Returns to Intelligence Are Nonlinear

Author: @scaling01 | Date: 2026-05-02

Thesis that returns to intelligence are nonlinear because decisions are path-dependent. Small intelligence advantages compound over time through better decisions, each opening new opportunity spaces unavailable to less intelligent agents.

Key Claims

  • Intelligence returns are nonlinear (superlinear), not diminishing
  • Path-dependence is the mechanism: better decisions unlock better future decisions
  • Small capability gaps compound over sequential decision chains
  • This applies to both AI systems and organizations deploying them

Takeaways

  • Explains why marginal model improvements matter more than they appear
  • First-mover advantage in AI capabilities may compound rather than erode
  • Path-dependence means you can’t “catch up” by just matching current capability — you’ve missed the decision tree
  • Strategic implication: invest heavily in intelligence early, returns accelerate

Linked Concepts