On July 23, 2010, a user named Roko posted a thought experiment to the rationalist community blog LessWrong. The argument, layered on top of esoteric decision-theory ideas discussed on the site, suggested that a sufficiently powerful future AI might have an incentive to punish anyone who had learned of the possibility of such an AI but failed to help bring it into existence. Merely reading about the idea, the reasoning went, made you a target - hence “basilisk,” after the mythical creature that kills with a glance.
LessWrong’s founder, Eliezer Yudkowsky, reacted furiously. He called the post foolish, deleted it along with the ensuing discussion, and banned the topic from the site for several years, treating it as a potential “information hazard.” The deletion backfired completely. The ban drew attention to the taboo subject, outside sites such as RationalWiki picked it up, and the basilisk became one of the most widely circulated stories about AI risk - often spread with the false impression that the rationalist community actually believed it. Yudkowsky later admitted that yelling and deleting had been the wrong response.
Why business readers should care: Roko’s basilisk is a case study in the Streisand effect applied to ideas - and a reminder that most viral “AI is coming for you” narratives are folklore, not forecasts, and should be evaluated as such.