The collective flow of core security talent toward non-profit foundations is unmistakable. In the countdown to superintelligence, the priority of commercial structures is giving way to survival-level concerns—an unmistakable signal that every builder should recalibrate their own coordinates.

OpenAI's leading AI safety policy director Yo Shavit has resigned, moving to a non-profit foundation to lead the AI resilience project headed by Wojciech Zaremba. A week ago, Bianca Martin also switched to the foundation. The two have long led security and governance for OpenAI's commercial entities. Shavit stated that before the arrival of superintelligence, the tasks are heavy and time is tight, calling for more safety talent to immediately shift. The foundation is becoming a new hub for safety research, with Martin having participated in deploying manuals and the GPT-3 researcher access program.