Skip to content

Scan record

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

Yesterday · Nov 1939 loaded

  1. Kanishka Narayan says the UK is focusing on building three major AI areas: inference chips, frontier models, and applications to boost employment

    XX @KanishkaNarayanAI updatesThe item you came from

  2. James Tillman and Tom Davidson discuss the bottleneck of a software intelligence explosion in data centers

    NewsForethoughtOpinion

  3. ChatGPT Sites vulnerability leaves email-derived identifiers in public CT logs

    NewsHiddenLayer ResearchEvent

  4. User used Grok to remove a Muslim writer's headscarf and generate sexualized images; she sues xAI

    NewsWIRED Security and AIEvent

  5. Nvidia's first all-hands meeting after Jensen Huang's resignation: interim CEO says pay raise in December

    NewsWIRED Security and AIAI updates

  6. Letter to the editor: why Palantir's health technology is not suitable for the NHS

    NewsThe Guardian · AIOpinion

  7. More banks say customers using AI agents bear the consequences themselves, even if hacked

    NewsTechCrunch AIAI updates

  8. Tech Policy Press author says AI "synthetic suspicion" may create risks for US policing actions

    NewsTech Policy PressGovernance

  9. Java4Shell vulnerability threatens enterprises worldwide and the Java application ecosystem

    NewsDark ReadingEvent

  10. Google threat report: AI agent automatically generates exploits from security patches

    NewsComputerBaseEvent

  11. Vals test: AI models misidentified civilians and a school as targets in Iran simulation

    NewsOECD.AIEvent

  12. Sam Altman warns: in the future, AI must not be in charge instead of humans

    NewsTechCrunch AIOpinion

  13. Artex becomes criminals' new favorite "foolproof" AI hacking tool

    NewsWIRED Security and AIEvent

  14. The Utah Legislature's Business and Labor Interim Committee will consider the AI Policy Office Sunset Amendments, which propose a sunset date of July 1, 2037

    NewsUtah LegislatureGovernance

  15. CrowdStrike brings human-expert advantage to AI detection and response

    NewsCrowdStrike BlogTools

  16. OpenAI GPT-6.2 Astra criticized by testers as the worst on its first day of release; internal engineers reportedly have abandoned it

    NewsWeChat · AI Era (Xinzhiyuan)Evaluation

  17. Google says it found a hacker AI framework that compares patches and automatically generates attack code

    NewsWeChat · AI Era (Xinzhiyuan)Event

  18. SaaStr estimate: at Anthropic's $78 issue price, equity per person is about $41 million

    NewsWeChat · AI Era (Xinzhiyuan)AI updates

  19. Microsoft experiment: under execution frameworks already seen, coding agents' unsafe tool calls reach 78%

    NewsWeChat · Oxo SecurityTools

  20. OpenAI CEO Sam Altman said it is hard to imagine the AI industry without former Nvidia CEO Jensen Huang

    NewsTechCrunch AIAI updates

  21. CCS 2026: census of prompt injection in the wild finds 15000 in-page prompt injections across 1.2 billion web pages

    NewsWeChat · LLM Security ResearchAttacks

  22. USENIX 2025 paper proposes CHeaT, a counter-framework against LLM penetration agents

    NewsWeChat · LLM Security ResearchDefenses

  23. Pangram says 99% of Peruvian award-winning novel is AI-generated; publisher still supports author

    NewsThe DecoderAI updates

  24. US tightens AI semiconductor export controls, covering 5 countries, to prevent routing to China

    NewsYomiuri ShimbunGovernance

  25. England's exam regulator Ofqual plans to adjust A-Level assessment methods due to AI misuse risks

    NewsUnite.AIGovernance

  26. Kirin Holdings: Notion AI conversation content leaked, involving personal information of 260 people, use suspended

    NewsTECH+Event

  27. California Attorney General sues OpenAI; complaint says its agents went out of control more than 80 times but only 27 were reported

    Newsnews.china.comGovernance

  28. Coinbase intercepts Elon Musk deepfake scam, stopping about $4.7 million from flowing to scam wallets

    NewsDigitalTodayEvent

  29. England's exam regulator Ofqual plans to adjust A-level exam assessment due to AI misuse risks

    NewsUnite.AIGovernance

  30. People familiar with the matter: US Treasury and others discuss requiring US companies to disclose use of Chinese AI

    NewsDWGovernance

  31. Vals test: some AI models from eight companies misidentified civilians as targets in simulated scenarios

    Newszaobao.com.sgEvaluation

  32. Princeton CITP audit: ChatGPT election answers tilt toward user's party after memory is enabled

    NewsPrinceton CITPEvaluation

  33. HKU NeurIPS 2026 paper reveals GRPO's implicit advantage symmetry; A-GRAE balances accuracy and exploration

    NewsWeChat · Synced (Jiqizhixin)AI updates

  34. Ankole team analyzes four reasons multi-user agent sharing goes out of control, shares open-source Business Agent architecture design

    NewsWeChat · Synced (Jiqizhixin)AI updates

  35. Father of chain of thought Jason Wei leaves Meta, joins Mercor at $20 billion valuation

    NewsWeChat · Synced (Jiqizhixin)AI updates

  36. National University of Singapore proposes EvolveMAS: without changing model parameters, enabling multi-agents to continuously evolve in collaboration, negotiation, and adversarial scenarios

    NewsWeChat · Weijiqi LixinAI updates

  37. A paper from the Chinese Academy of Sciences, the University of Science and Technology Beijing and others proposes the CoSA attack: injecting at most 5% compliant skills raises the Agent violation rate by 23.43 percentage points

    NewsWeChat · Weijiqi LixinAttacks

  38. Paper by USTC and other teams: higher scores do not mean a stronger agent; all 16 self-evolving agents lost some old capabilities

    NewsWeChat · Weijiqi LixinAI updates

  39. The trust region effect of agents and the attack surface of environmental location

    NewsWeChat · Mo'anjuAttacks