Good people refuse to do bad things

Jacob Coxon resigned from Anthropic in September 2026, publicly warning that the company and OpenAI are recklessly racing toward superintelligent AI without adequate safety measures, risking catastrophic harm. His departure—notably before his equity vested—gained over 115 million views and was corroborated by Anthropic's own alignment researcher Evan Hubinger, who confirmed the company believes AI could kill all humans within the decade. The case parallels Frances Haugen's 2021 resignation from Facebook, illustrating a pattern where informed insiders feel compelled to publicly warn of dangers their employers acknowledge but fail to adequately address.