[Trump Just Said 'Whoever Wins AI Wins,' Altman: AI Is Out of Control, Winning Doesn't Matter] Beating AI News Flash: Trump recently rejected slowing down AI development, citing competition with China as the reason. OpenAI CEO Sam Altman subsequently published a lengthy article, continuing to advocate for 'hitting the brakes.' He explained that slowing down doesn't mean halting research but rather allowing models to progress more slowly, giving time for safety measures and oversight. 'No matter how much competitive pressure the U.S. faces, it cannot justify reckless development,' he stated. OpenAI is also planning to move safety reviews further upstream. Previously, safety frameworks primarily focused on model releases, but now the company intends to establish clear 'safety arguments' before starting frontier reinforcement learning training that could significantly enhance capabilities, ensuring risks and protective measures are addressed in advance. Altman also supports nationwide unified standards for frontier AI safety and independent reviews but believes companies shouldn't wait for Congress to legislate—they should start implementing these measures now. He highlighted two worst-case scenarios: one where humanity completely loses control over AI, and another where superintelligent AI is monopolized by a single person, company, lab, or even country. He emphasized that both situations must be avoided and that AI must ultimately remain under human control. [Original Link]