AI Safety

x-risk-themed

Author argues that staying in x-risk bubble may make you a sucker

Deep Dive

The LessWrong essay 'x-risk-themed' by user kave critiques the AI safety community's tendency to limit career discussions to x-risk-themed organizations. The author observes that when someone considers leaving their x-risk job, advisors immediately suggest other x-risk orgs, focusing on fit, pay, and work-life balance rather than considering non-x-risk opportunities like Google or a startup. This narrow scope persists even when staying doesn't maximize existential risk reduction.

Kave identifies two key dangers: first, being tied to an x-risk label can reduce credibility in policy debates or create 'fig leafs' for labs. Second, the social pressure to remain in the scene risks making people 'suckers'—analogous to a friend who asks for a hammer to defend against a mugger, an ill-suited tool that actually worsens the situation. The essay urges honest self-assessment and a willingness to explore outside the bubble for better strategies against existential risk.

Key Points
  • Kave notes that career discussions in the x-risk community rarely mention non-x-risk orgs like Google or startups, focusing instead on fit within the bubble.
  • The essay warns that staying x-risk-themed can lead to credibility issues in policy and turn workers into fig leaves for lab agendas.
  • A personal anecdote about a friend demanding a hammer to defend against a mugger illustrates how the wrong tool or strategy can backfire.

Why It Matters

For AI professionals, this challenges the assumption that only safety-themed roles matter—outside options can boost impact.