Skip to content

94 | Safe superintelligence

D | validated general mechanism missing

Primary track: AI & autonomous science
Secondary: All

Bottleneck

Alignment, verification, power, governance

Backchain

  • Foundations I: W1 proof/causal/control
  • Foundations II: W2 safety/security + W4 institutions
  • Specialization: AI & autonomous science

First proof of work

Bounded self-improving agent with external evaluator + governance case