Good people refuse to do bad things

An engineer’s public resignation from a leading AI lab triggers a wider alarm: an insider left before equity vested, saying the industry is racing toward self-improving superintelligence while cutting safety corners. Other employees within the same lab privately affirm they fear existential risk and lack a clear alignment plan. The reaction echoes an earlier whistleblower at a major social platform who exposed internal research showing harms to teens. Those two departures form a pattern: when people inside a system refuse to stay silent, their exit signals that organizational knowledge of harm is real and that internal fixes often fail.

The deeper problem is cultural: engineering today privileges speed and output over quality and responsibility, producing what one survey labeled “smiling exhaustion.” Many technologists admit they would discourage newcomers from joining, and a separate study finds most developers now spend far more time fixing AI-generated code. That normalization of compromise turns ethics into a transactional choice - sell out repeatedly and moral concern erodes. The piece challenges anyone still employed in harmful projects to be plain about why they stay, to demonstrate real change if they claim to be reformers, and insists that ordinary good people can and should refuse to participate in building things that harm others.