The Case File Library
Publicly documented quality failures, dissected for lessons — case file by case file. Recalls, groundings, outbreaks: what actually failed, and which discipline would have caught it.
Chernobyl: how a delayed safety test met an unforgiving design
On the RBMK, the emergency shutdown button briefly made the reactor run hotter. A look at how a postponed turbine test met a design that amplified every mistake.
St. Jude Riata: how a silent wear mechanism defeated every monitoring system
Here is an uncomfortable observation: every monitoring system attached to the Riata lead worked exactly as designed, and roughly 128,000 patients were still carrying cables wearing through their own…
Uber Tempe: how suppressed braking and a phone killed Elaine Herzberg
The situation Uber's Advanced Technologies Group was testing on public roads in Arizona because Arizona asked the fewest questions. California's regulator had all but shown them the exit; Phoenix…
Jeep Grand Cherokee: how a packaging decision outran crashworthiness
Here is an uncomfortable observation from two decades of gateway reviews: the fuel tank never wins a packaging argument. It has no advocate in the studio, no line on the option sheet, no seat at the…
Flint: how a skipped corrosion inhibitor poisoned a city
Nobody in Flint shipped a bad part: a water-source swap skipped ~$100/day
Three Mile Island: how a misleading indicator caused a partial meltdown
When a control panel indicator shows what was
Virgin Galactic: how a missing interlock turned human error fatal
I have sat in enough 8D review rooms to know the pattern. The room goes quiet. The timeline goes on the wall. The discussion drifts toward "operator error." Someone reaches for stricter procedures, a…
Southwest 1380: how a fatigue fracture defeated engine containment
Borescope inspections caught visible wear for years—until they missed the micro-fatigue growing underneath. How a latent defect
DC-10 cargo door: how a false-lock design killed 346 people
The DC-10's cargo door design relied on a visual check that confirmed the lock cycled—not that it actually engaged—leading to two crashes and 346 deaths.
Apollo 13: how a dropped tank and a voltage mismatch caused a deep-space failure
Here is an uncomfortable observation: catastrophic failures rarely come from a single dramatic event. They come from a chain of small decisions, each defensible in isolation, that collectively strip…
Heartbleed: how unvalidated infrastructure became a global failure
Heartbleed exposed how the internet's most critical security infrastructure ran on unaudited, underfunded code — and why nobody asked the questions that matter.
Knight Capital: how a dead code flag bankrupted a firm in 45 minutes
When a dead code flag reached one old server out of eight, it woke an algorithm that bought high and sold low, bankrupting Knight Capital in 45 minutes.