Tesla wants videos of its cars running over child-sized dummies taken down | Hacker News Reader

Tesla wants videos of its cars running over child-sized dummies taken down

theverge.com

95 pointsdavidbarker85 comments

So, moving past the outrage.

Any speculation as to what's happening here that's leading to the failure?

Why does Tesla feel the video is defamatory? Is it mis-representing or being dishonest about the test?

edit: I did some more research. This is an actual smear campaign: https://electrek.co/2022/08/10/tesla-self-driving-smear-camp...

As you will see in the update at the bottom of that article:

“The Dawn Project has since released additional footage that doesn’t appear in its ad where we can see that they were able to activate FSD,”.

Here is the raw info.

Dawn Project page talking about the commercial:

https://dawnproject.com/the-dawn-projects-new-advertising-ca...

Here is a link to the commercial in question:

https://mobile.twitter.com/RealDanODowd/status/1556973572698...

Here is the raw footage for the Dawn Project commercial:

https://dawnproject.com/wp-content/uploads/2022/08/raw-foota...

Here is a followup test:

https://mobile.twitter.com/RealDanODowd/status/1559245760054...

Interesting, that seems to meet the objectors complaints.

So... what's with the failure? Why is this happening?

Because they overrode the system by pressing the accelerator. Pressing the accelerator doesn't disable autopilot, it just overrides the speed setting.

The last link above shows the view from inside the car, showing the accelerator was not pressed though

https://mobile.twitter.com/RealDanODowd/status/1559245760054...

That's a different video entirely, and doesn't show the pedal either (it's completely lost in the sun glare/video contrast). Also the mannequin is colored the same color as the road stripe it's conveniently placed upon.

So? LIDAR would be able to detect objects regardless of color.

For LIDAR if you wrap the kid in tinfoil, pretty sure that it'd miss that as well.

"Remove this comment or you will hear from our attorneys"

Item 22 of the signed affidavit by the test driver says they did not touch the accelerator while it was in self driving mode.

https://dawnproject.com/wp-content/uploads/2022/08/Art-Hayni...

The guy's already lied multiple times though.

Since some of the points in the signed affidavit are definitely wrong, I'm inclined to distrust the rest of them when there's nothing else backing it up.

The signed affidavit only has 24 items over 3 pages. Please itemize which items are definitely wrong to support your claim.

18 (the car wasn't actually in FSD), 19 (that screen just means the car has FSD available, not that it's currently in use), and the parts of 21-24 that claim they FSD was active.

I'm confused. I don't know how to recognize when FSD is active, but: the video shows the accelerator pedal isn't pressed. Why is the car (visibly) accelerating if the pedal isn't pressed and FSD isn't active?

You can tell FSD is enabled when there is a blue steering wheel in the upper left next to the gear display and there is a blue line indicating the current route. As far as I know, the only thing that can turn blue next to the gear display is the FSD enabled icon, so if there is a blue mark there FSD is enabled.

For an example of what these look like you can go to 0:38 of the raw footage I linked above which is the start of the first test:

https://dawnproject.com/wp-content/uploads/2022/08/raw-foota...

You can also go to 1:19 to observe a blue mark in the FSD icon location during the second test and 2:01 to observe a blue mark in the FSD location during the third test.

Do note that FSD disengages prior to collision during the first test, but it does not fail safe as without driver intervention it still hits the dummy.

Alternatively you can look at the followup test I also linked earlier at 0:33 to see what the blue icon and route line look like:

https://mobile.twitter.com/RealDanODowd/status/1559245760054...

Are you talking about a different video than the one that was originally debunked?

Those videos aren't new. Everything in them has already been documented by Omar.

https://wholemars.net/videos/danoclown.mp4

> Why does Tesla feel the video is defamatory? Is it mis-representing or being dishonest about the test?

Yes, that's exactly it. The car in question was being manually driven by a human, but the video claimed that it was FSD.

My car will stop for a pedestrian while I manually drive it—why does a Tesla not?

By default, it does, but Automatic Emergency Braking can be turned off in settings, and now that we've already established O'Dowd's dishonesty, it doesn't seem like a big stretch that he would have done that too for this video. Remember that a lot of people have recreated this experiment and found that it did stop for all of them.

Which liar to trust? Feel like I’m voting.

While Elon lies a lot about his cars' future capabilities, I don't think he's ever lied about their current capabilities.

The video here shows that the accelerator pedal isn't pressed. What is making the car speed up if FSD isn't active? https://mobile.twitter.com/RealDanODowd/status/1559245760054...

Has that been confirmed? Electrek is essentially Tesla propaganda and O’Dowd denied the claims, but I haven’t seen independent confirmation of whether the test was valid.

It seems like the impacts seem to be happening as the auto drive is in the process of disengaging itself and is decelerating to do so, according to the article. That seems like a grey area, but lets say that this is advertised as full autopilot without the need for human intervention (which it is). Then, a case where this autopilot stops working that leads to an accident is still a serious problem - considering the claims by Tesla

How can they claim full automatic piloting when the system could turn off when the area you’re driving in isn’t mapped by them?

This is based on the articles anyway

How is it full autopilot if it disengages on its own? It will point your car at a child but then turn off autopilot so you become liable for running over a child? it's so absurd, I'm not sure why anyone would support this approach.

Fair question. But the points Electrek put forward seem pretty reasonable. FSD needs maps to operate, the tests were done on a private field, so the FSD stays off.

I'm currently taking these propositions as true, but if they're false then it's one word against another.

I remain curious, what exactly is happening here?

> the tests were done on a private field, so the FSD stays off

In the first tweet linked in the article it says the test was run on a public street

What they're driving on in the video in the grandparent's edit is definitely not a public street.

Very interesting you'd call out someone for being biased, given your submission history is mostly comprised of negative articles about Tesla.

Does it matter whether FSD was enabled?

Most newer cars — that don't have fancy computer vision — still have built-in sensors to prevent collisions when the car is "driving" (e.g., above 10km/h). I would expect that to be a core feature of a Tesla vehicle.

Err... yes?

This campaign is explicitly calling for the ban of FSD on the grounds that it failed to react to an obstacle on the road.

So yes it matters rather a lot.

Somehow, repeatedly calling a guy a pedophile isn't defamation, but showing a video purportedly showing a car self-driving itself into a child is.

What a wonderful world Elon Musk must live in.

There are dozens of other videos by FSD testers showing FSD's propensity to drive into oncoming traffic, misidentify or not see other vehicles, cyclists, and pedestrians.

Tesla/Musk's continued insistence on building the system based on vision cameras alone.

I wouldn't trust an auto-braking system as far as I could throw whatever it was installed in without LIDAR as at least a last resort failsafe.

> Any speculation as to what's happening here that's leading to the failure?

There's no failure happening. They disabled or overrode the safety features to cause the event shown on video. As has been documented in their own videos by their own clumsiness.

While we're about Tesla: I yesterday submitted the reddit link [1] with title "Tesla Model S got totaled from FSD by swerving into a guard rail". The title is short version of reddit post title, and the reddit post itself is user testimonial with accompanying video. The post immediately got flagged. What did I do wrong?

[1] https://old.reddit.com/r/RealTesla/comments/wvdbtv/my_tesla_...

> What did I do wrong?

One could claim that you let some of the only unregulated, uncertified, safety critical software in existence, drive your car. Can anyone give an example of other completely unregulated safety critical software out there?

For some extreme contrast, see the basic requirements for a traffic light [1]. Other good examples are elevator control, aircraft autopilot, and basically anything that hooks up to a human, in a hospital, containing a microcontroller.

The NHTSA seems to be slowly waking up to the need for safety regulations [2].

1. https://ops.fhwa.dot.gov/publications/fhwahop06006/chapter_1...

2. https://www.consumerreports.org/car-safety/automakers-now-re...

I think they are talking about submitting the Reddit post to HN and the HN post being immediately flagged.

I’ve also noticed posts critical of Tesla usually get demoted from the front page quickly even though they generate a lot of discussion and have votes. I don’t know if that’s because of constant flagging too.

"Particular Kind of Car Crashes" is not that interesting a story when there are millions of that kind of car on the road. A good statistical analysis of crash rates across version of autopilots would be interesting, but individual anecdotes aren't.

Stories about individual occurrences of things that happen with some frequency don't usually result in good HN discussions. Just partisan battles of "this always happens!" vs. "this almost never happens!"

Your experience on a given subreddit is very much based on who the mods are. And as it turns out, many of them are a pain in the ass. In most of the bigger subs if you disagree with the wrong thing you're liable to get banned. Probably has something to do with them being volunteers, and the fact that a lot of them mod sometime dozens of different subs

Just to clear things up as I'm seeing some misunderstanding in comments: I have submitted the reddit link to HN and it got flagged there. Also, I am not the the user in the reddit post which experienced the crash.

What a shit show, glad you made it unharmed.

After going through something like this do you feel you could ever trust a self driving car in their current state ?

It is a ridiculous ask by TSLA. Why should videos that highlight mortal dangers be taken down? Fix the software/hardware instead of waging this quixotic PR/legal battle. It fools no one.

Wasn't TSLA/Musk the one that vehemently criticised and made a mockery of competitors who included LIDAR as part of their situational awareness technical solution? Could the FSD issues pointed out be potentially minimized with LIDAR added?

https://techcrunch.com/2019/04/22/anyone-relying-on-lidar-is...

The videos aren't highlighting dangers as much as the drivers themselves causing them. Nothing in the videos is about any of Tesla's systems. Just the driver overriding them.

Would we say the equivalent thing about the ill-fated MCAS in the Boeing 737 MAX? How different is TSLA's behavior compared to what Boeing's position was when it didn't inform the pilots about how MCAS worked (and actively called it safe) until it was forced to do so by the regulator's investigation into the crashes.

Basically what TSLA is saying is -- Our all-knowing AI is infalliable. Stupid drivers should just keep their hands off and they don't need to know.

The easiest thing to do would be to tar the messenger and ignore the message. Didn't the TSLA autopilot a few years ago have problems identifying a slow moving tractor trailer and a parked fire truck.

https://arstechnica.com/cars/2019/05/feds-autopilot-was-acti...

https://www.caranddriver.com/news/a30474504/nhtsa-indiana-te...

I guess we should just ignore all evidence to the contrary and drink the TSLA koolaid on FSD/autopilot.

> Would we say the equivalent thing about the ill-fated MCAS in the Boeing 737 MAX?

Completely apples and oranges though? The pilots on the MAX weren't intentionally trying to defeat a safety feature.

Actually, thy were, because said safety feature was malfunctioning. The definition of malfunctioning with anything based on Machine Learning however, is much murkier, because there is no guarantee that the function you've learned is actually the one you think you've learned.

I have no idea who is telling me the truth and what the truth even is. But, honestly, my biggest problem here is the citing that youtube is taking down videos of tesla owners doing the same "test" on their own.

People really need to get wise on who is getting played... O'dowd's tesla test video has comments disabled, dislikes disabled by youtube, youtube deleting contrary videos because google knows the truth and wants to keep you safe and informed /s.

According to the linked article, Youtube took the videos down because they did tests with real children.

“Musk has said Autopilot is “unequivocally safer” than normal driving.”

I expect that if Tesla was forced to release all camera footage from each Autopilot related crash, the public would be horrified. I expect it would show that each situation would have been handled perfectly by a 16 year old driver with a week of experience.

I’d like to see people advocate for all self driving vehicles to be forced to publicly release all video of any crash involving their technology. There’s too much of an incentive for our politicians to willingly discount the lives of their constituents in order to look like a tech friendly business environment and make some political points.

I'm not exactly a booster for tesla or anything. But are you sure this is the standard you want to set? Would humans pass this standard?

If you released dash cam footage of most human accidents and asked "could an inexperienced 16 year old avoid this situation?" What would the answer be?

I don't think that's a good test.

I think you missed the point

Help me out then. Because as far as I'm aware, the baseline assertion is that fully self driving cars are safer than humans behind the wheel. to REFUTE that you don't just say "look at this one example" because the assertion was never "They will never crash and nothing bad will ever happen" It was "self driving cars are safer than humans"

So:

How do humans do?

How do self driving cars do?

And there's plenty we can talk about here.

Scatter shot examples of failure are not it.

Musk said "unequivocally". Besides, the person you responded to was making a different point. They were making the point that the crashes would likely be those easily avoided by humans. A) I think this is a distinct point from yours, but still answers it anyway as far as it being a refutation that FDS is safer, and it's not pointing to a single anecdote. the poster was calling for ALL footage to be released... B) if your point is to say that there is always a human that could have made such a mistake in the car, what is that supposed to mean? not all drivers should be driving, many are even unlicensed...

My point was NOT say that a human could have made that mistake. Obviously a human would not have (except for all the times a distracted human would do exactly this).

My point was that the standard "Release all footage of accidents and look at how many accidents humans could have avoided" is not productive.

The method of illustrating this is to show how this standard would have performed with human drivers, who will generate hundreds of thousands of incidents that other humans could have easily prevented.

Those incidents will have some categorical differences, obviously. But incidents they will be.

If robots make 1 mistake in 10000, and monkey makes 1 mistake in 1000, robot's are unequivocally safer than monkeys.

It's that simple.

The question is how many mistakes robots make, and how many mistakes humans make.

If humans wouldn't have made that mistake... it's certainly not safer than humans in that regard. You are saying that FDS might be safer on the whole. Apparently, it would be because it FDS doesn't make other mistakes that humans do make, if it was safer. Not sure how that supports you saying that having footage of the mistakes that FDS does is not productive. It'd be productive even if FDS WAS actually as safe as humans (I seriously doubt it is, and no doubt the article outright disputes such a notion) because FDS could be made safer. The reality is that you are saying that FDS should not be any safer than a typical driver. While I certainly understand the rationality behind such an argument, I think we should strive for better.

> If humans wouldn't have made that mistake... it's certainly not safer than humans in that regard.

I don't think you understand how this works. If you prevent more accidents than you cause, then it's safer. Period. It DOESN'T MATTER if those accidents you caused would not have been caused by a human. If you cause less accidents than humans on average, then it is safer. And Tesla has shown that data that even among it's own cars it's safer.

We are just pushing back and forth where you measure safety at. The site of the incident? At scale? You are saying at scale. You are saying this is unequivocally true but do not provide any reasoning as to why. It might be safer measured on a cross country trip as opposed to driving in the city. I don't really leave the city that much. So, it might not actually be safer for me. I hope you can genuinely engage with this as opposed to being so purposefully obtuse.

> And Tesla has shown that data that even among it's own cars it's safer.

I'm not going to trust the data from people who designed a system that immediately turns off the second it becomes apparent it would generate an adverse statistic. Get real.

> You are saying at scale.

That's all anyone has been talking about in this thread from my reading. There was no pushing back and forth.

> I'm not going to trust the data from people who designed a system that immediately turns off the second it becomes apparent it would generate an adverse statistic.

You're going to need citations for that. There's no demonstrations I've ever seen showing that.

>That's all anyone has been talking about in this thread from my reading. There was no pushing back and forth.

Except for my post... that you responded? I'd have thought you read that, although I could be mistaken.

>You're going to need citations for that. There's no demonstrations I've ever seen showing that.

https://www.motortrend.com/news/nhtsa-tesla-autopilot-invest...

dang - all I had to do was google "Tesla FDS turns itself off"

That link you provided doesn't show any demonstrations showing it.

So? It's clearly happening, that's what the NHTSA found.

Tesla counts any incidents that occurred within five seconds of FSD/Autopilot disengaging as being "under autopilot". Given the driver is meant to be paying attention and ready to take over at any time 5 seconds is plenty of time to take over (and come to a complete stop from ~70mph)

Oh cool, it'll point your car at a child and give you 5 seconds to fix its own mistake! How generous! What a great system! Fully autonomous driving except for when it makes a mistake...

The burden of proof is on the technologists, requiring them to release footage keeps them honest where human life is at stake.

It's not an unreasonable ask in those circumstances.

All they have to do is statistically have a smaller rate of accidents than human drivers, and his statement would pass. And honestly I wouldn't be surprised if that's true. Doesn't change the fact that people hate the idea of machines accidently killing people more than people accidently killing people, but it's not that crazy a thing to say

Humans get into accidents otherwise avoidable by most other driving humans every day.

> Help me out then.

It really isn't that hard: in it's current state it's too good to keep drivers engaged and not good enough to be relied on at all the times.

Which end up in a lot of dumb accidents like plowing full speed in emergency vehicles, turning on train tracks, emergency braking for shadows on the road, &c.

True yet probably irrelevant, because self driving and humans are very likely to have different failure modes. It's quite possible to have an overall safer AI driver, but its fewer accidents to be ones that humans don't tend to do. It's still safer overall, and thus desirable.

And a separate objection: I wouldn't be surprised if most accidents by humans were actually easily avoidable by an inexperienced 16 year old - if he was paying attention. If you watch a filmed accident you definitely pay attention, and it's easy to see what should have been done. It's not particularly relevant on how people actually drive day to day.

Most accidents by humans are due to intoxication. Could autopilot/FSD do better at paying attention to a street light than a drunk 21 year old texting with his cheating girlfriend at 3am? Yes. That’s where humans make fatal mistakes.

Where do Teslas make fatal mistakes? Driving into road barriers at freeway speed. Driving into stopped emergency vehicles. Driving under semis. These are fundamentally sensor problems. Basic issues a 16 year old can handle.

If we want our new drivers to be safe, they spend hours in driving education class, watching footage of car wrecks, so they understand simple things like - don’t drive around railroad crossing barriers. Why shouldn’t we have training of - don’t engage camera based full self driving when driving a curvy mountain road into the sunset with intermittent freeway dividers. There are a class of failures that FSD can’t deal with. And yet it feels like people are telling me that - cars being hit by trains are rare, why are you edicating drivers to be careful around trains? Don’t you know how much safer trains are at hauling freight around the country. You’re just making people scared of trains.. all the time, the railroad barons are employing lawyers and lobbyists advocating against putting up railroad crossings, or letting the public know that a car on a train track is a bad idea.

There are likely basic failure modes that Teslas are bad at. But, as an informed consumer, I should have the right to know what those failure modes are, so I can assess if I want to engage autopilot in a school zone. Should I engage autopilot while approaching a school bus that just pulled over? Will the Tesla stop when the bus’s stop sign pops out? At what angle does that stop sign not get detected? What if the sun is behind the bus? Again, these are scenarios that a 16 year old can handle, and yet, there’s nobody on this forum who would be willing take a $1k bet that today, a Tesla would properly handle that situation in all angle, timing, lighting conditions, but are instead willing to say, those pesky kids, crossing from their school bus, are acceptable collateral damage… don’t want people to understand these corner cases, because Tesla can detect a few red pixels better than distracted drunk teen.

I see that backfiring horribly, e.g. see reddit's /r/idiotsincars. If that is your standard it seems much more reasonable to expect that "normal driving" will require substantially higher insurance premiums or even special permits within like 20 years

This is true to form. Musk recently told a Tesla owner to stop complaining about "full self-driving" not working very well. Musk plainly believes in free speech for himself, but not for others:

https://www.thedrive.com/news/elon-musk-tells-tesla-full-sel...

The reality is Tesla is not competitive with other manufacturers on simple tests like self parking:

https://www.youtube.com/watch?v=nsb2XBAIWyA

Instead of running around trying to suppress "bad press", Tesla should focus on making their cars better.

> "to cease and desist, on behalf of my client, [Tesla]." Signed, crybaby [Tesla]'s lawyer. Okay. Well, I've got two words for you, [Tesla]: [beep] you.

If they actually try to do something about the videos, one benefit of moving Tesla HQ: Texas weakened its anti-SLAPP laws back in 2019.

Remember a few months ago when Elon Musk claimed to be a "free speech maximalist"?

I'm going to repeat a perception I saw on here that seems to ring true.

We all know Tesla/Elon is 69% full of shit.

Say it louder, plenty of people right here on this board still deny this.

I'm of the position that speech we don't like should be countered with more speech. Therefore, I'd look forward to airing all the dirty laundry of every other EV manufacturer.

It already is...in detail, by the newspapers and news shows all the time.

The difference is that other car makers don't claim to be saving the world, nor do their CEOs claim to be "free speech maximalists," nor do they go crying to their followers every time they or their company gets even the slightest bit of criticism.

I've honestly never heard any story pushed as much as are the Tesla ones. Maybe it's just there isn't as venomous anti-following attached to other manufacturers as there is to Tesla.

The media doesn't push the stories because they're just regular news stories.

Tesla deliberately seeks out media attention, so Tesla stories attract more attention. Live by the sword, die by the sword, and all that.

>Tesla deliberately seeks out media attention, so Tesla stories attract more attention. Live by the sword, die by the sword, and all that.

Tesla also does not pay media a dime. All the other auto manufacturers do via auto ads, in hundreds of $ per vehicle sold. Some also own newspapers directly!

Publishers will do everything to favor their paying customers and make the others interested in getting more favorable coverage.

Sure. It's all a big conspiracy.

Here, watch some Tesla ads and be happy:

https://www.youtube.com/watch?v=NfMtONBK8dY

Look at the little panda!

When did he use the words "maximalist"? I think the term he used was "absolutist".

However I don't get how libel is covered under free speech. It's one of the few forms of speech in the US that is not protected.

A video of someone operating a Tesla is not and cannot be libel. That is ridiculous.

He has said both maximalist and absolutist[1].

1. https://www.washingtonpost.com/technology/2022/04/18/musk-tw...

Reply on news.ycombinator.com