Another week, another federal inquiry into Tesla's driver-assist software.
The National Highway Traffic Safety Administration has opened yet another investigation, this time into whether Autopilot failed to detect and respond to parked emergency vehicles.
It is at least the fourth such probe in recent years, and the pattern has become grimly familiar: a crash, a headline, a promise that the software will improve.
Meanwhile, Tesla keeps shipping the feature to more drivers and calling it something close to autonomy.
The company's own manuals say otherwise, buried in fine print that most buyers never read.
Drivers are told to keep their hands on the wheel and their eyes on the road, even as the marketing whispers that the car can handle it.
That gap between the brochure and the disclaimer is where people get hurt.
We have quietly accepted a bargain that no other industry could get away with.
Imagine a dishwasher that floods your kitchen once a year, but the box says it is "full self-cleaning." Imagine a lawnmower that occasionally lunges at the neighbor's dog and the company responds by saying the owner should have been watching.
We call the software a "beta," and suddenly the rules change.
The deeper problem is not the technology itself.
It is the culture we have built around it, this idea that convenience is worth almost any tradeoff, that we can roll out half-finished tools and let the public serve as the test track.
Every American who shares a road with a Tesla is now part of that experiment, whether they signed up or not.
Automakers are rewarded for announcing features, not for withholding them until they are safe.
The people in the middle just drive to work.
There is a version of this future that works.
Driver-assist systems that actually assist, that intervene in emergencies without pretending to replace human judgment, could save thousands of lives a year.
That version requires honesty, though—about what the software can do, about what it cannot, and about who is responsible when it fails.
Right now, we have the opposite: a feature that is sold as magic and defended as mere assistance, whichever excuse fits the moment.
That is marketing with a body count, and it is happening in broad daylight while we scroll past the headlines.
The real question is not whether Tesla can fix its software.
It is whether we, as a society, will keep letting companies define safety down until it means whatever survives the lawsuit.
Our take: innovation should not require a body count to prove itself.
Final Thoughts
Until these systems are honest about their limits and held to the same standards as every other product on the road, "Autopilot" will remain one of the most dangerous misnomers in American life.