Spec-driven Development: How AI Changed Everything (And Nothing)
summary
Argues for AI Unified Process and spec-driven development that keeps structured requirements as the source of truth and uses AI to generate diagrams, entity models, code and tests; demos include reverse-engineering Spring Pet Clinic into system use cases and an entity model, then generating code and tests that run in CI pipelines. Uses Vaadin, Jooq, Cloud Code skills and skills and MCP servers as the agent harness. Claims modernization across 1,200 DB tables and 500 microservices, tests at ~100 ms vs 2–3s, and Gartner’s ~5% net speedup while warning about prompt injection and human review.
AI-written summary · about
chapters (16)
- 0:00Spec-driven development intro
- 4:15Model vs spec-driven dev
- 7:03AI Unified Process explained
- 10:18Domain model & use cases
- 13:09Use cases vs user stories
- 16:03Reverse engineering Spring Pet Clinic
- 20:55Agent harness & code gen
- 23:24Reverse engineering systems
- 26:17Demo: use case implementation
- 31:54Setting up AI development guidelines
- 34:39Human review & risk management
- 36:31AI impact on architecture
- 38:40Self-contained vs microservices
- 41:08Testing use cases
- 41:28Future of spec-driven dev
- 42:53Conclusion & Q&A
related talks
44mSpec-driven Development: How AI Changed Everything (And Nothing)
45mSpec-Driven Development: the New Era of Development
47mSpec-Driven Development with AI Agents From High-Level Requirements to Working SW
18mUsing Spec-Driven Development for Production Workflows
25mDemystifying Spec Driven Development
47mThe Renaissance Developer: Building Better, Safer, & Quicker with AI-Powered Spec-Driven Development
37mSpec Driven Development: AI Assisted Coding With Doc at the Core
28mHow To Use Spec-Driven Development for Production Workflows