# Claude Code

#entity #tool #ai

## Summary

Anthropic's agentic coding tool (CLI/agent). The vault's first *tool* entity — created 2026-07-29 because the corpus has begun assigning it roles well beyond writing code, and because it is the harness this vault itself runs on.

## Current Understanding

Three distinct roles the corpus gives it:

| Role | Claim | Source |
|---|---|---|
| **Build engine / commoditizer** | "Everyone has 10–15 great products sitting in Cloud Code" — products become cheap, distribution becomes the scarcity | [[2026-07-26-how-to-build-a-billion-dollar-company-2027]] ([[oskar-hartmann]]) |
| **Full GTM stack** | Landing page, invented company identity, scraped prospect list, cold-outbound script, then the product — the human only picks the service and closes | [[2026-07-29-start-a-business-with-claude-code]] ([[dan-martell]]) |
| **Domain tutor / leveler** | "Whatever business you want to be in, Claude will also tell you how to do it" — collapses the no-expertise barrier to starting | [[2026-07-29-start-a-business-with-claude-code]] ([[dan-martell]]) |

Adjacent, not about the product itself: the [[2026-07-06-sebastian-interview-ai-and-software-engineering|Sebastian interview]]'s enterprise-harness thesis (companies will pay for compliant, managed wrappers around exactly this class of tool — [[future-of-engineering-work]]), and [[eugene]]'s bring-your-own-harness practice.

## Evidence

- Vibe-coding product glut ("Cloud Code" = this tool, near-certainly) — [[2026-07-26-how-to-build-a-billion-dollar-company-2027]]
- The 5-step $0→$1M playbook with Claude Code executing steps 2–4 and the build — [[2026-07-29-start-a-business-with-claude-code]] ([[dan-martell]], owner-attributed 2026-07-29)
- Capability counter-evidence: measured AI coding gains are modest and novice-tilted; experts slowed in RCT — [[2026-07-18-ai-productivity-adversarial-evidence]] via [[ai-productivity-evidence]]

## Related Pages

- [[ai-market-shift]] — the roles table there is the market-side generalization of this page
- [[sell-before-build]] — the playbook this tool is cast as executing
- [[future-of-engineering-work]] — the harness market it anchors
- [[eugene]], [[dan-martell]], [[oskar-hartmann]], [[overview]]

## Contradictions / Uncertainty

- The corpus's claims about the tool trace to exactly **two voices** — [[dan-martell]] (GTM stack, leveler) and [[oskar-hartmann]] (build engine/commoditizer) — both promotional, zero demonstrated cases; the one empirical page in the vault ([[ai-productivity-evidence]]) qualifies the capability premise. The leveler claim, however, targets novices-in-a-domain — the segment the same evidence says gains *most* — so the evidence cuts less against it than against expert-productivity claims.
- "Cloud Code" in the Hartmann source is a transcription of this product's name — near-certain, marked as an inference.

## Next Questions

- Is there any documented (non-promotional) case of a business whose GTM assets were substantially Claude-Code-generated and which reached revenue?
- Where is the line between "Claude tells you how" and regulated/physical domain knowledge it can't stand behind?