Tesla autopilot involved in far more crashes than previously reported

Speaking of crashes, I wonder if Tesla was on autoplilot when it invested in crypto…

6 Likes

Congrats on your due diligence, but I don’t think that somebody assuming that an extremely expensive product that is called “Autopilot” would be self-piloting is a damned fool, especially when the CEO does as much talk about autonomous automobiles as Musk does.

11 Likes

Announcing a special Tesla upgrade: a windscreen that goes completely opaque when it senses an imminent crash. This spares the occupants the alarm and stress of contemplating their death or dismemberment, something which that autopilot also seems unable to deal with.

16 Likes

I persist in thinking that a person who ignores the evidence of their senses — it takes ten minutes behind the wheel to work out that “Autopilot” while a good driver-assist system is a piss-poor autonomous driving system, twenty minutes if you’re a slow learner — because of branding and the rantings of a known nutter, is indeed, a damned fool.

Maybe you have a different definition of “damned fool”. That’s between you and your insurance company.

ETA: besides, you’d think someone with the handle “Aeroplane” would know an autopilot isn’t and never has been an autonomous flying system, but requires pilot attention.

5 Likes

Is that a solder sucker?

2 Likes

That’s a great guess; and I do love playing with solder suckers; the “phut” is so satisfying.

The item in the photo is a restrained bolt “stunner”; a manual, lower strength version of the heavy weight, compressed air device used by Javier Bardem in “No Country for Old Men”.

7 Likes

Brilliant. Wish I could give you ten likes!

4 Likes

Every day I meet people at my job who have crashed their cars while using their cellphone. Not one of them, EVER, sees the causation. Today, driving to the lake, i saw a guy swerve off the highway, recover, and continue talking on his handheld cellphone.

I think jgs might be the only guy who realizes.

And I think cars should be harder to drive, not easier. And driving tests should be harder. But that’s just the trauma nurse in me hoping to save lives.

11 Likes

:cry:

Prior to seeing @euansmith’s post, I would have said you can’t fix stupid.

I would agree, if I thought hard-to-drive cars would result in less mayhem. But data seems to indicate that as cars have gotten easier to drive, fatalities per mile driven have generally fallen.[*] Correlation, causation, blah blah, but still.

Do you think there’s some reason [**] to believe that hard-to-drive cars would make the roads safer?

Yeah, but then we end up talking about how we’ve built a world where if you don’t drive, you don’t work, and if you don’t work, your family goes hungry and homeless. And, before we start dialing up the driving tests, maybe we should actually keep people with DUI convictions off the roads? Or uninsured motorists? /Grumble

But if I had three wishes I wouldn’t mind spending one of them on making everyone a good and considerate driver.

[*] Not for the last two years, but the overall trend is still distinctly downward.
[**] Ideally supported by data.

4 Likes

Hence EM’s GOP-leanings. Properly greased, they’d help him get through (or around) repercussions whether cars or rockets.

3 Likes

Cars are also built to perform better in accidents (air bags, crush zones) and more and more people wear seat belts, so I’m not sure easier to drive is the variable that can explain fewer auto fatalities. Someone who is driving a manual transmission is probably more engaged in their actual driving experience.

7 Likes

Sure! But by the same token, “harder to drive” lacks even suggestive correlation, AFAIK. There are enough variables that anyone short of a statistician working in the field, who claims they know for sure, is fooling themselves.[*] Another factor to consider: cars accelerate faster than they used to (much!) and have higher top speeds.

Speaking as a long-time stick driver (I went direct from decades on stick to EV, I’ve never owned an automatic transmission) I’m not convinced that manual transmission corresponds to attention to the road. Like anything else, muscle memory takes over once you have enough experience – and when you’re a new driver, lack of experience with shifting means you have one additional thing to divert your attention from what’s in front of you. So, “engaged with their actual driving experience”? Yeah maybe, depending on what you mean by “driving experience”. But a safer driver where it counts? Data please, because it doesn’t jibe with my… actual driving experience.

[*] And I can’t say about the hypothetical statistician either, but maybe.

3 Likes

The one gauge that they really should have on the dash:

confidence-level-meter-confidence-level-scale-arrow-measuring-device-icon-sign-tachometer-speedometer-indicators-177883790

In other words, how sure the program is that its data describes the actual road situation. (Checking to see if its confidence level is accurate is a problem for another day.)

“Wow, that intersection back there, I just had no clue! I mean, there were lots of confusing things going on, and I wasn’t really sure what the traffic lights were doing, or where some of the cars were going. Took my best guess and kept going.”

I’m sure many drivers would like to know when the “autopilot” is doing computerized white-knuckle driving, at night in the fog with faded lane paint.

11 Likes

PS most people on the road are fools. Auto design safety must take that into account.

5 Likes

“What do you mean, I’m not helping!?!”

3 Likes

Yep - I have autopilot (got it for ‘free’ on a used model) and basically for me it’s the same thing as Volvo’s much more conservatively named ‘Driver Assist’ feature. The naming and branding are pretty terrible.

Where I use it is stuff like ‘I’m stuck in stop and go traffic, and I want the car to handle the constant stop and go part’ or ‘I’m on a long ass highway and I’m going to change the radio / etc’. It’s an assistant but I would never close my eyes or just take a nap or something like half of the folks duped into thinking it’s magic are doing.

4 Likes

Must it? When do you think that will start? Because in the history of the automobile, I’m not aware of one ever having been idiot-proof. Not even close.

I mean, an idiot-proof car would be nice to have, but I suspect it’s kind of like a truly secure operating system. (You know the old saw, you can truly secure a computer if you remove power, cast it in concrete, and sink in in the Mariana Trench. Otherwise, the best you can get is secure-ish. So it goes with making safe multi-ton chunks of metal moving at high speeds, Larry Niven’s Safe at Any Speed notwithstanding.)

1 Like

Considering that for the most part we can’t explain how sample-trained neural networks even come to any specific decision this would definitely be a challenge. I do like the idea of there being a smaller “Confidence Confidence” meter right beside the “Confidence Meter”. And so on, recursively, until it’s just meters all the way down

5 Likes

Error bars

12 Likes

We just take the number of non-crash minutes, divide it by the number of crash minutes, and display the result on an analog meter without a scale. Marketing tells me that it’s an “advanced Bayesian inference engine”…

(edit: I definitely left a divide-by-zero error in my laughable oversimplification for comic effect; not because I’m an incompetent hack. Definitely. We, um, pad the number of crash minutes with a positive nonzero ‘scalar safety offset’, for safety, and then divide the number of non-crash minutes by the enhanced safety number of crash minutes. Yes, just like I said the first time.)

3 Likes