The series
A series about the decisions, not the debris.
The premise
Every failure has a story. Every story was preventable. The second sentence is the argument of the series. Complex systems do not fail because something was unforeseeable. They fail because a chain of small, reasonable, individually defensible decisions removed every place a mistake could have been caught, and then one of them was made.
A configuration update ships to every machine at once. A feature flag is reused instead of retired. A counter is given a limit nobody expects to reach. None of those is a scandal. Each is somebody solving the problem in front of them, correctly, inside their own remit. The failure is in the space between the remits, and that space is what this series is about.
Each episode reconstructs one catastrophe from the primary record and stops at the exact point where the chain gave way. Not the moment of the accident; the decision that made the accident available. An incident earns an episode by the strength of its record, not the size of its headline. The video is the argument; this site is the evidence.
Reliability should be the default. It is not a sentimental position: reliability is a property you get by refusing to remove the places a mistake can be caught, and every episode here is a record of what it cost to learn that.
Method
Start from the primary record
Investigation reports, accident dockets, court findings, regulator orders. Not coverage of the coverage. Where a claim rests on a single source, the episode says so.
Reconstruct the chain, not the moment
The moment of failure is almost never the interesting part. The episode works backwards through the decisions that made it inevitable, and marks each point at which it could still have been stopped.
Name where it gave way
One sentence, stated plainly, at a specific timecode. Not a list of contributing factors, but the place the load path actually broke.
Publish the working
The full transcript and every source sit on the episode page, at a permanent URL, so the argument can be checked rather than taken on trust.
Standards
These are fixed for the whole series rather than decided per episode, and they are published here so they can be held against us.
- Blameless
- The subject is the system, never the individual. Rank-and-file employees are not named. Executives and public figures appear only where their public statements are themselves the story.
- The sourcing standard
- Every stated figure traces to a published postmortem, a regulator finding, a court record, or two independent reputable reports. Where a root cause was never officially confirmed, the episode says "reported as" and not "was".
- The fact boundary
- Each script carries a fact box, and it is a boundary rather than a summary: nothing outside it is stated as fact on camera, or in a caption, a title, or a description on any platform the episode is published to.
- Corrections, never silent
- A factual error, once verified against the record, gets a correction note on the episode page and a pinned correction wherever the episode is published. Nothing is quietly deleted.
A factual error can be reported to corrections@thepointoffailure.com.
Some episodes concern loss of life. Those are made and presented in a graver register throughout: slower, without promotional framing, and without speculation about the people involved.