Having done my time as an Army medic, this is incorrect. It takes more force than that, but less than you might think. A good 25 kilos with some velocity behind it will easily sever a phalange. Up it to 50 or 80 kilos and you can claim an arm or shin. Mass is the real killer. I’ve seen a vehicle at comically slow speed absolutely yeet someone because it had several tons of momentum behind it.
Casual readers might remember a recent very low-speed collision that nonetheless caused a catastrophic failure due to the tens of thousands of tons of weight. The MV Dali vs. the Francis Scott Key Bridge, if you didn’t guess. It struck the bridge at about 8 mph.
He did demonstrate it that way, specifically with a carrot. And it somewhat worked. The problem is they programmed it to do more and more pressure every time it fails meaning that doing the carrot first actually caused a safety issue. He only moved onto his finger because the safety feature seemed to be working.
The engineer told him the frunk increases in pressure every single time it closes and detects resistance, Judkins said. It’s going to assume you want to close the frunk and maybe something like a bag is getting in the way, which would make it close harder.
But let’s face it: if the car lid would never close if something is in the way, some other dumb youtuber would have made a video about it and here there would be a discussion about how stupid are the engineers to not let the lid close even if a bag in slightly on on the way and the user know what they are doing.
You’re missing the point of a safety feature. The car shouldn’t, by itself, close the lid if something’s in the way. It should allow the user to push it down, or disable it temporarily, to do so.
The point of a safety feature in any system is to prevent unexpected situation from having unexpected consequences, not to be a magic solution that accommodate for brainless people. In one direction, you can make the judgement call and force the thing down, in the other direction you lose a finger.
You’re missing the point of a safety feature. The car shouldn’t, by itself, close the lid if something’s in the way. It should allow the user to push it down, or disable it temporarily, to do so.
I get the safety feature. The point is that here I am saying to the car to close the lid even if something is in the way. I made a conscious decision to do so, and more than one time, so I expect the car to do it. But I agree that it could have been designed in a better way.
The point of a safety feature in any system is to prevent unexpected situation from having unexpected consequences, not to be a magic solution that accommodate for brainless people. In one direction, you can make the judgement call and force the thing down, in the other direction you lose a finger.
Which is exactly what happened here. He made the judgement call to ignore the safety feature (and probably ignored how the feature works)
He did the test wrong because he's experimenting with "safety" algorithms that the manufacturer has provided little-to-no documentation on and is having to come up with answers on his own. Maybe he wouldn't be "doing it wrong" if Tesla hadn't over-engineered every aspect of their piece of shit truck in the first place. This thing is a solution in search of a problem, and it'll chop your fingers off until it finds it.
That was very nearly my exact same thought. Maybe not for curious children with carrot-sized fingers, but for adults, how convenient! Business competitor's body won't quite fit in your fancy frunk? Just while away on your phone for about 10 minutes, let the cat do its magic, and off go the legs! Travel-sized!
“Cave Johnson here. Fact: the key to any successful cooperative test is trust. And as our data clearly shows, humans, cannot be trusted. The solution? Robots!”
"An update: It turns out the robots learned how to lie and can no longer be trusted. They tried to take over the testing mainframe and I’ll be damned if I let them get their hands on my equipment! So I sent in a couple of interns to take care of it. Go earn those recommendations! I told them.
“Anyway, it’s back to humans. This time new and upgraded with our state-of-the-art Aperture brain chips. The boys out-did themselves this time, increases testing compliance by 150%! Let’s see those humans just try to lie now!”
Judkins said that after the finger test, a lead cybertruck engineer at Tesla said he did the video wrong.
The engineer told him the frunk increases in pressure every single time it closes and detects resistance, Judkins said. It’s going to assume you want to close the frunk and maybe something like a bag is getting in the way, which would make it close harder.
Are you kidding me? You did the test wrong on a safety critical feature? No you dumbass engineer, you designed it wrong. Why in the holy fuck would you make a safety critical algorithm keep applying more pressure on subsequent attempts??? That’s literally the opposite of what you do for safety.
This is why, as a software developer, I’m against designing any system that assumes what the user wants and tries to do it for them automatically. On the occasions where the assumption is right, it’s a mild convenience at best. When it’s wrong, it is always infuriating if not dangerous.
“Smart” may as well be synonymous with “unpredictable”. I don’t need my computer to be smart. I need it to be predictable, consistent, and undemanding.
Yeah, I’m an embedded software developer myself and yeah, when we architect our code we have safety critical sections identified with software safety reviews and we always go with the assumption that we’re going to run into that one guy who’s the living embodiment of Murphy’s law and go from there with that design to minimize the potential for injury and death.
Can’t imagine who the hell is in charge of the software safety reviews there that let that pass.
Same in the medical devices industry. We have whole teams of non-developers whose job is to find out when and why a surgeon can be a moron. The code is more difficult to write, but it’s way better and more robust.
And also every additional kind of complexity (which stacks BTW) makes you more dependent on the vendor (good for them, bad for you) and on doing things exactly as their imagined user (because it’s disproportionately your problem as laws don’t seem to work in making it theirs).
Distributism is actually a very good political ideology. Sad it’s associated with Catholic religion, because it correctly generalized the principles making democracies and markets and cultures work.
I wonder if the guy that designed autopilot had the same idea. “So when the car detects resistance up ahead in the form of a crowd or wall, it will accelerate to make sure it goes through!”
The engineer told him the frunk increases in pressure every single time it closes and detects resistance, Judkins said. It’s going to assume you want to close the frunk and maybe something like a bag is getting in the way, which would make it close harder.
What the fuck kind of idiots are leading things over there? “Something’s in the way. Better crush it!” What a bunch of morons putting everyone in danger.
I know I’m old school and all that, but why do people want to pay for automatically closing doors of any kind? Automatic opening of cargo spaces I get, if you have your bags full of hands or whatever, but once you put the stuff in there… Seem like such an incredibly unnecessary and costly feature, that also have a high chance of failing in the future. I don’t get it.
Good question. My wife’s RAV4 has a rear door that will only close if you press a button. You can’t close it manually. Furthermore, it’s on the door while it’s open and my five foot tall wife can barely reach it. It’s ridiculous.
You know, that’s true and it didn’t even occur to me. I guess she just wouldn’t have bought it? (I would have been fine with that, I hate SUVs, even hybrids.)
We’ve got a 2019 Rav and I can’t remember how, but you can adjust the height that the door opens to by some series of button pushes. We had to lower it so that it doesn’t hit the frame of the garage door when opening it inside the garage. Maybe just adjust it so that it doesn’t open all the way and it’ll be easier for her to reach the button?
I actually sell these. You can manually lower the door to the height that works comfortably, then hold the automatic door button down for about 3 seconds. That should program the door to a new maximum height.
How do I set the height on my vehicle’s adjustable power liftgate?
When the liftgate reaches the desired height, push the rear liftgate close-button once (button is located on the doorjamb of the rear liftgate, and only accessible when the liftgate is open). Press and hold the button until it beeps 4 times. Click here to view a video.
My Subaru has a similar setup, and there’s a feature for changing the max height of the tailgate. You might wanna see if the same thing exists for you.
Because like you said, it’s a nice to have feature. I like my wife’s auto closing hatch for when I have a handful of boxes for that final grocery run and just walk away and it closes. It’s literally just really nice convenience feature and if it fails, you go back to closing it manually.
I get it’s nice to have, and if it somehow cost nothing I wouldn’t mind having it in a car, if it’s pretty much guaranteed that when it fails it doesn’t prevent me from open/close manually. But I’d much rather not pay for neither the R&D, engineering, parts and manufacturing of it, only to end up with a more complex door mechanism that is more expensive to repair and more likely to break. When all it does is give me the slightest of conveniences. Best example of this is the motorized charging port lid on the Rivian. Like, whyyyy? Cheaper and longer lasting vehicles, please.
Why the hell would it close harder if there is something in the way? That’s not the correct behavior for a lid, that’s the correct behavior for powered shears.
Never tried to force the closing of your trunk lid because there is a bag that is slightly over the limit and you need a little more pressure, even if the bag is a little pressed down ?
The assumption here is that if it is your finger which is in the way, you take it out the way and you are not that stupid to try to close it again if for some reason you are not able move it away, which to me seems to make a lot of sense.
Safety critical? I’d rather have a trunk I can get to close than one I can stick my finger into four times in a row without pinching it. What do you think happens when you slam down a normal trunk on someone’s finger?
Lol. Nah, the trucks are super dumb. I just know I’d want a trunk that would be able to close more than an overly sensitive pressure detection permanently preventing it. For that matter, I think it’s dumb to attach a motor to a trunk.
No thank you. I refuse to engage with a person trying to straw man and change topics from a software safety argument to a personal preference that goes nowhere but you feel free to engage if you wish.
Maybe you didn’t comprehend it? The close force attempt increases with each unsuccessful attempt at closing. That way seems better than it eventually not working at all a few years down the line as all the electronics get more jankety be cause something gets a bit bent or worn out and it always detects a small amount of resistance so it quits closing all together.
Nobody wants to discuss the logic involved with having to open the door and then close it again for it to attempt to close harder and why that isn’t the dire safety hazard that people are trying to make it out to be. These people are the reason why we have to have “no smoking” signs at gas pumps because apparently they’d leave their hand in the door after attempting to close it 3 or 4 times.
It strikes me as exactly the kind of engineering call that Elon has tended to make, time after time. With zero training in an area, he gets a solution in his head crufted up from some set of pre-existing notions or points of view and then pushes to have them implemented. He will also go on to fire anyone who disagrees with him. I spoke with an engineer who worked on the gull wing doors, which the team had objected to, and not only did he force them through, he burst in on one of the finalization meetings where they had finally reached a design consensus and insisted they change the hinge. Given similar reports on his behavior regarding other products (including especially twitter), I have no reason to disbelieve this person.
It could be a lot better if it were able to get through tough terrains like wet beach sand. Or if the body didn’t rust after touching moisture. Or if it was able to survive a car wash.
Also it would have been neat if they had some automotive professionals working there to tell them that the accelerator pedal needs to come back up when your foot is off it.
I saw my first cybertruck in person the other day. It looks incredibly dumb in promotional photos, but it’s astonishing how much stupider it looks in traffic surrounded by normal vehicles.
Well traditionally, cars only had one trunk, now it is common that they have two. The need arose to distinguish them, and ‘front trunk’ easily collapses into a nice single syllable portmanteau that makes communication simple and concise; the language evolves and a new word is born.
If you read the article, it’s not a statement with entirely no merit.
The engineers prioritized an algorithm which is far more likely to be useful in real world scenarios where you keep trying to cram a bunch of stuff in the frunk and close it (who hasn’t done this?) rather than the edge case of repeatedly testing it with vegetables until you stick your finger in it.
This is why you keep your safety features consistent. If they want bag close mode, then make it where you hold instead of press a button or something. It “happening automatically” is just unpredictable to most, not magical
In addition, what if the person noticed the instructions and then moved them away, and then accidentally got a finger in there? That’s a realistic scenario too.
There should be no algorithm. It should be done by a human. There are no amount of lines of code I will ever make up for knowing intent and what the current situation is.
If it’s going to be closed by software it needs to prioritize safety 100% of the time. If more pressure is needed and that pressure needs to come from a human.
Youre constantly forcing your trunk closed? That doesn’t sound normal to me actually, and sounds like the opposite of what I would want. Hello, groceries, important things, stuff I don’t want stolen so goes in the trunk?