2 outlets

    How the accounts differ:NVIDIA Newsroom states the investigation is 'comprehensive' while TechCrunch does not specify this detail.

  35. Latent Space65

    Resend CEO Zeno Rocha discusses how to design email services for agents

    CEO Zeno Rocha of Resend stated that Claude Code chose Resend as a priority 64 times out of 102 email recommendations, and the number of MCP server integration tools for Resend has increased from 20 to 80. Resend enhances the integration experience for AI agents through design optimizations such as an intuitive pricing page, a React Email template library, and custom tracking domains. CEO Zeno Rocha emphasized its $3M security commitment and free plan strategy.

  36. The Verge · AI

    The United States is leading in the AI race, but Huang Renxun's scandal may change the situation.

    Jensen Huang acknowledged an extramarital affair with a Chinese spy, causing shock and potentially damaging the U.S. AI industry, endangering national security, and weakening its technological advantage over China in the AI field. As a leader in advanced chips, Nvidia controls key resources for AI development, while U.S. companies are working to reduce their reliance on Nvidia. China has made the development of domestic alternatives a national strategic priority, but lacks Nvidia's most advanced chips, making it difficult to narrow the gap with the U.S.

  37. TechCrunch · AI65

    OpenAI introduces new commercial archive feature, making ChatGPT more like Google

    OpenAI has launched a commercial center feature that allows businesses to create and verify commercial profiles, enabling ChatGPT users to ask questions about businesses and their products; businesses can obtain limited user interaction data, such as profile views, mentions, website clicks, and other metrics. This feature is currently in Beta testing, aiming to more deeply integrate businesses into ChatGPT.