Beware Even Small Amounts of Woo

A viral essay by researcher J Bostock, posted on the rationality forum LessWrong, issues a stark warning to the tech and AI community: the embrace of 'woo'—a cluster of neo-pagan, Buddhist-adjacent, and tarot-ish beliefs popular in liberal-left tech circles—poses a critical threat to clear thinking about artificial intelligence. Bostock analogizes woo to alcohol or a well-adapted religion; for most people in everyday life, small amounts are socially lubricating and harmless. However, he argues that reasoning about AI and the technological singularity requires 'world-class epistemics,' a state where any reliance on intuitive, feeling-based 'just vibe with it' reasoning becomes a dangerous liability.

The core of the argument is that woo trains a 'mental motion with poor form'—following unexamined intuitions toward strong feelings. While smart individuals can usually subordinate intuition to logic on well-trodden topics, AI presents a novel, immensely complex environment with no clear logical paths. Bostock observes that several otherwise sharp thinkers exhibit 'uncharacteristically poor thinking' specifically when discussing AI, a failure he links to woo-adjacent mental habits. The post concludes that in the pre-singularity era, where missteps could be catastrophic, the tech community must be vigilant in excising even small amounts of epistemically corrosive belief from its reasoning processes.