LW - DeepMind's "Frontier Safety Framework" is weak and unambitious by Zach Stein-Perlman | The Nonlinear Library | Podwise