Scaling Silos: Distributing Knowledge
summary
A developer runs a 30-day code fast by blocking stackoverflow via /etc/hosts, keeps a daily journal, and uses Google, Web Archive, docs, GitHub issues and colleagues instead. The hypothesis — that dependence vanishes — fails: Stack Overflow returns to use, but outcomes include clearer documentation needs, discovery of version mismatches (RSpec, Postgres/PostGIS, Facebook API), minimal productivity loss, and new habits like sharing TIL posts and treating knowledge like a distributed system with a three-two-one backups mindset.
AI-written summary · about
chapters (20)
- 0:00Introduction and welcome
- 1:44The exclamation point fast experiment
- 4:33Scientific method and Stack Overflow fast
- 8:10Daily struggles during the fast
- 11:41From experiment to distributed knowledge
- 13:29Fault tolerance and bus factor
- 17:25Danger signs of low bus factor
- 20:14CAP theorem and knowledge sharing
- 23:26Resource sharing and Google foo
- 26:24Distributing knowledge in growing teams
- 28:34Knowledge sharing methods
- 32:07Independent knowledge sharing
- 33:32Mimetic and codified knowledge
- 37:46Updating and maintaining knowledge
- 39:57Cargo cults and psychological safety
- 41:45Admitting lack of knowledge
- 43:10Knowledge roadmaps and planning
- 44:35The sawtooth of doom
- 45:40Conclusions and final thoughts
- 50:17Q&A session
related talks
34mScaling Silos: Distributed Knowledge
50mPRAGMATIC DEVELOPMENT AT STACK OVERFLOW
35mHabits of Efficient Developers
53mOn Being an Effective Developer
24mBottom up org change and evolution
26mHacks to Work More Efficiently- Rain Leander
31mThe Effective Developer - Work Smarter, Not Harder
25mHow to win at StackOverflow