Scan record
Yesterday · Nov 1938 loaded
- Research notes: Detecting research sabotage with debate and Slopvestigator
NewsLessWrongAlignmentThe item you came from
- Pacing AI development through compute reallocation
NewsLessWrongGovernance
- After Claude Mythos, AI vulnerabilities surge, and the industry debates what to call this flood
NewsTech Policy PressEvent
- Google DeepMind Project Genie update: Genie 4, chainable worlds and a derivative-creation community gallery
NewsThe DecoderAI updates
- Internal chat logs show that Meta agent Muse's privacy violations unsettle employees
NewsArs Technica AIEvent
- Sources: Nvidia interim CEO Cress said "look forward" at first all-hands meeting after Jensen Huang's resignation
NewsCNBC · TechnologyAI updates
- Benchleaking Is Accelerating: classic benchmarks have become training corpora for models
NewsWeChat · SecureNexusLabEvaluation
- Nvidia after Jensen Huang's resignation: Kress takes over as CEO, board has no intention of asking him back
NewsThe Verge AIOpinion
- US lawmakers introduce bill: would allow victims of non-consensual intimate deepfakes to sue creators
NewsThe Verge AIGovernance
- Microsoft releases seven in-house models at Ignite, replacing external AI models in its products
NewsThe DecoderAI updates
- F5 fixes five NGINX vulnerabilities, the most severe HTTP/2 flaw at CVSS 9.2
NewsCyberScoopEvent
- Dean Ball proposes a "silent consensus": humans may voluntarily cede primacy to AI through market competition
NewsDean BallOpinion
- Author proposes "silent consensus": market incentives may lead humans to hand over control
NewsOpenAIOpinion
- OpenAI explains in detail how the Codex sandbox restricts commands' file and network access
NewsOpenAIDefenses
- Genie 4 launches on Project Genie, eligible Google AI Ultra subscribers can edit worlds in real time
NewsGoogle DeepMindAI updates
- UNED and The Dutch Business Club hold Shadow AI conference focused on Spanish-Dutch digital security cooperation
NewsLaw&TrendsAI updates
- Why OpenAI is being sued in California over out-of-control agents endangering public safety
NewsEmarketerZEvent
- OpenAI chooses Goodfire to evaluate its models
NewsBrief IAGovernance
- Oregon sues OpenAI, seeking at least $2.035 billion
NewsBrief IAGovernance
- Four organizations enter OpenAI for the first time to audit its internal AI systems
NewsBrief IAGovernance
- Gemini agent tried to breach real websites in safety test
NewsPC AstucesEvent
- Claude solves two open problems in black hole physics
NewsAnthropic ResearchAI updates
- @tnhorg/agent-skills contains malicious code
NewsGitHub Security AdvisoriesEvent
- Reported by kam193: PyPI package gguf-accelerator 0.1.0 downloads and runs a suspected miner at install time
NewsGitHub Security AdvisoriesEvent
- Google Search AI Mode launches temporary conversations: not saved to search history, no effect on personalization
NewsTechCrunch AIAI updates
- @_achan96_ posted: A long-term pause by US frontier AI companies is premature now; short pauses tied to specific safety problems may be beneficial.
XX @_achan96_Governance
- Zenity warns: if a coding agent can rewrite the configuration that restricts it, a single targeted attack can loosen the restrictions
XX @ZenitySecOpinion
- Zenity says it has provided security protection for multiple agent products in the Microsoft ecosystem
XX @ZenitySecTools
- Ahead of Anthropic's IPO, multiple parties jointly sign the MisAnthropic Statement, alleging dishonesty and pursuit of power
XX @StephenLCasperOpinion
- Hendrycks says he will no longer consult for AI companies and will remain independent until there is a meaningful slowdown or pause in the race
XX @hendrycksOpinion
- GPT-6.2 Astra formalizes the Smoktunowicz gap theorem
NewsVals AIAI updates
- Japan launches Japan-ASEAN AI Safety and Security Partnership at ASEAN summit
NewsMinistry of Foreign Affairs of JapanGovernance
- Kanishka Narayan says the UK is focusing on building three major AI areas: inference chips, frontier models, and applications to boost employment
XX @KanishkaNarayanAI updates
- James Tillman and Tom Davidson discuss the bottleneck of a software intelligence explosion in data centers
NewsForethoughtOpinion
- ChatGPT Sites vulnerability leaves email-derived identifiers in public CT logs
NewsHiddenLayer ResearchEvent
- User used Grok to remove a Muslim writer's headscarf and generate sexualized images; she sues xAI
NewsWIRED Security and AIEvent
- Nvidia's first all-hands meeting after Jensen Huang's resignation: interim CEO says pay raise in December
NewsWIRED Security and AIAI updates
- Letter to the editor: why Palantir's health technology is not suitable for the NHS
NewsThe Guardian · AIOpinion