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
- Instrumental Convergence Theory — Nick Bostrom isolates basic convergent instrumental goals, proving that any intelligent agent will n...
- Markov Decision Processes — Richard Bellman formalizes the mathematical framework for modeling optimization problems where outco...
- What Is the Human Soul in a World of AI and Perfect Machines? — You can break the body down into atoms and the brain into synapses, but you won't find love, faith,...
- 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...
- The Value of a Handshake in a World of Digital Contracts — We live in the era of the frictionless transaction. With a single click, a biometrically scanned th...
Discussion: