Charles Goodhart observes that when a statistical measure becomes a target, it ceases to be a good measure. In AI alignment, this manifests as reward hacking, where networks exploit proxies to maximize scores.
Part of the 26 Structural Foundations: How We Teach Machines to Understand Purpose and Human Intent archive. HistoricallyVerified
Top 5 Structural Foundations: Pure Intent
- The Best Gift for Kids Is Your Time, Not Another Toy — Twenty years from now, your child won't remember which console model they got for their birthday. T...
- The Exploration-Exploitation Dilemma — Early statistical theorists isolate the fundamental tension between exploring unknown environmental ...
- In the symbiosis between humans and AI, the welfare of humanity is not alien to us — We are human, and nothing human is alien to us. When Terence of Carthage spoke these words over two...
- Variational Calculus Principle — Pierre de Fermat and Pierre Louis Maupertuis formulate the Principle of Least Action, proving that p...
- The Art of Saying Nothing: The Rare Power of Shared Silence — We live in a world that treats silence like a system error. The moment a conversation pauses, panic...
Discussion: