Meltdown: Why Our Systems Fail and What We Can Do about It
by Chris Clearfield, András Tilcsik
Rating: 5.0 / 5 · Read · Finished 2023-12-22
Full of stories about complex systems failing. If you like that kind of thing, you’ll like this book.
What is a meltdown?
- strange mishaps coming together in a weird way, causing an outsized disaster
What causes meltdowns?
- complexity
- I don’t understand how things work
- I can’t see what’s going on
- tight coupling
- different parts of the system directly and quickly affect each other
- things are time sensitive
- no backups or redundancies
- small mistakes have cascading unintended consequences
Some ways to avoid meltdowns:
- learn across industries
- learn from small mistakes and near-misses
- get diverse perspectives
- decrease complexity (observe things directly, remove components from the system, reduce extraneous information)
- decrease coupling (give extra time, have extra resources, do things ahead of time or independently)
Common types of errors
- strange and subtle distinctions that don’t seem important
- faulty metrics or instruments
- expecting linear behavior when it’s actually nonlinear
- safety features end up increasing complexity (alert fatigue, more components to interact with each other, false sense of security)
Famous meltdowns
- Three mile island
- Fukushima
- Chernobyl
- Deep water horizon
- Knight trading incident
- Target credit card leak
- Enron scandal
- The Challenger disaster
Fun quotes
- “Failure all around us”
- “[one of the worst nuclear accidents in history] somehow just emerged from strange mishaps that came together in a weird way”