Fixing the Ultimate Legacy System: You
summary
Software development fails when mental models age; fix by slicing work into hour-sized stories, developing a tracer-bullet end-to-end thin thread, committing to a single mainline so code stays continuous deployable mainline, and running CI/CD with rapid user feedback. Avoid long feature branches and pull-request handoffs. Empirical gains include studies showing ~50–60% productivity lift from tools like GitHub Copilot. Expect prompt-driven assembly (image models, Copilot) to become a new role: prompt engineering, coordinated by a three-track attack of deliver, discover, refine.
AI-written summary · about
chapters (13)
- 0:00Start
- 9:59Amygdala
- 12:26Communication and Learning
- 27:49Fundamental Attribution Error
- 28:26Need for Closure
- 31:23Free Information Flow
- 33:42Corporate Accounting
- 34:17Learning To Adopt System Thinking
- 35:03Virtuous Cycle
- 37:55Black Swan Event
- 41:44Moving More towards System Integration Instead of Custom Code
- 42:09The no Code Movement
- 52:49Do You Believe that Systems Engineers Have To Start with Quantified Critical Qualities like Security Usability Adaptability
related talks
23mAn ensembling approach to making legacy code habitable
01h 03mHow To Do Kick-Ass Software Development
01h 00mThe Influential Software Engineer
48mThe Effective Software Architect - Work Smarter, not Harder
21mNeed for speed - Moving fast as a software engineer
50mThe Effective Developer - Work Smarter, Not Harder
52mMaximizing Productivity
01h 02mGetting Unstuck. Strategies for Surviving in a Rapidly Changing Post-Agile World