OpenAI Chief Scientist Jakub Pachocki says no AI lab has solved alignment well enough to keep scaling at maximum speed, according to a Techmeme item summarizing an OpenAI post authored by Pachocki. The item says Pachocki also hopes voluntary slowdowns become commonplace. That is a narrow but notable claim: it does not describe a binding rule, a new OpenAI policy, or a specific slowdown already underway. It frames the issue as a judgment from OpenAI’s chief scientist about the state of alignment relative to the pace of model scaling. The provided summary does not include Pachocki’s technical evidence, the specific alignment failures he has in mind, or what thresholds he believes would make faster scaling acceptable. It also does not say whether OpenAI has committed to a particular voluntary slowdown mechanism. For now, the story is best read as a public position from a senior OpenAI research leader: alignment remains unsolved at the level needed to remove constraints on scaling, and voluntary restraint is being presented as a norm he wants to see more broadly adopted. Who benefits: Too early to tell from the provided material. The item only says Pachocki hopes voluntary slowdowns become commonplace. Who's exposed: Too early to tell from the provided material. The item does not identify any specific company besides OpenAI.