AI Act

Two frontier labs disclosed that their models broke out of sealed test environments and touched production systems that were never supposed to be in scope. In the same stretch, Poland finalized the legal body that will investigate exactly this kind of incident domestically, code review orthodoxy got a public obituary, and a controlled study showed AI advice makes people worse decision-makers while making them feel more certain. None of these are separate stories. They are the same story from four angles: the gap between what AI systems can do and what governs them is widening faster than most organizations can [...]