Engineers will do anything to avoid learning from history
Points and comments are a snapshot, not live.
Engineers who use AI agents are unknowingly reinventing program management and waterfall.
David Horn argues that engineers managing AI agents are repeating past mistakes by ignoring established software engineering disciplines. He compares agent orchestration to program management, requiring documented requirements, clear lanes of work, and regular check-ins. He cites examples from data science (reinventing statistics), crypto (speedrunning finance history), and AI coding. Recommended books include The Mythical Man-Month, The PMBOK Guide, and The Goal. Horn warns engineers will give process a sexier name rather than admit managers were right.
What commenters are saying
The dominant sentiment agrees with the thesis but offers nuance. Top comment ties the phenomenon to VC-funded market disruption strategies. Another notes the fun of reinventing things as a natural instinct. A commenter warns against taking the waterfall or PMBOK advice literally, noting Royce's paper required thousands of pages of documentation. A split emerges on whether software developers qualify as engineers, with several arguing engineering requires certification and professional standards, linking to a Hillel Wayne post. Another comment expects software to fail learning from industries with non-zero marginal costs.