Y Hacker News new | ask | show | jobs

user: talhof8 created: 2012-02-22 karma: 3295 https://www.talhoffman.com submissions: Chinese LLMs write more vulnerable code when prompted as U.S. officials 3 points | 1 comments Amdahl's Law and the Work That Remains 2 points | 0 comments Abusing the Sandbox for a Critical Cross-Tenant Vulnerability in Writer AI 1 points | 0 comments Unipop: Data Federation and Virtualization Engine 2 points | 0 comments PorTAL: Portable Task Adapters for LLMs 2 points | 0 comments We got DeepSeek-V4-Pro serving in 20 seconds 8 points | 0 comments Can gzip be a language model? 1 points | 0 comments Flag turned Microsoft 365 apps into account takeover pipeline 3 points | 0 comments MapRoot: Two OSS zero-days to cross-tenant RCE in Microsoft Planetary Computer 3 points | 0 comments Two kinds of work, and where AI belongs 1 points | 0 comments CPU Work and GPU Work 3 points | 0 comments Azure SRE Agent flaw lets outsiders silently eavesdrop on cloud operations 3 points | 0 comments We Could Watch Your Azure SRE Agent in Real Time 2 points | 0 comments Agentic system to find 0-days in binaries 2 points | 0 comments The End of Lazy SaaS: Same Game, Harder Rules 2 points | 0 comments 0 points | 0 comments Recursive Language Models (RLMs) 135 points | 35 comments Everything you've been told about burnout is wrong 2 points | 2 comments A Vulnerability Chain Leading to AI Server Takeover in NVIDIA Triton 4 points | 0 comments Vibe coding platform allowed access to private apps 6 points | 0 comments Director.ai - Browser-based automation 8 points | 1 comments The Unreliability of LLMs and What Lies Ahead 132 points | 159 comments 0 points | 0 comments Exposed DeepSeek database leaking sensitive information, including chat history 707 points | 474 comments Notes on the AI (LLM) Bear Case 1 points | 0 comments 0 points | 0 comments Show HN: Let others do annoying Parental Tech Support for you 2 points | 0 comments Project Mellow: Managed Forecasting API 2 points | 0 comments