Tesla Is a Software Company Trapped in a Car
The update channel is the thesis. The factory is the constraint.
[ essay ]
Tesla’s car is a shipping container for software. Over-the-air updates, a screen as the dashboard, Autopilot and FSD as the argument people actually have: those are software motions. Body shops and factories are the trap. You cannot iterate a fender the way you iterate a checkpoint.
I do not work at Tesla. I read the public company the way I read any vendor that ships a computer with wheels: 10-K language about autonomy and energy, plus the fact that the hardware has to be legal on a road. The software thesis is real. The manufacturing thesis is also real. It is slower, and worse at pretending a bad quarter is a failed A/B test.
A company that thinks in deploys still has to stamp metal, pass crash tests, and service a fleet that rusts. That constraint is the trapped part. The opportunity is the same constraint. Once the computer is in the driveway, you can tax it with features without a dealer visit. Other OEMs noticed. The copy is now weather: every legacy brand has a software slide. Few treat the vehicle as a node that phones home for the next build.
If you read Tesla as only Detroit, you miss the update channel. If you read it as only a model lab, you miss the factory. Hold both. The company is software. The product still has to not fail at a wet junction.
— JV · Dark Heart Labs.