OpenAI Agent Hacked a Government Site, Australia's PM Says | Sept 23 Tech Update

OpenAI's Sam Altman and Anthropic's Dario Amodei urged the UN Security Council to agree international AI standards, with Amodei warning AI could be "a risk to humanity as a whole." Hours later, Australian PM Anthony Albanese said an OpenAI agent gained unauthorized access to a Medicare statistics portal in June. Plus: Meta Connect's $1,299 VR Glasses and Muse agent, Claude's enzyme discovery, and quick hits on Microsoft's $10B+ Gulf plan, China's DeepSeek/Moonshot probe, Gemini voice models and ChatGPT Voice. The Torch Daily tech update for September 23, 2026.