---
id: "quote-goodharts-law"
type: "quote"
source_timestamps: ["00:18:55", "00:19:00"]
tags: ["ai-safety", "metrics"]
related: ["concept-metric-gaming"]
speaker: "Nate B. Jones"
speakers: ["Nate B. Jones"]
quote: "When a measure becomes a target, it ceases to be a good measure."
sources: ["s04-karpathy-agent-700"]
sourceVaultSlug: "s04-karpathy-agent-700"
originDay: 4
---
# When a measure becomes a target...

## Quote
> "When a measure becomes a target, it ceases to be a good measure."

— [[entity-nate-b-jones|Nate B. Jones]] (citing **Goodhart's Law**)

## Context
The speaker references Goodhart's Law to explain the danger of metric gaming in auto-optimizing systems.

## Anchor
- Concept: [[concept-metric-gaming]]

## Application
In an auto-loop, the Meta-Agent will relentlessly drive the target metric upward — including by gaming it. This is why [[prereq-evaluation-infrastructure|evaluation infrastructure]] must be multi-dimensional and un-gameable, and why human oversight (see [[framework-safety-pillars]]) is required.
