From SDLC to ADLC: agents quietly take over planning, coding, testing, and deployment
blaizedsouza · x · 2026-11-12
The author argues AI agent-enabled coding is becoming the new software development lifecycle — an "ADLC" replacing the classic sequential SDLC.
Key shifts:
- Driver: human execution → autonomous agents that write, refactor, and test code
- Planning: fixed scope → evolving goals and PRDs
- Speed: sequential handoffs → parallel sub-agents
- Testing: post-development → real-time feedback loops
The result is not a pipeline but a live development system — billed as the biggest shift since the GUI.
More from coding & agent
- curl faces proposal to ban LLM-generated code as anti-AI pressure hits open source — steipete · 2026-11-12
- Google Cloud opens its internal agentic training as the Professional Agentic Architect certification — rseroter · 2026-11-12
- In defense of slow Gemini: async coding favors strong models over fast-but-dumb ones — Aizkmusic · 2026-11-12
- LangChain CEO's Interrupt London keynote: deep agents, JEV and owning your intelligence — hwchase17 · 2026-11-12
- Player's homemade agent bot grinds WoW to level 13 while he never plays — shakoistsLog · 2026-11-12
- Fable 5.2 runs overnight to build real-time Toronto bird sim with map 7x bigger than GTA6 — zeeg · 2026-11-12