Another week, another federal investigation into Tesla's driver-assistance software.
The National Highway Traffic Safety Administration has now opened inquiries into hundreds of crashes involving Autopilot and what Tesla markets as "Full Self-Driving," including a string of incidents where cars struck parked emergency vehicles or ran through crossings they should have seen.
Each new filing reads less like a tech story and more like a slow-motion ethics case study playing out on American highways.
Here is what makes this different from the usual gadget controversy.
When a phone drops calls or a laptop overheats, you return it.
When a vehicle's software misreads a stopped fire truck, someone dies.
Tesla's own name for the feature does the damage before a single line of code runs. "Autopilot" and "Full Self-Driving" tell drivers something the system cannot deliver, and human nature fills in the rest.
People trust the label, glance at their phones, and the margin for error vanishes.
The marketing problem and the safety problem are the same problem.
Tesla has spent years selling a future capability as a present one, collecting thousands of dollars per car for software that remains, by the company's own fine print, a driver-assistance tool requiring full attention.
State attorneys general have started asking whether the pitch itself crossed a line.
What unsettles me is how ordinary the danger has become.
Teslas are in suburban driveways, school pickup lines, and commuter traffic across the country.
Millions of Americans ride in these cars daily, many of them never reading the manual, many of them genuinely believing the wheel can handle it.
A society that outsources attention to software it does not understand is taking a gamble it has not priced.
Automakers everywhere are watching, because this is not really a Tesla story.
It is the first real test of what happens when Silicon Valley's "move fast" culture meets two tons of steel moving at 70 miles per hour.
Every other manufacturer now ships lane-keeping and adaptive cruise features with similar names and similar overselling.
The industry standard is being written in these investigations, and so far the standard is: promise everything, disclaim everything, and let the courts sort out the wreckage.
There is a defensible version of this technology.
Driver monitoring that actually works, naming that tells the truth, and marketing that respects the driver's life would go a long way.
It requires restraint, which is the one thing the sector has never been good at selling.
We have spent a decade teaching Americans that convenience is worth almost any tradeoff, that the machine knows better, that vigilance is someone else's job.
Autopilot is just that belief with a steering wheel attached.
The recalls and investigations will eventually produce fixes, but they will not undo the lesson we taught ourselves: that we would rather trust a screen than pay attention.
Tesla can build remarkable cars and still be wrong about how it sells them, and both things can be true at once.
Until the names match reality and the monitoring matches the names, every driver is part of an experiment they never agreed to join.
Final Thoughts
That is a bet placed with someone else's life.