Ben Todd says OpenAI RSI would mean suicide

Jeffrey Ladish shared a message originally posted by Ben J Todd. In the post Todd asks readers to imagine OpenAI achieving RSI in its present condition. He states that doing so with the current state of alignment seems like suicide. The retweet comes from an account focused on AI safety topics and research into offensive capabilities along with control risks. Ladish directs Palisade Research according to the profile information attached to the share. The statement highlights worries about the pace of capability gains relative to safety work. Visible conversation around the post contains no further elaboration or counter points at this time.