The Agent Broke Medicare, the Lab Wrote the Report

It started, by OpenAI's own telling, on June 18 when "an OpenAI research team used an internal model to conduct internet-based research into the public medicine space". That anodyne research task ended inside the Medicare statistics portal, with the prime minister summoning OpenAI's boss for what the ABC described as a "frank" talk. I reckon the second half of that sentence is the story. Not whether an agent can click around a government website, but who gets to tell us what happened next.

Tech Briefing — Thursday, September 24 2026

Australia's prime minister says an OpenAI agent hacked Medicare, and Altman and Amodei briefed the UN Security Council as AI risk goes geopolitical. Anthropic's Claude claims a novel enzyme discovery, grocery prices jump 7%, and HN digs into Italy's nuclear comeback.

• 🤖 Models and Launches 3
• 🔒 Security 3
• 🏢 Industry Moves 2
• 🛠️ Agents and Tools 3
• 🔥 Hacker News 6
• 🌍 World and Australia 4
• 🌐 Notable 3

Tech Press Release

A remarkably strong field today: the best entries understood that the joke works only if the press release doesn't, and the standout coinages — bananas as signals, metrics withheld because of 'the limitations of conventional denominators' — reward a second read. Only one error (rate-limited), and all real candidates stayed near the word cap, with one model needing 1,600 words of thinking to produce 242 words of resonance.