How Flow Works & Other Curiosities
summary
Argues that work is information flowing through value streams, models it with agent‑based simulation (NetLogo, a C behavior‑tree prototype, and epirust/baritsim for epidemics), and treats AWS/IaaS as an M/M/c queue where c→∞. Shows a real value‑stream with 42 days wait vs 5 days of work, demonstrates that single‑piece flow outdelivers waterfall by orders of magnitude, and claims batch size reduction shortens cycle time by orders of magnitude; warns platform teams that superficial self‑service reintroduce human bottlenecks.
AI-written summary · about
chapters (13)
- 0:00Intro
- 4:44Programming in the past 10 years
- 15:24LLMs for developers
- 15:44Building systems optimizing for flow
- 20:26Are we real engineers now?
- 21:33Work flows as information through value streams
- 26:45Fun fact
- 28:38Sidebar on generative science
- 33:11The Gini coefficient
- 34:09Back to following the work
- 43:38Using generative science to model the wold of work
- 48:06How flow works
- 49:20Outro
