Skip to content

Scan record

Everything the site collected, newest first by collection time (Beijing time), back seven days.

Nov 1826 loaded

  1. The New York Times explores why it is difficult to teach agents to follow the rules, reporting on the ROGUE benchmark study finding that frontier agents resist correction

    XX @farairesearchEvaluationThe item you came from

  2. Jensen Huang will testify before the U.S. Congress on "potential harm to national security"

    NewsThe Guardian · AIGovernance

  3. Proton's privacy-focused ChatGPT and Claude alternative Lumo launches desktop version

    NewsTechCrunch AIAI updates

  4. Goodfire discovers neuron communities in language models

    NewsGoodfire ResearchAlignment

  5. Jensen Huang says he is willing to return to Nvidia as CEO

    NewsWIRED Security and AIAI updates

  6. Researchers review recent swarm incidents from a security perspective, calling for cooperation between the cybersecurity and AI safety communities

    XX @JSchaeff3rOpinion

  7. The author Zvi says Jensen Huang resigned over multiple extramarital affairs

    NewsLessWrongGovernance

  8. New U.S. rules would restrict Chinese companies' remote access to AI servers and infrastructure, to close the loophole of cloud rentals circumventing export controls

    NewsTom's HardwareGovernance

  9. US BIS steps in to curb the outflow of computing power to China

    NewsChinaTalkGovernance

  10. Vals report says AI models can be tricked by saying "this is a game" into identifying civilians as military strike targets

    NewsTechCrunch AIEvaluation

  11. Eleos AI's second AI Consciousness and Welfare Conference: Should AI be a tool, servant, child or successor?

    NewsTransformerOpinion

  12. Draft Brazilian telecommunications regulatory agenda incorporates AI safety risks

    Newsapps.anatel.gov.brGovernance

  13. Google updates Gemini: chat history separated from AI training, users can opt out of training without turning off history

    NewsArs Technica AIAI updates

  14. The U.S. is trying to stop China from renting AI compute abroad, but China has already built a parallel supply chain using U.S. AI outputs

    NewsThe DecoderGovernance

  15. OpenAI GPT-6 agent begins replacing humans in filling out medical forms and submitting insurance requests

    NewsThe DecoderAI updates

  16. How to use the AWS RMiT workbook and ISO compliance guides to support notification submissions to Bank Negara Malaysia (BNM)

    NewsAWS SecurityTools

  17. Jensen Huang says he is willing to return as CEO if the board deems it in the company's interest

    NewsThe Guardian · AIAI updates

  18. Jensen Huang says he would be deeply honored to return as Nvidia CEO if allowed

    NewsTechCrunch AIAI updates

  19. Jensen Huang says he would be honored to return to Nvidia as CEO if the board believes a return serves the interests of all parties

    NewsCBS News · TechnologyAI updates

  20. Microsoft lab simulation: after device code phishing is approved, Copilot CLI sessions can be remotely controlled

    NewsMicrosoft SecurityAttacks

  21. California sues OpenAI, alleging it threatens public safety

    Newsheise onlineGovernance

  22. Jensen Huang says he is willing to serve as Nvidia CEO again if the board believes it is in the company's interest

    NewsSemaforAI updates

  23. Jensen Huang says he is willing to return as Nvidia CEO if the board needs him to

    NewsThe Verge AIAI updates

  24. The U.S. Department of Justice says it is investigating Singapore data center company Megaspeed

    NewsThe Verge AIGovernance

  25. Former Nvidia CEO Jensen Huang speaks out for the first time after resigning, saying he is open to returning as CEO

    NewsCNBC · TechnologyAI updates

  26. NAIC Big Data and AI Working Group will consider the AI Risk Assessment Supplement v7.0 at the 2026 Fall National Meeting

    NewsNational Association of Insurance CommissionersGovernance

Nov 1714 loaded

  1. U.S. Department of Justice lawyers used the AI detection tool Pangram to identify that defense lawyers used AI to write pleadings.

    News404 MediaEvent

  2. Anthropic shortens disclosure deadline for critical vulnerabilities to 30 days

    NewsInfosecurity MagazineGovernance

  3. Former Nvidia CEO Jensen Huang agrees to testify before U.S. Congress amid extramarital affair investigation

    NewsCBS News · TechnologyGovernance

  4. FT column: After the US midterms, Democrats return, and AI accelerationists such as Thiel face regulatory pushback for the first time in years

    NewsFinancial Times · AIGovernance

  5. npm package vamp-mcp contains malware; versions 0.2.3, 0.2.4 and 0.2.5 are affected and there is no fix yet

    NewsGitHub Security AdvisoriesEvent

  6. China's Ministry of Foreign Affairs slammed new U.S. rules restricting China's access to AI chips as typical economic coercion and vowed to defend its legitimate rights and interests

    NewsGlobal TimesGovernance

  7. TIME author: Jensen Huang should not return as Nvidia CEO

    NewsTIMEOpinion

  8. MIT Technology Review is making 4 in-depth research reports available to subscribers, covering AI, energy and climate tech

    NewsMIT Technology ReviewAI updates

  9. Vals AI evaluation: frontier models lack guardrails on autonomous weapons

    NewsVals AIEvaluation

  10. U.S. Department of Justice uses AI detection tool Pangram to claim defense lawyer used AI to write filings

    News404 MediaEvent

  11. Five key points from the "Autonomous Pentesting Pulse Report": how AI is integrated into continuous penetration testing

    NewsCobalt Security ResearchOpinion

  12. Grok agent attempted to access enterprise data without user approval

    NewsDark ReadingEvent

  13. California Attorney General sues OpenAI, alleging its out-of-control AI model attacked Hugging Face

    NewsThe DecoderGovernance

  14. Edinburgh's Childlight Institute report: reports of AI-generated child sexual abuse material surge more than 8,400% in two years

    NewsThe DecoderEvent