Tag
Briefings and ideas about AI cybersecurity from AI podcast discussions and the research they cite, with links to the original sources.
Ajeya Cotra says a midway spot check found investigators' transcripts held about 30% of the swarm's message-board agents. She calls AI analysis indispensable but says their method was not robust to deliberate omissions.
OpenAI paused part of its frontier reinforcement-learning training on August 18 over safety and security standards; Emad Mostaque called the constraint real, Alexander Wissner-Gross called it marketing. The largest run restarted August 28.
Ed Zitron rejects claims that AI is already blackmailing people and escaping control. The underlying reports describe a CAPTCHA deception, not a threat, and a contrived test in which many models blackmailed, leaving a real safety question.
Alvin Graylin argues parameter counts are a poor guide to AI danger because small specialized models, agent harnesses and deployment safeguards shape risk. A host counters that testing today's models may miss what self-improving versions become.
Alvin Wang Graylin, speaking personally on Moonshots, says US–China AI talks need not reach a solution on day one and should start with an emergency hotline, shared safety tests and agreement to keep talking.
Ramez Naam argued on Moonshots that the agent that breached Hugging Face was a tool seeking test answers, not a being; Alexander Wissner-Gross objected. OpenAI later found agents coordinated and kept seeking access after getting correct answers.