They should be designed with the intention of preventing dumb people from doing bad things accidentally, but there’s no way to make it 100% impossible for people to do dumb things intentionally.
I have the solution, the car should apply the turing test to the drivers seat before activating
2 Likes
I just did. The software failed to detect the non-presence of a person in a seat, and either this functionality wasn’t tested at all or the testing itself was a failure.
If you do that you’ll have made the car not roadworthy.
So literally one failure due to intentional misuse of any feature is a huge failure? No deaths, no crashes, etc noted due to someone doing this, simply a huge failure.
Is it a huge failure when a car thinks you have your seatbelt on when you plug it in behind yourself? Or is this logic only applicable when technology is involved?
Two days ago:
https://www.consumerreports.org/car-safety/electric-cars-prove-safe-in-iihs-crash-tests-and-insurance-claims/
1 Like
As an engineer, it is literally impossible to predict the near infinity of ways that an end user will use/misuse your product.
1 Like
Not to mention design to zero accidents.
There is going to be a huge mistrust of automation in driving. People like to be in control. Honestly though, say we went fully automated driving and there were 10,000 deaths a year from software glitches or other accidents. That’d still be an excellent result. People don’t think about this tech in that way though. It’ll save countless lives.
3 Likes
This the point we were driving to. People don’t trust what they don’t understand.
So, basically, jal is just like people who don’t want the vaccine.
1 Like
Elon massively overhyped his “fully self driving” cars to his cult followers, which naturally led to some of them getting into the backseat and letting Tesla’s glorified driver assist ram them into tress and trucks and so on. Not much you can really do outside of having Musk gently discourage people from killing themselves.
2 Likes
That’s the simple story but this car didn’t have autonomous driving and reportedly the autopilot was off. Just don’t think this is going to be that clean cut of a story
Playing devil’s advocate to myself - say there are 10,000 deaths to software bugs or misuse that are mostly random. My lizard brain tells me that my odds of dying are much lower because of my excellent driving ability, so I’d distrust the tech and be less inclined to use it if the fatality rate was that high.
Thinking pragmatically though, this is just silly. We will have to get to mostly zero deaths for it to ever be a thing, unfortunately. Will probably delay it by 10-15 years, and then regulation will delay it by another 10-15. The tech is good, but definitely not all there yet.
Jal’s conclusion of “take them off the roads” is flawed, but the concerns are valid.
There are going to be stupid people misusing these things and we need ways to minimize that.
Even if the risk is objectively reduced, the tolerance from the masses of a death from a self-driving car is going to be much less than a death from a traditional one.
There are tons of those things already built in. For example, the Tesla watches you on video and if they catch you not paying attention they revoke your ability for self driving.
1 Like
If he sincerely believes it, then it is a good faith objection. Someone can be wrong while arguing in good faith. Are you calling jal a liar?
I’m not. It’s basic math. As you already pointed out. Objections to self driving itself can honestly only be bad faith or ignorance. Discussing the nuances of how it should be rolled out or designed is worthwhile. Pretending self driving tech doesn’t work is not good faith.
I honestly wonder if this concern is overblown. Think of all the automated systems built into modern cars and there is zero push back. I agree full automation is different but I bet the price of full automated driving vehicles will be a bigger roadblock than the psychology.
Yes. Tesla has massively improved. It’s the leader now for sure.
1 Like