[purplethinking]

Also, when it comes to cars, and probably other devices/vehicles in the future, they are increasingly operating themselves. You can buy FSD for Tesla and drive for hours in mixed highway and city streets without having to intervene. When you do intervene you can take control for 15 seconds and then give back control to the system. At that point, why put in buttons to optimize the experience for human drivers? This is true for other cars as well, but to a lesser extent, but the direction is clear.

[mmooss]

> At that point, why put in buttons to optimize the experience for human drivers?

Less optimization results in more accidents, injuries, and deaths.

[purplethinking]

Not if the car drives itself 99% of the time.

[consteval]

It doesn't though, and it never has, and I'm doubting it ever will. FSD is flawed from the conceptual stage. Roads are complex, constantly changing, and difficult to navigate. Other types of transportation, like rail or air, are trivial to automate in comparison. And then at a hardware level, Tesla also messed up. Pretty much any idiot could've told Tesla that a camera-only system won't work, but here we are.

FSD is a very small step above adaptive cruise control. It's more of a novelty than anything. I certainly wouldn't trust using it, and I don't really care what numbers say either. Tesla doesn't really play fair, being deceptive is a core part of their business. It's no surprise then that FSD auto shuts off right before accidents. We actually have no idea how safe it is, and I'm not going to be listening to what the guy selling them has to say.

[purplethinking]

Here is FSD navigating Manhattan: https://www.youtube.com/watch?v=sD3KSD620qk

"FSD is a very small step above adaptive cruise control"

Yeah right

[mmooss]

However much time the person spends driving, less optimization for their driving will result in more negative outcomes.