Tesla is damaging the idea of self-driving cars in spite of the media, not because of it.

Self driving cars were billed as the cool new thing. Then Tesla started charging people for "Full Self Driving" with a system that regresses from basic LKA that was shipping in the mid 2000s in terms of safety by using public roads for beta testing an intentionally hamstrung* L2-billed-as-L5 driving stack.

It's frustrating to watch as someone working in the AV space, and it's silly to act like this one case was isolated enough that the fact it was wrong changes any of that reality. We know Tesla's stack has killed people: it got normalized when the first few times it happened they got away with it by victim blaming. By 2021 it was already accepted, this incident was already business as usual.

*humans just need eyes so self driving cars just need cameras

The thing is, I would have thought that Tesla's stance on repairability alone would have made their brand radioactive to the average Hacker News commenter, way before the Pedo Guy accusations or Elon Musk burning a bunch of money to make Twitter worse. Tesla is everything we claim to hate about Apple.

If Apple owned Tesla, there wouldn't be enough room in this comment to list all the reasons HN would hate it:

1. What do you mean, the OS and all its software is proprietary? Why can't I sideload my own fart app?

2. Highest margins in the industry? No we won't praise their business genius, we'll complain that they're charging a Tesla Fanboi Tax.

3. Like you said, RIGHT TO REPAIR, this is John Deere stuff, only it's a car and not a tractor.

4. Shit build quality would have us raving about butterfly keyboards 2.0.

And so on, and so forth. Steve Jobs had a "reality distortion field." So does Elon Musk, and it's aimed squarely at a certain very obvious demographic.

From a business perspective, a demographic bigger than the HN readership.

> a system that regresses from basic LKA that was shipping in the mid 2000s in terms of safety

Cite? Everyone likes to throw around the word "safety" as a qualitative thing, but the only data we have points to the system being extremely safe.

Mid 2000s LKA was passive systems in the US market: they were backups that vibrated the seat or steering wheel, or provided mild outputs. That meant they couldn't actively put the driver into danger, making them inherently safer than a system that will gladly drive onto a sidewalk

Now what about active systems from back then? Almost 20 years ago Lexus had learned the lessons that Tesla kicked and resisted until about 2 years ago on. You need to watch the driver to ensure they're full attentive: https://lexusenthusiast.com/2007/09/08/a-look-at-the-lexus-l...

In markets with LKA, that same module was used to ensure that drivers were actively watching the road while it was activated.

In other words a 2007 Lexus did a better job on the one thing that Tesla is still getting dinged on (https://www.consumerreports.org/car-safety/tesla-driver-moni...)

Tesla blamed a man for playing Candy Crush when his Tesla drove into a barrier and killed him. Lexus had solved the problem that killed this man over a decade prior.

A 2007 Lexus would have used it's Driver Awareness Module to recognize the man was not paying attention and stopped lane centering. The solution was not complex, not expensive, iirc Tesla even had interior facing cameras by then but was still resisting turning them on until much more recently.

Tesla put their not-marketing ahead of human life, because if they had enabled more aggressive awareness monitoring than capacitance sensing or silly steering wheel jiggles, Elon wouldn't get to tweet about how "you're just there for regulatory reasons"

-

Also slightly off topic, but I think my biggest internet pet peeve is when people ask for a citation on a complex problem that requires critical thinking. I can't chew and pour the conclusion for you, but you're free to research before making low effort commentary like "Cite"

Are you saying a Lexus has never had a driver fatality from inattention? That seems ridiculous. What are you actually claiming here? If Lexus is safer than Tesla that seems like something someone can measure. Have they?

> Are you saying a Lexus has never had a driver fatality from inattention?

That appears to be a deliberate misreading.

The GP was pointing out Tesla refused to follow well known best practices from around the beginning of Model S creation. Despite it being pointed out to them as a reckless choice numerous times.

They never said Lexus was perfect, they showed evidence Lexus appears to have taken the risks involved seriously in a way Tesla hasn’t, as evidenced by numerous recommendations from NHTSA that they did not follow or followed belatedly.

I gave a concrete example of a systematic decision that Tesla is making that has resulted in avoidable deaths. I showed how a comparable system was trivially capable of avoiding the same cause of death due to its manufacturer making a decision in the opposite direction of Tesla: Putting human life above the marketability of a piece of tech.

If you want to somehow "interpret" into saying no one has ever died driven distracted in a Lexus, you're free to, but that's an absurd leap I won't bother engaging with. There's always some room for (mis)interpretation, but I think you're taking a football field on this one.

>Now what about active systems from back then? Almost 20 years ago Lexus had learned the lessons that Tesla kicked and resisted until about 2 years ago on

Lexus/Toyota doesn't "learn lessons". The are simply ultra-conservative about adopting anything so when anything is a dud or the V1 version sucks they wind up looking smart.

... what and utterly bizzare form of sour grapes.

"They're not so smart, they just took a conservative approach to the life or death problem of drive attentiveness that resulted in their V1 solution outperforming Tesla's V3.x solution which came out 16 years later"

I mean for all their risk averseness, the Corolla got driver awareness monitoring before Tesla did, I don't think your point really works...

> And the fact that this was media disinformation won't matter much, because the damage to the Tesla brand and the idea of self driving cars in the public's collective consciousness has already been done, and won't easily be undone by these newly revealed facts.

The "damage to the Tesla brand and the idea of self driving cars" wasn't caused solely by this one incident.

> "A lie makes it halfway around the world before the truth even gets its boots on."

It's not a lie to say cigarettes kill people by lung cancer, even if it's later proven that one smoker who died in some study actually got lung cancer from a different cause.

It wasn't "media disinformation", it was the local constable and his accident investigators being wrong.

https://www.khou.com/article/news/local/tesla-spring-crash-f...

And even in that article, there's an addendum a couple of months later stating that a person was seen getting in the driver's seat and that the NTSB reenactment wasn't able to engage the autosteering when trying to replicate the crash.

I don't know what more you want from the press. They reported what they were told by the authorities. And then when new information came in, they reported that as well.

> it was the local constable and his accident investigators being wrong

Possibly it's the media's job to question when law enforcement comments seem implausible?

True, but are you questioning the NTSB as well?

What general reporter would have the experience and evidence to override an authority speaking with that confidence? They're reporting news. They report the law enforcement side and then the Tesla side. They updated (at the top of the article) twice as more information came to hand.

What most people actually want is better attention from the public, not better reporting from the media.

Related to https://en.wikipedia.org/wiki/Brandolini%27s_law

Conspiracy theory: Autonomous driving is being undermined by traditional car companies who see that if it is successful, it would hugely change personal transportation and greatly reduce total sales of vehicles.

> greatly reduce total sales of vehicles.

This is one of those things that always struck me as a weird, poor argument... how is autonomous driving supposed to suddenly make everyone want to stop owning their cars and use taxis instead?

When you own a car, you typically only use it a small percentage of the day (it certainly is true of the average owner). When you're offering a ride service, that one car is being utilized a much higher proportion of the day.

Ignoring the driver, taxi/ride-hail/autonomous-ride services are much more efficient uses of a vehicle over time. (And this is just a side note, but obviously if you don't have to pay for a human driver, you make the cost of operating the vehicle much lower per kilometre, and then it becomes much more attractive, from a purely financial POV, for most people than owning their own vehicle.)

When you own a car, you don't want to wait 20min before it's ready for you. You want to bring it in remote places where it will stay so that you won't have to wait 2h for it to come back. And you want to leave your stuff in it.

If you are ready to wait for a car to come pick you up, or to plan your trip based on the schedule of the car, and to pack your stuff in one bag that you can carry along...

... then, my friend, you are ready for public transportations :)

Those might be considerations for you. For many others, they'll love their car out earning revenue when they don't need it, and plenty of people who want door-to-door clean transportation without having to wait at a bus stop and without having to own a car.

Much like car sharing today, you'll have the option to keep the car at a remote trailhead or whatever, so it will be there in 2 hours. You'll just have to continue paying for it. Otherwise, you can end your ride and take your chances that it will still be there when you get back. In areas outside the "home" area, you'll have no choice but to keep paying since you can't "return" the car there.

People that live in cities, where this is most likely to catch on earlier, already can't leave anything in their cars unless they want to come back to a smashed window.

> People that live in cities, where this is most likely to catch on earlier

in cities where the public transport sucks, you mean? I just don't understand who would ever want to use a car inside Paris or Amsterdam, for instance.

The wait for an Uber is typically much smaller than the walk to the transit station + wait there (I don't think I've ever waited 20 min), you aren't restricted to one bag, and there is no schedule. The wait would get even smaller if there were 10x as many cars participating on that platform.

Car sharing is inconvenient because you first have to get to the station. Public transport often doesn't take you where you actually need to go, you're stuck to a schedule, you can't transport anything big, and it's often slower than a car. Cheap robotaxis fix that.

More importantly, I think many people have a car because there are some trips that just aren't realistically feasible without, and then they use the car because they already have it, have paid the fixed cost, and it's marginally better than public transport. With cheap robotaxis, they might use public transport for 90% of the trips because now they don't have to choose between owning a car and being unable to do those remaining 10% of trips.

I guess it depends a lot on the quality of your public transport.

Where I come from (not the US), in cities you have a bus/tram/metro every 5-10 minutes, bringing you everywhere, for a fraction of the cost of an Uber (which is why nobody uses Uber).

Places that don't have good public transport are quite remote, usually there are no Uber there, and if there is a taxi service it is quite far away so you have to wait and it's expensive.

Finally, moving one person at a time in a big heavy vehicle is less energy-efficient than having them in public transport (or better: taking the bike!). So public transport + bike again wins there, I think.

But again, it depends on the quality of your public service.

If a couple percent of cars were in this kind of system, they'd be waiting all over. It would be much less than 20 minutes. They could even set up equivalent "stations" where there's almost always a car parked within a couple blocks and it can get there before you do.

I don't go on a lot of long-term trips, but if I did that would still be a reason to get one car for a household instead of multiple. For rarer trips I would be happy to have a rental drive itself to my house.

If you mean day trips, then none of those issues are notable with a competent self-driving fleet except "one bag", but I don't think I've ever needed more than one bag for a day trip...

Really it sounds like you describe an efficient public transport system :). Of course not every city has that (I hear public transport kind of sucks in the US), but still it's important to know that other countries have that, without autonomous cars.

You'd have to run a bus every three minutes to get the same speed of pickup at stations.

But also those busses wouldn't be able to come to your door, or drop you off directly at your destination.

So I don't think I'm describing efficient public transport. I'm describing something like the platonic ideal of taxis in a dense city, but without the huge number of drivers.

I would say that metros in Paris are much, much closer to 3min than to 20 (20 sounds like the middle of the night).

But now imagine if every passenger instead took an individual car? That would require a lot of cars.

Yes, an individual car is more comfortable if you can afford it and if not too many people have one (otherwise you need big big roads to avoid traffic jams). But really my ideal is closer to metros/trams than taxis.

If you bring out metros, you can get to those numbers, but now you're at a tiny number of places, and most cities just aren't dense enough for that even if they're going for great public transit.

> otherwise you need big big roads to avoid traffic jams

Right, that's the main downside, but there's a lot of areas where they'd fit.

I think trams and buses work well, too. Of course buses are getting closer to autonomous taxis, but it still feels more efficient to share a bus than to have one person per taxi :).

Then of course they can be used together. In my city there is a system of rental cars that you leave at different stops: you have a card and can just take the car if it's there. The car doesn't come to you, but still it's a pretty good solution I think.

Anyway, autonomous cars have to get reliable before they can actually be used with public transports, I guess.

This seems to run counter to how humans work, or at least how Americans work.

Already many people would benefit financially from using public transport or ride-sharing, or even carpooling, but they don't. Because they want their car! Many people are paying too much to fuel up vehicles that are much too large for their normal use, but they want their truck!

Very, very, very few people think rationally about vehicle ownership, and telling them, "You know, your car can go do other things while you're working," just doesn't fly. Why would I want someone else in my car? Why would I want to have to keep my car spotless for others? What if they don't for me? What if something happens and I want to go to lunch early, or there's an emergency and my car is 20 minutes away? These aren't rational concerns for most people, but even if a company managed to truly deliver level 5 automatic driving, I can't imagine the majority of people would use it to let their cars go any farther than the nearest parking spot without them.

I don't see how that lowers the amount of sales of vehicles. In fact, it seems like vehicles would be used more and need replaced more often.

I think you're missing another important aspect of car usage: demand is quite bursty. A substantial portion of the workforce works 9-5 jobs, and that means that the chief demand is shuttling people to their jobs at the same time. A fleet sized to carry that demand is going to have poor utilization for the rest of the day--no better than 50%, and probably somewhat below 25%--simply because there's nothing for those cars to do.

That's not bad. Those cars still need to be charged, maintained etc

Consider this: our family has three cars for four people. We often have all three in different places, but rarely are all three _being driven_ at the same time. If a car could ferry itself autonomously between destinations, it’s a huge gain for logistical flexibility and we could probably get away with two or even one car.

If using a taxi becomes cheap and convenient, the need to own a car goes down.

I don't use a car near enough to justify owning one, but have considered getting one just because taxis are unaffordable for any bigger trips and carsharing can be a pain in the ass (mostly having to get to/from the carsharing station, inspecting for damages, and the lack of flexibility).

Solve those problems, and I will actively not want to own. Parking is expensive.

Alternate conspiracy theory: Tesla does not market its cars, while traditional automakers spend billions advertising dollars per year. Reporters work for organizations that rely on advertising.

This isn’t really a conspiracy theory. Any time you see a positive puff piece listed as being written by “staff” or similarly vague term… that’s a marketing piece that is a paid placement.

I don't know if that's true of Autonomous vehicles (it might be), but I think you'd have to be pretty naive to think that isn't happening to Electric Vehicles to continue the trillions and trillions of dollars that flow into the oil and gas industry.

I have the feeling that automotive manufacturers themselves wouldn't have anything to gain by undermining driverless cars.

The insurance companies that underwrite them, on the other hand.

You don’t need a conspiracy for the media to report that “new thing is dangerous”. Those stories practically click on themselves.

Seems completely wrong to me. Commoditize your complement. If you make using a car cheaper (by removing the labor of driving), people will use cars more.

Also, empirically, traditional car companies are actually investing heavily in self-driving tech. Their capabilities actually surpass Tesla’s.