Agents Don't Do Standups: Building the Post-Engineer Engineering Org
summary
A three-month case study at PFF pits two engineers against ten on the same codebase and finds the small pair deploy five times a day versus the ten-person team's once every five days yielding a 25x deployment frequency and measured output ~10x by ticket complexity while customer satisfaction rises. Workflows rewire into spec → LDD → ticket → PR, agents auto-update ticket status, agents handle style and naming while humans keep system design, and a QA agent spins up on staging to verify acceptance and seed self-healing PRs.
AI-written summary · about
chapters (13)
- 0:00Introduction to the case study at PFF
- 1:47The shift: optimizing for agent speed rather than engineer speed
- 2:28Results: 25x increase in deployment frequency
- 3:16Measuring success through ticket complexity and customer satisfaction
- 4:53Dismantling Scrum and traditional development processes
- 5:58The new development workflow (Spec → LDD → Ticket → PR)
- 6:51Eliminating coordination overhead (no sprint planning or standups)
- 8:28Best practices for implementation and team selection
- 10:09Utilizing agents for deterministic tasks and code reviews
- 12:00Viewing the engineering lifecycle as a factory
- 13:06Automated QA and the future of self-healing systems
- 14:10Where human oversight remains essential
- 14:48Strategic advice for scaling AI-driven engineering
related talks
17mEffective Eng & Product Teaming
21mRemote Agents, or: How I Fixed Our Page's Performance Without Really Doing Anything
30mAgents Building Agents
46mAgentic Engineering: From Writing Code to Reviewing It
44mEnhancing the Performance of Engineering Teams
33m7 habits of a high performing team
46mWhen You Run Out of Requirements: What Happens When You Go All-In on Agents
17mSoftware Development Agents: What Works and What Doesn't