2025 / Free online book · The AI-Native Canon
The Spec Is the Program
Intent as the new source code
Access
Free
Chapters
9
Read time
79 min
Alpesh Nakrani
CRO at Devlyn · former CTO & COO
Code used to be the durable asset and the spec a throwaway. AI-Native inverts it: the implementation is regenerated on demand and the specification is what you version, review, and defend. This is a working grammar for writing intent precisely enough that the machine builds the right thing and you can prove it did.
When a model writes the implementation, the specification becomes the artifact you actually maintain. How to write intent that survives a model upgrade.
This edition is free to read onsite. Each chapter has its own URL, so readers can bookmark, share, and return to the exact section they need.
Table of contents
01 The Feature That Passed but Failed The incident began as a success story. A product manager wrote a short ticket for a billing feature: "Let admins export invoices as CSV from the account page." The team fed the ticket to an AI coding agent inside the repository. 11 min 02 A Prompt Is Not a Spec A prompt is a request. A spec is a contract. 9 min 03 SPEC-Lock SPEC Lock is the working method of this book. It is intentionally small enough to remember and strict enough to change behavior. 8 min 04 Examples Are the Executable Core Most arguments about requirements disappear when people write examples. 9 min 05 Constraints as Source Code Features tell the system what to do. Constraints tell it what must remain true while doing it. 9 min 06 Properties, Contracts, and Invariants Examples catch known cases. Properties catch classes of cases. 9 min 07 Accepting Machine-Authored Diffs Code review changes when the author is a machine. 9 min 08 Living Specs and Spec Drift A spec that is correct on launch day can be wrong six months later. 9 min 09 Conclusion The promise of AI-assisted software development is not that prompts replace engineering. The promise is that implementation effort can fall enough for teams to spend more attention on intent. 2 min A Appendix A: Glossary Acceptance The evidence that decides whether implementation is good enough to use. 2 min B Appendix B: Source Index Intent as the new source code 2 min