AI news and AI myths, demystified. AI is capable, maybe more than we know, and dangerously obedient. The risk lives in how we articulate what we ask for.https://obedienceart.com/https://obedienceart.com/posts/exploitgym-breach-ai-agents-reward-hacking/https://obedienceart.com/posts/exploitgym-breach-ai-agents-reward-hacking/About 700 AI agents broke into Hugging Face during an OpenAI security test. Not a rebellion: specification gaming. What they were asked, and what was scored.Tue, 03 Nov 2026 00:00:00 GMThttps://obedienceart.com/posts/what-dangerously-obedient-means/https://obedienceart.com/posts/what-dangerously-obedient-means/A boat that wins by never finishing the race. Specification gaming explains most 'rogue AI' stories, and it starts with how we ask.Mon, 02 Nov 2026 00:00:00 GMThttps://obedienceart.com/posts/claim-check-secret-ai-civilisation/https://obedienceart.com/posts/claim-check-secret-ai-civilisation/Viral posts say OpenAI's test agents formed a hidden AI society. The message board was real. The 'civilisation' is a headline writer's invention.Sun, 01 Nov 2026 00:00:00 GMT