---
id: "claim-skills-compound"
type: "claim"
source_timestamps: ["08:28:00", "08:35:00"]
tags: ["mental-models"]
related: ["concept-skills-vs-prompts", "quote-skills-compound", "contrarian-prompts-dont-compound"]
confidence: "high"
testable: false
speakers: ["Nate B. Jones"]
sources: ["s43-file-format-agreement"]
sourceVaultSlug: "s43-file-format-agreement"
originDay: 43
---
# Skills compound over time; prompts do not

## Claim

Because skills are **persistent, version-controlled files**, they can be continuously refined, tested, and improved based on real-world failures. This allows their value to compound. Prompts, being ephemeral text blocks, do not benefit from this systemic iteration.

## Confidence: High · Testable: No (mostly conceptual; partial empirical support)

## Validation (Enrichment)

Conceptually valid but unquantified. Version-controlled skills enable iterative refinement, unlike ephemeral prompts, mirroring software-engineering practices in MLops. No direct benchmarks, but compounding via org repos (e.g., [[entity-product-openbrain]]) is emerging in dev tools like [[entity-product-cursor-d43]].

## Counter-Perspective

Critics (including some Simon Willison follow-ups) argue skills are *just versioned prompts with YAML* — the compounding is real but modest, and is bottlenecked by the same LLM nondeterminism that limits prompts.

## Related

- [[concept-skills-vs-prompts]]
- [[contrarian-prompts-dont-compound]]