OpenAI reportedly disbanded its preparedness team as part of a 'streamlining' process

"Streamlining" is the corporate world's gentlest word for amputation. By reportedly dissolving the team responsible for gauging catastrophic model risks, OpenAI hasn't just cut overhead—it has muted the internal alarm system that was designed to wake us up before the fire starts. In the race to ship AGI, the safety railings are apparently the first thing thrown overboard. This is more than a reorganization; it's an explicit declaration of priorities. When the very unit tasked with asking "what if this goes horribly wrong?" is absorbed into a general velocity push, the message to the rest of the industry is clear: risk assessment is a bottleneck, not a buffer. It signals a grim convergence—where frontier labs start treating existential safety as a PR problem to be managed rather than an engineering problem to be solved. As competitors sprint with reckless abandon, rigorous safety analysis becomes a competitive disadvantage. We're now seeing the terminal phase of the AI arms race: the cost of vigilance is viewed as profit