Mercedes beats Tesla to autonomous driving in California | Hacker News Reader

Mercedes beats Tesla to autonomous driving in California

theregister.com

464 pointsbelter750 comments

What do we want from self driving cars?

Naturally I wand a car I can summon from the pub that will drive me the 20km, over country roads, to my house while I mix cocktails or have a nap.

That may never happen, current technology does not seem up to it.

Then what? Modern cruise control, that keeps my car a constant distance form the car in front (unless it roars off at a speed higher that what I am comfortable with) is great. Helpful, especially in city driving.

I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

Stopping at a parking spot, push a button, and park my car for me, less useful but still worth it.

But unless I have L5, and can turn my back on the road as I drink my vodka drink, (would that be L6? The impossible dream....) I am not interested in anything that lets me take my hands off the when whilst driving on any road. It seems too dangerous.

As a computer programmer I make my living building complex machines, and I have a very deep distrust of machines built by my comrades, at Tesla or Mercedes.

I can easily imagine that multimodal GPT-style LMs running in real-time could do all the common sense thinking that seems to be lacking in modern self driving. That future could be just a 100x optimization away.

Edit: multimodal. Does no one remember the vision + language version of GPT4 that microsoft demoed?

real-time LMs that can be reasonably allowed to take part in life-or-death situations feels a lot further away (to me) than a 100x optimization.

Why do you feel this is a language problem?

LM = language model

I fail to see how that is relevant to computer vision

You want Waymo then.

> I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

Our Bronco Sport has this (steering wheel vibration) in addition to lane assist. I don't like the latter, so I appreciate that I can enable/disable the two modes independently.

> Our Bronco Sport has this (steering wheel vibration) in addition to lane assist. I don't like the latter, so I appreciate that I can enable/disable the two modes independently.

I know it's way more stuff for them to test, but being able to enable/disable individual features is so huge. My blind-spot warning on my Mazda is super sensitive (plus I drive mostly in traffic so any time I put my turn signal on there is a car next to me). I want to be able to turn it off, but since it uses the same sensors as my RCTA, they can only be enabled/disable together. I love the RCTA, so I just have learned to ignore the beeping that invariably happens every time I signal...

I want cheap trains everywhere.

You can’t have them both “cheap” and “everywhere”.

Why not?

You don't get the economies of scale needed to justify high-quality, regular rail service along the more quiet routes.

Look at China's high speed rail network. Only a small minority of routes manage to break even, and the rest of them are a huge financial burden that the CCP just accepts because achieving a huge engineering feat creates a sense of national pride.

Cheap trains everywhere doesn’t have to mean high speed rail everywhere.

You can have slower cheaper rail infrastructure that serves less traveled routes.

Fair enough, but then nobody wants to use it because it's slow.

I like using the train in America.

The main issues I have are that the trips aren’t as frequent as I’d like and often don’t leave on time.

It also doesn't have to be profitable. Transportation is a service that's critical to economic function. There's no reason not to shift the cost of taxpayer funded vehicle infrastructure to transit infrastructure.

What those utilizing the transportation network are willing and able to pay to use the network will be directly proportional to how critical the usage of that transportation network is to economic function.

So there is never any economic need to transfer the cost of using the transportation network from those using it, to taxpayers at large. In fact, socializing costs of transportation is what can cause economically non-viable transportation services to grow in usage, and in doing so, burden the economy with loss-making transportation.

    > So there is never any economic need to transfer the cost of using the transportation network from those using it, to taxpayers at large.

Like most things in a thriving, functioning society, socializing the costs is what makes it work. Let me reframe your statement:

    > So there is never any economic need to transfer the cost of using the education system from those using it, to taxpayers at large. In fact, socializing costs of education is what can cause economically non-viable education services to consume greater amounts of resources, and in doing so, burden the economy.

Yet this same education system is what has produced a society capable of enormous economic output over the relatively short history of public education.

Your argument is the most basic of libertarian nonsense that would argue that in past decades would argue that tax dollars collected in UT shouldn't fund interstate highways in CA.

>Like most things in a thriving, functioning society, socializing the costs is what makes it work.

What is it about transportation wherein socializing the costs of using it would make it work?

Toll roads work all around the world, and better than the routes where the costs are socialized.

>Yet this same education system is what has produced a society capable of enormous economic output over the relatively short history of public education.

The modern education system arose largely in the 1930' "high school movement", which was almost all locally funded and controlled, and thus subject to very little socializing of costs.

As the costs of providing education have become increasingly socialized, the spending efficiency has decreased as rent-seeking institutions like teachers unions have taken over:

https://academic.oup.com/qje/article-abstract/111/3/671/1839...

>Your argument is the most basic of libertarian nonsense that would argue that in past decades would argue that tax dollars collected in UT shouldn't fund interstate highways in CA.

I never made a libertarian argument. I made an appeal to basic economics. That triggered your anti-libertarian ideological defense mechanism, which has been instilled in the populace by the rent-seeking institutions that rely on centralized control:

https://www.hoover.org/research/california-state-government-...

These rent-seeking mechanisms are rapidly taking over the economy:

https://ourworldindata.org/grapher/social-spending-oecd-long...

And as they do, the belief systems that they rely on, like anti-libertarianism, become more pervasive.

Libertarianism has been exposed as bullshit over and over again. To dismiss these well-documented concerns as "ideological defense mechanisms" is just sad.

FWIW homeopaths and chiropractors use the same sad defense.

> As the costs of providing education have become increasingly socialized, the spending efficiency has decreased as rent-seeking institutions like teachers unions have taken over:

I have nothing else to say but to point at this comment and say that calling a teacher's union a 'rent-seeking institution' is nonsense. As a libertarian you should be cheering for unions because they are the natural consequence of workers using their power in a free market to come together and negotiate as a block.

>calling a teacher's union a 'rent-seeking institution' is nonsense.

The evidence, an example of that which I provided, shows that they are.

May I ask if you or your relatives are in a union? If so it may be hard for you to be objective about this.

>As a libertarian you should be cheering for unions because they are the natural consequence of workers using their power in a free market to come together and negotiate as a block.

The unions that exist today are not operating in a free market. They leverage laws that mandate the employer - which in the case of the public sector, is the government - to engage in collective bargaining with them, to the exclusion of all other parties.

This provides them with an extreme barrier to competition from other workers, which in turn allows them to extract economic rent.

No, I nor are any of my relatives are in a union. That will not be an avenue of argument for you.

Second, the businesses that exist today are not operating in a free market. They leverage laws that allow them to prevent employees from leveraging their ability in engage in collective action. Additionally teachers are not required to join a union; the fact that teachers join the union and engage in collective bargaining is a result of their natural collective power. If there is a refusal to engage in collective bargaining then they can strike, both of which are natural consequences of free market unionization.

I am still not seeing where you are claiming there is 'economic rent' other than the fact that unions exist and you seem to dislike that. Given your post history, that seems to be where you actually lie.

>>No, I nor are any of my relatives are in a union.

That's a relief. When someone has a financial conflict of interest in how the public perceives a particular political issue, then it becomes very hard to appeal to their reason and objectivity.

>>That will not be an avenue of argument for you.

I'm not looking for an avenue of argument. I'm looking to reach a consensus based on the evidence.

>>Second, the businesses that exist today are not operating in a free market. They leverage laws that allow them to prevent employees from leveraging their ability in engage in collective action

None of that is true. There are no laws that allow them to suppress their employees' ability to engage in collective action. Employees are free to quit, boycott a company, make collective bargaining a condition for their employment, etc. There is no law that deprives them of any of their contracting rights for the benefit of the employer.

I am open to seeing arguments/evidence demonstrating otherwise.

>If there is a refusal to engage in collective bargaining then they can strike, both of which are natural consequences of free market unionization.

1. The law prohibits a company from negotiating with anyone but a union if a majority of a work unit vote to unionize. This prohibition extends to workers who did not vote to unionize, as well as applicants outside of the company's workforce who did not participate in the vote. This law, which the political left celebrates, is a blatant violation of contract liberty.

2. The law prohibits companies from requiring, as a condition of employment, to not unionize. By prohibiting these so-called Yellow Dog contracts, the law also blatantly violates contract liberty.

3. The law prohibits a company from firing workers who strike: if a company hires replacement workers and the strike ends, the company is not only required by law to employ the workers who ended the strike, but to continue employing the replacement workers.

Being restricted from engaging in hiring replacement workers without conditions restricts the company's free market rights, in order to make it as difficult as possible for companies to disassociate themselves from unions. In a free market, one would be free to disassociate themselves from a union.

>I am still not seeing where you are claiming there is 'economic rent' other than the fact that unions exist and you seem to dislike that.

Unions are able to derive above-market wages by restricting the contracting rights of the employer to negotiate with parties other than themselves. I've already explained that and you seemingly totally ignored me, which is concerning, given it make the prospects of a rational discussion quite dim.

Unions extracting economic rents is a very widely studied and well understood phenomenon:

https://scholar.google.com/scholar?hl=en&as_sdt=0%2C5&q=coll...

Not everything needs to be done for purely economic purposes - that includes the creation and maintenance of transportation networks. Societies are free to have broader sets of objectives.

Of course. I was pushing back against the claim in the original post, which argued for socializing transportation costs on economic grounds:

>Transportation is a service that's critical to economic function.

Transportation is not amongst the subset of goods/services where having individual market actors pay for its production leads to economically sub-optimal under-production.

Whether transportation should be subsidized for other reasons is an entirely different subject.

I am a bit less convinced about the "optimalities" in practice, because differences in funding costs, risk preferences etc. can lead to inefficiencies in the real world that a state actor could bridge.

>>because differences in funding costs, risk preferences etc. can lead to inefficiencies in the real world that a state actor could bridge.

I don't understand, could you elaborate?

For example: corporations have a lot of secondary objectives beyond profit such as limited volatility of profits that can suppress beneficial activities. State actors can take different risks and absorb such volatility by funding activities and then recovering via taxes (simplistically). In the financial markets this gets often arbitraged away, but in slow and high cost markets, those arbitrageurs aren't necessarily forming.

Similarly, time preferences can stop long-term beneficial activities by corporate actors.

In the end, it is about bridging different preferences. There are also issues of resource pooling and coordination, but those are easier to overcome.

Yes, market friction like what you note is possible. But in theory, those points of friction and inefficiency can be overcome by the market with innovations in financial instruments.

So I think there is an opportunity here.

Like many people, I cross a train track on my commute. There even used to be a commuter rail line that stopped there. Now it is all freight.

Like many freight lines, it carries a lot of "bomb trains" running oil between the fields and the refineries.

Eventually -- not tomorrow, probably more in the 30-50 year range -- those bomb trains are going to stop running.

That is going to leave a lot of track underutilized, and adding passenger rail back to these lines is one way to keep it cost effective to keep them open for freight. You have some overhead to re-add minimal stations -- concrete pads for disembarking, roofs optional -- but the tracks already existing covers a lot of the cost and complications.

Most cell phone towers don’t break even … they are just expanding the network because of pride …

I mean that's a bus or Bus Rapid Transit, because a train requires dedicated infrastructure and is a fixed route. A bus can route around problems where a train cannot.

In Boston they're trying to connect North Station and South Station, a distance of about 5 miles. The cost is at least $10 billion.

https://en.wikipedia.org/wiki/North%E2%80%93South_Rail_Link

A bus is still limited to highway speeds for any significant distance. It solves the issue of not needing to drive, but it doesn’t provide a sufficiently faster alternative.

$10B because it was neglected for a century. But that's peanuts compared to CEO compensation. So tax the rich and get it done.

According to labor union association AFL-CIO, the highest CEO compensation in 2021 was $835 million[1], less than a tenth the cost of the $10 billion estimate for the North-South connector.

Also, I would expect at least a few cost overruns.

[1] https://aflcio.org/paywatch/highest-paid-ceos

Yeah, the highest paid single individual person on that list gets only a measly 8% of the total estimate per year, therefore we shouldn't tax the rich.

The extreme mental gymnastics some people will go to to rationalize the current state of the world absolutely blows my mind.

If you want to change the world positively, the first step is accepting the current state of the world is what it is.

They were pointing out that you jumped from the CEO compensation in general they mentioned, to just pointing at one person, which does not reflect the state of CEO remuneration (and the wealthy in general, which they also referenced), at all.

We did: https://massbudget.org/fairshare/

That's still about 5 years of the total raised revenue alone to meet the minimum for just this one connector.

Roads are terrible, bus suspensions are terrible and the noise and ride quality are correspondingly also terrible.

> I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

The Toyota CH-R my wife had did that. Above 50 km/h / 30mph the steering wheel would gently "shake" if you tried to switch lanes without using your blinker.

Several brands/models offer that, sadly sometimes as an extra (FWIW I think it should be illegal to offer that as an extra: it should either come stock or not be offered at all... Safety features shouldn't be something you need to pay for).

It's not at all self-driving though.

I also enjoyed the side mirrors indicating (using some kind of yellow LED) when there was a vehicle in the blind spot.

Safety features cost money; it doesn’t bother me for them to be optional, especially since the alternative for novel ones is more likely “not offered” or “everyone is forced to buy them” rather than “given away for no charge”.

Safety features also save money, though? Through the safety. Makes sense that we could mandate they exist.

You can (and we have mandated some). Most of those that are now mandated followed a pattern of being developed as an optional, premium-charge feature, which is unsurprising and probably drives much of the development of new safety features.

What you want is called a horse!

But I want a faster horse.

...that eats electricity not hay!

> I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

Lane departure warning systems do exactly that. My mazda does this and basically between smart cruise, lane departure warnings, and big old alarms when someone in front of me suddenly hits the break, driving is SO MUCH MENTALLY EASIER. And don't get me started about blindspot alerts and HUDs oh man so amaze.

I still think Level 2 assist is worse than no assist. But level 1 assist is fucking amazing.

I agree about the parking thing. Its such a novelty thing like "hey hey hey bro check it out" and press a button. "ooooh! Aaaaah" meanwhile I can do the same maneuver in say 10 seconds flat since I've been doing it so long. Honestly maybe in 10 years the feature will be so smooth people will question how to even park without it, like rear cameras.

I rented a car with that kind of assistance yesterday and I found it annoying that the steering wheel give you some resistance when you want to go back to the main lane after overtaking a car on the opposite one!

I don't think it does that if you use your turn signal, atleast that's how my car behaves.

You need to use turn signals. Using those the algos know what line crossings to expect. My dad went through this because it forced him to start signalling. This is also probably why I'm seen more and more BMWs use turn signals... maybe it wasn't faulty turn signal wiring after all (/s)

Thanks for that. I definitely use turn signals when changing lane to overtake but indeed I don't necessarily do it when returning to the main lane because it is implicit. I am not supposed to drive on the opposite lane indifinetely once the overtake is complete and I am faster than the other car so no warning is needed.

Yeah, its a good habit to have anyways that you turn, you signal at least 3 seconds ahead of time, that way drivers around you can react. Predictability is king on the road.

Recently I was reading that in Europe, most cars sold don't even have automatic transmission, yet the drivers there are much safer than than people in the US or Australia.

There was some speculation that because they shift gears manually, they have to pay more attention to the road and can't do things like drink coffees while driving. They also enjoy driving a lot more than we do.

I wonder if other technology would have a similar effect of ultimately making driving less safe and enjoyable. I've never driven a super-modern car, but I do know that I zone out a bit when cruise control is on...

Lead footing is a phenomenon exclusive to automatics. Basically, you are on a 35mph road maybe, you lead foot some, the car shifts from third to the overdrive gear, now you are going like 45 or 50mph with the engine barely making any noise, you thinking this is just fine. On a manual, you'd know if you were in third and went to 50mph—the engine would be howling at another couple thousand rpm depending on the car. Going into your overdrive gear would require a conscious effort to shift into it.

I think it’s a lot of factors because when driving an automatic ‘99 Honda Accord, it’s pretty obvious when you’re not driving 35mph anymore.

That said, after driving the same vehicle for a while, at least for me, I can usually guess how fast I’m going.

That‘s just silly - I drive my (manually shifted) car basically solely in the highest gear in non-city traffic and have the same effect. Nothing to do with automatic, but with powerful cars.

The driver test is super hard and expensive in France. Never got it for this reason even after spending 1k euro on it. Came to the US and got the driver license in a week for like 200$ including lessons lol

> Never got it for this reason even after spending 1k euro on it.

Really? In order to pass the test you have to pass the theory test which is just about knowing the rules of the road. Yes many experienced drivers wouldn't always pass it first time, especially when it has been a long time they got their license and laws have changed but you get it quite easily by just ready the "rulebook" and do trial tests at your driving schools until you pass tests. It is no different than passing any easy test as a student.

As for the practice, you get something like mandatory 20 to 30hours with an instructor. And most of it is just about applying the rules you already know by having passed the test above. It takes less than a handful of hours to master clutch and gearbox, same for parallel parking, rest is applying the priority rules and showing the instructor...and then the examiner that you apply them and give a quick look in all directions and your blind zone whenever you change lane or direction. It is more about feeling comfortable and confident on the car than learning really.

Me and a bunch of friends failed the test the first time, and then I had to wait 6 months to pass it again. If you try to pass it without going through a school it’s the same wait time.

The booklet is also pretty big, and the trial tests are pretty hard with hard questions.

In the US I just went there and passed it on the spot. Super easy. No way anyone with half a brain could fail it. The code book is so small, there are barely any rules compared to France. Worse than that you’re barely supervised. Some lady yelled “no phones allowed” to a girl passing the test on the computer with her family around her lol.

Then I just drove with my gf who already had the driver license (legal in the US). Took 2 hours of lessons. 2 weeks later I was passing the driving test. Super easy as well.

> There was some speculation that because they shift gears manually, they have to pay more attention to the road and can't do things like drink coffees while driving.

I am highly skeptical of this. I am sure auto insurers in the US have looked at the statistics, and if it were true, they would be charging higher premiums for automatics than manuals.

There could be a shortage of data to control for obvious things like disproportionately many sports cars having manual transmissions.

I doubt it. There are so many vehicles, especially historically, from Toyota Tacomas, Jeep Wranglers, Honda Civics, to yes, Porsche 911s that are available with both AT and MT. Considering the number of vehicles insured in the US, and the fact that you can do paired comparisons like this, if there is a meaningful difference, insurers would see it.

edit: link from State Farm, a large US insurer. https://www.statefarm.com/simple-insights/auto-and-vehicles/....

Studies haven't really shown which transmission type is safer. Automatic and CVT transmissions do allow you to keep both hands on the wheel at all times, while manual transmissions may require more attentive driving. Your driving habits will play a large role in your safety. Safety features available for both automatic and manual vehicles can make a difference, too.

I’m still skeptical there’s enough data to distinguish from the obvious selection effects. The difference in safety between manual and automatic 911s probably tells you essentially nothing about whether a random driver would be safer driving a random automatic or a random manual.

It is hard to distinguish indeed, since there is no noticeable difference.

Jokes aside, as a data scientist of a long career, I don’t share your skepticism. It would be a hilariously easy task to extract this information from the data with the variables involved and the records insurance companies are able to keep.

As an expat from Germany I also highly doubt this.

After a recent 3 week driving trip through Europe, I can anecdotally back this up. It’s not even the manual transmission, but also the much smaller roads with no shoulder where at times you meet a car, have to slam on the brakes, and decide in the moment whether you or the other car will back up to a turn out. You have to pay constant attention and there’s little room for looking at phones, eating, etc. In the US we have such large roads and shoulders that you can zone out and are more easily tempted to take your eyes off the road.

This is backed by infrastructure studies. It's called traffic calming and it's done on purpose in places with advanced infrastructure such as the Netherlands.

In the same vein, I encourage neighbors to park their cars on the street rather than in their garage or driveway so the neighborhood street becomes narrower for traffic due to a row of parked cars on each side of the street. Works better than posted speed limit.

That's just a bad idea. Cars passing through the street won't see the kids and kids won't see the cars.

It is kind of a moot point when so many cars in the USA end up crunching people on the sidewalk or their own driveway...where the kids usually are.

I'd rather have one of those people who drive distracted crash into my parked car than on my lawn.

Demonstrably incorrect on my street, which has hotly contested on-street parking that narrows the roadway to a single lane. Cars have to stop and reverse into parking spaces to pass each other. Yet children play ball games in the street, and stop to move out of the way of the cars, with the cars barely needing to brake.

Children are not mindless suicide machines; they learn at a tremendous rate from their environment. The children you see around you might not survive on my street, but that's because they haven't lived there for all of their lives.

UK here. I have driven both autos and manual cars.

It barely takes any mental energy to drive manual. After you learn it it's just not a big deal.

I drink coffee whilst driving.

Main difference is that American roads are big, grid layout, traffic light to runway to traffic light style things. UK roads weave inbetween buildings. It requires more attention.

> Recently I was reading that in Europe, most cars sold don't even have automatic transmission

I don't have any numbers regarding sales, but after renting 10s of cars in Europe during the last decade, it seems like they almost never have any manual cars anymore. In some countries, the entire fleet doesn't even include any manual ones anymore, I was told the last time I wanted to rent a car.

That’s true for rentals, they don’t want idiots screwing up their gearboxes. But, most cars on the streets are manuals, and you can’t get a driver’s license if you don’t know how to drive one.

> But, most cars on the streets are manuals

I don't know, anecdotally, all my friends (besides me, if I count as a friend of mine) drive automatic.

> and you can’t get a driver’s license if you don’t know how to drive one.

This is definitely not true in a bunch of countries, where you get to chose if to take a license for manual or automatic, where if you chose manual you can drive both and if you chose automatic, you obviously only get to drive automatic ones.

That is only true in poorer European countries with older, cheaper vehicle fleets. Most new car sales in more affluent countries no longer have manual transmissions. Modern automatics with 8+ forward gears or CVTs are more fuel efficient than manuals and easier to integrate as part of hybrid systems. And of course EV sales are growing rapidly: they don't use conventional transmissions at all.

New car sales - of course, but average car age even in Germany is around 10 years, it gets better only in wealthy small countries like Switzerland or Luxembourg.

I seriously doubt there’s a majority of automatic in Germans and French’s individual buyers. In fact most people that don’t rent cars don’t are very hesitant when required to drive one (if that ever happen)

You can get an automatic-only license in the UK AFAIK. But I don't know anybody that has that.

Even in Italy?

Just now I went to Avis.com , set pick up in Rome tomorrow, drop off in Rome the day after, and it said there are 6 types of cars available, and 5 of those types have manual transmissions.

I was really shocked in italy by the standard of driving. I was fully expecting full on car honking road rage in cities and uber fast driving on the motorways/highways.

Florence was incredibly calm, respectful and considerate driving. The Motorways everyone was doing just under the speed limit.

My italian friends said that its not the case in the north.

Apparently in Italy the speed limit is strictly enforced by automated cameras. If you go more than 5% over, you'll get a fine.

Wait... Does the US not have speed cameras?

Hmm, some states do. I didn't know that. The majority don't. California doesn't.

For the ones that do, I would guess they have more leniency than 5%, but I don't know.

Call them instead and ask. Multiple times they have manual available on the website, but after choosing it and going there to pick it up, it turns out they don't even have manuals in their fleet! Incredible

Americans should look outward with humility to policies which are proven to be effective. However, that may be bridge too far, given the inward looking hubris which has bedeviled road safety in America.

> in Europe, most cars sold don't even have automatic transmission

Something like 60% to 80% of new cars in European countries are now sold with automatic transmission. And this figure is growing year to year. It will take a decade or two for the old manual cars to get replaced, though.

This is a very recent phenomenon driven by hybridization and fuel regulation (auto get better mileage because they have more gear). But you are right, they sell a lot more auto nowaday than they used to.

I'm from EU, and i've been driving both manual and automatic transmission. I can testify that auto transmission is safer, your hands are 100% at the steering wheel. Most people cannot turn and change gear at the same time.

I've also spend quite a bit time in US driving. EU is safer as narrower streets/lanes and smaller number of lanes (on highways), closer parked cars and pedestrians/bicycles forces you to pay much more attention.

> Most people cannot turn and change gear at the same time.

That's just not true. If you cannot do that, you won't get your driver's license in most countries.

Sweden has very strict tests for the drivers license, and afaik you can easily pass those without ever turning (more than a fairly small bit) and changing the gear at the same time. With the planning you are supposed to learn this is kind of a natural consequence.

Completely anecdotal but I have never been the -source- of an accident in 30 years of driving. I have been hit 4 times by people being careless. I have always driven stick because I prefer the control of gearing and it kind of gives me something to do. If there comes a day I can't get a stick shift in a new car I will likely just keep fixing up older cars. The only I would replace that feeling if I can completely trust AI in cars to get me to work/destination without paying any attention at all.

Europe would be in a depression if you couldn't drink coffee while driving stick, which you certainly can.

Hehe that has nothing to do with having to watch the road because of stick shifting, that's the result of road safety design and setting very high demands on driver's permit. Where I'm from, it takes an average of 40 hours of driving lessons in which they drill you an driving safely, before you're able to pass the driver's license exam.

I'm sure it has more to do with the fact that roads are well-designed in Europe. While in the US it's not at all uncommon to have 4-6 lane roads in front of houses meaning drivers have to watch every house for a possible car pulling out of their driveway.

There a loads of cars with automatic transmission in Europe - especially the newer, medium to premium level cars. But it definitely isn‘t as prevalent as in the US.

It is no problem at all to drink coffee or even eat while driving a manual car, so no idea where this comes from ;)

Hmmm, being from Europe and having driven on a holiday in the US, I'd say that people are more careful and respect rules better in the US than in Europe (certainly than Southern Europe, but even in parts of the North). Driving in US reminded me of driving in the UK.

Of course, US is a big country and my perception could be not very typical. Extrapolating from one data point is dangerous. It's like comparing driving in Naples and Amsterdam. Both are Europe, but the driving experience is fairly dissimilar.

I agree, driving in the US (only went to Tennessee, Alabama and Georgia) is more relaxed than in Germany and driving is less aggressive (yeah, there is the occasional nutjob in Atlanta). I like the four-way stops in the US, which forces people to communicate with each other.

> There was some speculation that because they shift gears manually, they have to pay more attention to the road and can't do things like drink coffees while driving.

I don't think its the gears, Its a combination of education. You have to pass a reasonably intense exam that requires coaching and also pass written test (unless you're Belgium, although that might have changed)

Moreover road laws tend to be much more strictly enforced. And not just parking laws.

As an American who drove a manual car in Chicago for a few years, trust me, there are ways to eat and drink while rowing your own gears.

Most European countries have far higher barriers to entry for driver's licenses compared to the USA. A German driver's license costs around $2000 to obtain. I spent $2000 on my first car (in 2018). Combine high financial barrier with competent public transit, and people who would be bad drivers just don't drive at all.

If you live in a high traffic area having your car drive you through the stop and go is currently possible and a huge quality of life increase. Stop and go traffic is 100% the worst experience when driving, having that automated is awesome.

You reminded me of the biggest annoyance I have with the systems that work in stop-and-go traffic that I've tried -- they often leave too much space before they 'go' and where I live, that's just an invitation to have someone from the next lane over jump in. So the car waits too long, takes off quickly to fill the gap, and then has to brake suddenly because someone else already took that gap. You end up being yanked around, and if the driver behind you is a bit of a leadfoot can result in you getting rear ended. I've seen it happen a few times when I still commuted.

I’ve been using self-driving cars for a while in SF. I just order one in the app, it comes, drives me to where I want to go. It’s like uber except that it’s for free (for now) and no driver.

> What do we want from self driving cars?

Nothing. Absolutely nothing. I want public transportation infrastructure and far, far, FAR fewer cars. I think the US should strive to reduce car ownership by 75% over the next 20 years, at least in cities over a certain population size. I don't want any more money invested in self-driving cars. But I'm also a grown-up of a certain age and realize the bright-eyed 20-somethings don't understand the meaning of the word "consequences".

Yeah, self driving cars would be a step backward as currently imagined.

We need high speed rail to connect our cities.

They wouldn't be a step backwards, they'd be a step sideways.

What about inside those cities? Like from my apartment on this side of town to the other side.

What about the inside? I live in NYC and its metro and bus system is amazing, and even then it's still worse than what I experienced in Asia, we're for example in Seoul, I did not have to walk more than five minutes to be able to get on a train or bus and when I got off, I similarly did not have to walk more than five minutes to get to my end destination.

What about someone who doesn’t want to travel with other people in public transport, and also doesn’t want to walk the 10 minutes of the journey?

Then I guess it's their loss. Unless they're unable to use public transport, I shouldn't have to have the externalities of them using cars, such as carbon emissions, parking space that would otherwise be used, noise and light pollution, etc.

I guess the people having to deal with the externalities could consider them their loss.

They could pay more to cover the externalities, and enjoy relatively empty roads.

Town centers will soon be inaccessible by car, so they'll still need public transport occasionally for that.

> and also doesn’t want to walk the 10 minutes of the journey?

If they have a pre-existing medical condition: sure, they can use an alternative method of transportation.

If they don't want to walk for 10 minutes, they're bound to have a medical condition soon-ish.

Walking seems like a waste of time to me. It's a slow form of transportation and a very inefficient form of exercise.

Yet most people don't exercise, obesity is skyrocketing and personal responsibility is failing.

But hey, you got yours so who cares about others, right?

If you want avoid other people like that you probably shouldn’t live in a city.

subways, trams and buses. good systems can easily be faster than cars dinner thru don't have traffic and you don't need to find parking. for short distances, bikes are also great.

I know I'm just one data point, but for my commute to work, taking public transit is about 30 minutes slower than driving at rush hour.

Just another data point: walking to work takes 15 minutes for me. Same time by metro. Driving is 45-60 minutes.

Sound as if bike-roads are missing, a bike would reduce it to 5 minutes.

> We need high speed rail to connect our cities.

Perhaps we need to slow down a bit, in general?

If the rail isn’t faster than the car then it won’t ever see wide scale adoption. It’s not even remotely difficult to clear this bar when the political will exists. Why slow down when there’s no actual need to?

A car can go faster from Amsterdam to Berlin than the train, yet I still see people on the train.

I've paid more to take longer by train, for greater convenience and comfort.

I agree with your point in general, I'm just quibbling that speed is the decider.

Future transport is unlikely to be single mode - it’s almost definitely multimodal and will include both cars (why not self driving?) and rail.

Self driving cars are much more likely to be public transport than individually owned. It's also extremely likely that it requires ride-sharing to be viable, which will substantially reduce traffic as well as car ownership (I suspect there may be enough cost pressure that most self-driving vehicles are more like minibuses than cars, which would also help a lot with pollution and congestion, but this may require regulatory pressure). Every conversation about the transition to EVs should be caveated with 'assuming car ownership stays the same because self-driving doesn't work (and that most of the world still does not manage to provide traditional public transport options good enough to replace cars)'.

(And yes, all of this means that Tesla and most other traditional automotive company's stated commercial strategy for self driving makes little sense)

Autonomous driving is just another safety and convenience feature, like cruise control or anti-lock brakes. It won't fundamentally change anything. I see no reason to give up owning my own cars even if they can drive themselves. The biggest reason is that I use my cars like a rolling storage locker or mobile base of operations when I'm out doing stuff. There's no way I'm going to deal with the hassle of shared vehicles.

You may not want to deal with the hassle but I wouldn’t assume that applies to everyone.

A fully autonomous car (however unrealistic it may be) would be a game changer simply for the economics alone. Not paying full tax, insurance etc for a vehicle while retaining all of the transport possibility would be incredibly tempting to anyone on a budget. Not having to build housing with parking included could do great things for housing availability.

I’m not saying any of this will happen but fully autonomous cars definitely would open the door to a lot of opportunities.

The economics won't work outside some small niches. Any sort of car sharing service means a business is managing that service and charging high fees for it, in addition to passing on tax and insurance expenses to customers. The taxes and insurance on my own cars are cheap.

Car sharing (short term rental) services already exist and can be a great option for childless young people who live in a dense city and seldom leave it. We don't need autonomous vehicles to make such services work. But they're totally unsuitable for how most Americans actually live. Like am I supposed to summon a shared vehicle every time I need to take one of my children to sports practices or games and hope it actually shows up promptly? And what if I need to evacuate the family in a natural disaster? The whole concept is just totally impractical regardless of the economics.

What often gets forgotten is that demand is not equally distributed throughout the day, week, month or year.

Everyone needs to go to school or sports practice at about the same time, everyone wants to go out on Friday nights and not Monday nights, everyone needs a car for monthly shopping at around the end of the month and everyone takes their vacation at around the same time each year due to school vacations.

When demand goes up, not only will it cost more, but there is no guarantee that you'll even get a car on time.

Uber and suchlike work currently because almost everyone uses them rarely, not four times a day.

> Car sharing (short term rental) services

I would expect self driving car to have much higher utilization most of the time. In addition you have people who can't/don't want to drive and would be able to use the service.

> practices or games and hope it actually shows up promptly?

Effectively it would be a much cheaper (supposedly) Uber/Lyft/etc. alternative.

> sports practices or games and hope it actually shows up promptly? And

That's a bit of corner case and generally I would expect it would reduce the number of households which have more than 1 car rather than a full replacement for car ownership.

You're really missing the point. It won't actually be cheaper except for the few customers who rely mostly on public transportation.

Driving children to activities is hardly a "corner case". It's bizarre how disconnected many HN users are from how regular middle class Americans live. Very few married couples with children would be able to manage with one car; they often need to go different places.

I'm sorry, I wanted to quote:

> And what if I need to evacuate the family in a natural disaster?

When I said it's a corner case but pasted the wrong sentence.

> You're really missing the point. It won't actually be cheaper except for the few customers who rely mostly on public transportation.

Am I? Why would you not expect it to become cheaper overtime? Uber/etc. needs human drivers who need to be paid and car sharing is inefficient because because most cars end up somewhere where none needs them unless there is continuous bidirectional traffic which is pretty rare. Self driving in theory would solve both these issues.

> Very few married couples with children would be able to manage with one car; they often need to go different places.

Yes well they could in theory get two separate self-driving taxis they don't have to always stay in the same car together.

> Am I? Why would you not expect it to become cheaper overtime?

Ride sharing services are amenable to network affects. I.e. they tend to end in monopoly. And monopolies are not known for reducing costs for the customer. If some effective regulation could counter the network effect, i.e. make it easier and cheaper to switch between different sevices, then a monopoly could be avoided.

> Ride sharing services are amenable to network affects

Are they? Currently it's a low margin business with no real moat. It's hard to maintain high market share without continues subsidies.

> make it easier and cheaper to switch between different sevices

What barriers and/or costs are there for consumers which would make it difficult for them to switch between different services?

> Not having to build housing with parking included could do great things for housing availability.

This is a purely political decision that does not depend on the availability of autonomous vehicles. It won't happen because it would reduce property prices.

Wait what, garages are required by low in US houses?! Doesn’t that piss off people that don’t own one? Do they cheat, like building without the garage but saying daughter’s room is the reconverted garage?

It already happens. Tons of houses and apartments don't have garages, I'm living in one lmao

How far off median income are you

Autonomous driving would reduce car ownership but it almost would certainly make traffic worse. The cost to get driven around or have things delivered would be less while demand would stay the same.

Many traffic jams occur for no other reason than human behavior: https://www.youtube.com/watch?v=8qy6MrDcukk

Autopilots immediately get rid of that so I would disagree and argue the less humans the better. Have fun playing around with the "politeness" lever for example:

https://traffic-simulation.de/

3x as many cars using infrastructure not made for that type of increase in demand will cause traffic jams. No matter what the auto-pilot friendly meter is set at.

Traffic lights are a bitch, changing lanes is inevitable, incidents are inevitable.

Sure, but that's just one part of the traffic congestion problem. Add more cars, and whatever efficiency gains you gained from eliminating weird and inefficient human behavior rippling outward and causing random traffic jams will be more than outweighed by the presence of additional cars. Each additional car takes up ~15 feet plus the distance to the next car.

That adds up very quickly, and when you consider the presence of intersections and traffic lights, you aren't dealing with an isolated roadway where cars just...drive. There are very real limits on how many cars you can stuff in a given stretch of road without radically redesigning roads in a way that makes them entirely unusable for pedestrians and cyclists (i.e. replacing traffic lights with slot-based intersections[0]), meaning there's an upper limit to how much self-driving cars can limit congestion under the most optimistic scenarios.

When self-driving cars eventually manage to take hold, no one reasonably expects car traffic not to increase. Right now, you park your car and do things. There's no real opportunity cost, because you can't be two places at once. When your car can drive around on its own, there's now a real opportunity cost. If it can earn money for you with no human present, there's going to be a massive incentive to let it do so. Stores will use self-driving delivery vehicles--either their own or as part of their own delivery sharing services--for deliveries.

Will there hopefully be big improvements in pedestrian safety and decreased accidents? I think so, though they certainly won't eliminate either. But there will be an incredible amount of increased traffic that goes along with it.

0. https://www.youtube.com/watch?v=4CZc3erc_l4

"The cost to get driven around or have things delivered would be less while demand would stay the same."

Demand might also go up.

https://en.m.wikipedia.org/wiki/Jevons_paradox

Self driving fleets will be driving around to avoid parking while staying in the high use areas, and contributing a lot to traffic, perhaps more than if everyone just had their own car.

This isn’t a weird thing if you live in a place with lots of taxis. Beijing, for example, 25% of the cars are taxis, and if it isn’t really busy, you’ll be able to get one just by raising your hand on a minorly busy street.

Taxis are too expensive in most of the developed world to be that ubiquitous. But they aren’t uncommon in countries where labor is cheaper, which makes me wonder if robo-taxis in the developed world will be as common as human taxis in the developing world.

Yes, there are a some things that will push towards higher traffic, including induced demand as it becomes cheaper and easier to move around. This would need to be offset by ride-sharing (and perhaps to some extend by the reduced demand for parking, though I don't think this can really increase throughput much). Like I said, I think ride-sharing, especially with higher capacity vehicles, can offset this significantly, but it may require regulatory pressure to get everyone to play ball on it.

Ride-sharing isn't going to solve these problems. At most, you fit up to five people in a single car. Most will carry fewer people, perhaps only one or two.

If you try to scale up ride sharing vehicles so they carry more people, you now have to build specialized vehicles that make multi-stop trips where you pick up multiple parties on-demand and drop them off at their individual locations. You've now eliminated the advantages of ride-sharing or taxis--they pick you up where you want and take you directly where you want to go--and managed to recreate a smaller bus. Let's call it a small, self-driving minibus that's not a bus system but a ride share system.

Only it's much worse than a bus, because bus routes are optimized in a way that vehicles picking up and dropping off passengers at what amounts to random locations on-demand never can be. Plus, you're adding even more vehicles to the road in addition to all the other self-driving cars that are running around making deliveries, looking for parking or running off to the outskirts of the city to kill some time as they wait for their human passenger to summon them again, or hustling to try to pick up passengers to make some money for their owners.

Heck, you can't even have them drive on access-limited busways or bus lanes because, by their very nature, they can't stick to dedicated routes.

Consider this: the Siemens S700 streetcar carries over two-hundred people in its usual configuration[0] with a length of just under 95 feet, or the equivalent of around ~6 cars that are almost bumper-to-bumper and trams can be built to be autonomous.

Self-driving cars have their benefits, but they're not an alternative to actual public transport solutions no matter how we try to dress them up. In the end, and as others have pointed out elsewhere in recent years, every new transit idea eventually boils down to a train. Or, in this case, an actual bus.

0. https://assets.new.siemens.com/siemens/assets/api/uuid:3ce5a...

"Self driving cars are much more likely to be public transport than individually owened"

I've only seen this as a 'just so' greenwash pr story from the automotive industry to offset the projected 300% increase in total road miles driven anually. Please explain why peole would be compelled to ditch their multidude of private cars just because they now come with a 24/7 free driver included?

Indeed, that's the elephant in the room. Self driving cars by virtue of being self driving can be sent on trips by themselves without requiring the driver to be present. The presence of a driver is a limiting factor, once that one drops away there will be many more trips. Suddenly kids can use the car and all kinds of pick up and delivery jobs become cost effective in a way that they are not today.

Can’t wait for those human-less boxes rolling downtown carefree… what a fun time we’re gonna have painting and blasting their sensors.

Ah yes! Would love to know how the crowd that protests better public transit, pedestrian paths and bike lanes as “taking away space from cars” will respond to roads being further clogged with… empty cars?

We’re really going to spend billions and billions on further car infrastructure, more reduction in space-constrained cities towards pedestrian and bike infrastructure, and even the displacement of existing homes and communities for… empty cars?! If it weren’t so depressing, it’s a perfect indictment of the US’s car culture taken full circle.

To be fair, the current cars on the road are, on average, almost empty.

Don't know why you're being dv, I back this up:

- regarding my city (Paris) most sides are literally full of parked empty cars. I don't have the data but anywhere you go - apart the biggest boulevards and walkable very-inner-center - you'll see way more parked cars than driving ones. Indeed, 50% of space is allocated to cars while they represent 13% of the commutes [1].

- Vehicle occupancy was 1.5 in 2019 and is declining (US) [0].

0 https://css.umich.edu/publications/factsheets/mobility/perso...

1 (fr) https://www.lemonde.fr/les-decodeurs/article/2016/11/30/a-pa...

> roads being further clogged with… empty cars

Presumably the car is going to pick someone up. That means it isn't parked being useless. If this starts working at scale, supply and demand will pull ridesharing costs down in a way that reduces ownership.

It might go somewhere to pick up a package.

The problem everybody is ignoring with self-driving cars is that roads already are too expensive to maintain. Self driving cars will require tighter upkeep because they have lower margins for error, ballooning already ludicrous budgets to maintain the six million mile roadway network. Humans can drive drunk on wet gravel in the dark, but a computer needs full GPS, lidar, radar, and motion sensing cameras just to follow well painted paved roads. When there is no downtime and roads are constantly being used they will wear down faster and the computers will cause more crashes. Drains and manholes sink, paint fades, shoulders crumble, curbs deform, and bridges crack. Accelerate this by three, maybe four times and the entire nation will go bankrupt within a decade trying to pay to keep everything running just so autonomous cars can stay the default.

Because owning a car is expensive. And owning a self-driving car is likely to be even more expensive. If you can still get the same convenience for a far lower cost by just using a taxi, why would you continue to own a car? (Also, this is something the automotive industry does not want, as it will reduce their sales substantially. This is something that is claimed by self-driving companies which are not also automotive companies, and most people who analyse the economics of self-driving, especially looking at which technologies which are more likely to succeed).

rcxdude 22 hours ago | parent | context | flag | on: Mercedes beats Tesla to autonomous driving in Cali...

"Because owning a car is expensive."

That hasn't stopped families from getting multiple cars thus far, so why should that change?

"And owning a self-driving car is likely to be even more expensive."

I sincerely doubt that. Just think of all the potential for selling services and preload deals on what is basically a rolling networked computer with the user literally captured inside. You might have to pay more if you want the non-add supported version or the one without the bundled subscriptions.

I can think of two reasons on the top of my head:

1. Renting a car will become much more convenient. I can summon a car instead of having to search one 3 blocks away.

2. Selling a car with an L5-autopilot will have to include a manufacturer warranty/insurance in cases of crashes resulting from a fault of the autopilot. Tesla is already charging 10k (or 15k?) for their shitty self-driving-gimmick without a warranty. I expect L5-autopilots to cost much more and even come at a yearly subscription (which basically offsets the insurance and the required continued development). I don't expect the average person will (be able to) pay that subscription, when the far more reasonable solution is to rent a car.

This isn't what we need. I don't want to depend on the government for my primary means of travel or be limited by it.

Good thing all those roads aren’t planned, built, and maintained by the government!

The "Libertarians are like house cats" meme comes to mind

Who builds and maintains the roads for your car?

Not quite the same because they don't control as easily who drives on the roads. They can easily stop you from getting on the train though.

"They" routinely close or cordon off the roads & sections. They also suspend license privilege (obviously as allowed under laws & regulations).

This would be no different than cancelling the whole train or train route; & stopping or cancelling one's entrance pass to train station.

If the government ceased to exist, the roads would still exist. I doubt the train would continue to run

The roads would continue to exist in a road usable by modern cars for only about 10-15 years. Have you been to any failed state? :-)

What about a taxi? What about a private bus company? I'd argue that even if you don't like or trust your government there are environmental, economic, and infrastructure issues affecting the whole world that result from an overreliance on private cars. Surely you can find a way to maintain your liberty without having a car sitting in your driveway all day.

Hope you are digging your own fuel

As someone who feels the same way, I’ve lost all hope for the US in this regard. SUV/truck sales have continued increasing with absolutely zero mainstream discussion about preventing pedestrian deaths. It’s legal to have push bars and tow hitches (thus inhibiting crumple zones) with no purpose other than “protecting” your vehicle at the expense of lives.

Even ostensibly pedestrian friendly cities don’t have sufficient political willpower to hand street space over to other modes of transit at anything resembling a fast enough rate. Everything needs to sit in a “pilot” for two years before action is taken.

Excuse me, tow hitches are commonly used for practical purposes. I'm really interested in the case for making tow hitches illegal.

I think they mean hitches that people leave on their vehicles when not towing. I agree that people should stow their hitches if not actively towing for a variety of reasons.

Yeah, I’m not referring to useful ones being actively used for towing. I’m referring to people that leave them in while driving around not towing anything. Some are designed to extend upward as well with no practical towing purpose. Look up pictures of collisions after someone has a tow hitch in and you’ll see what I mean.

This ship sailed like 60 years ago (blame your parents, speaking of consequences). At this point, autonomous driving is the best we can do in places like Los Angeles or San Diego.

The ship hasn't sailed, it will just take a very long time to come back to port

If it's not out sailing, why are we waiting for the ship to return to port?

In the 70's, the city of Utrecht in the Netherlands was renovated from car central to bike central. That only took a few years, bike paths aren't that expensive.

I've been traveling around visiting US cities.

Cities like San Antonio, TX with its Riverwalk and Atlanta with its Beltline not only show every other city a path towards walkable and sustainable future but also how a city can invest tens of millions in walkable infrastructure that yields billions in economic investment and development.

Also, I don't have a car so I've been taking bus, train, plane, and occasional car rental to travel between towns and cities. Recently I've started taking buses. What I've discovered is that I can get on the bus in the morning, plug in my phone and laptop to AC power, and using my mobile hot spot work all day while on the bus. The buses travel on interstate highways which have excellent continuous mobile service. Amtrak trains are awful for working because lack of stable internet connection.

San Antonio barely has a functional downtown, there's a grim shell of tourist oriented restaurants and hotels immediately around the river walk but really struggles to be a center of business. There are few condo buildings downtown. Unlike Dallas and Austin San Antonio hasn't really seen a resurgence of it's downtown region.

Furthermore the river walk was built in the 1930s as a way to improve property values, and remodeled a handful of times over the decades, that's not a recent innovation

The San Antonio river walk has been extended 15 miles along the San Antonio river north and south of the downtown area. It has spurred billions in investment along its banks. This is exactly what my point is, there isn't a resurgence in the downtown region because new development is happening north and south along the river.

For the ship to come back to port, someone has to want it to come back to port.

I guess you don't know anyone that makes under $50k/yr. in the US. I'm guessing you probably just associate with other high-paid techies who can't understand why commoners just can't "eat cake". Trust me: when a bus route is shut down because rich people sniffle at a 0.2% tax for something they'd be too embarrassed to use, thousands suffer. If you've never had to switch from a 30 minute bus ride to a 1hr bus ride because of this, you will not understand any of this debate or its consequences, as I stated.

I was merely saying that for public transportation to be viable again in these cities, someone has to act - either by lobbying loudly and/or by putting their money where their mouth is.

Proposing that autonomous driving is the only way forward sounds pretty sad to me as it doesn't solve the main issue with cars: their environmental impact.

I didn't mean it would be easy to do, nor that it is decently doable to use public transportation in these cities' current state. Sorry if it came out that way.

You're certainly doing a great job making public transportation sound appealing to us "techies".

The ship hasn't sailed at all, very few things in life are really set in stone (for example fossils).

Doesn't mean a transition is easy or cheap. Nor will it be quick. After spending years and years both in the US and Western Europe, I wish for a healthy mix of cars and public transport. As it stands, this means significantly and heavily reducing cars in high density settlements where public transport is already sufficient or generally viable in the future.

I absolutely love cars and driving but I despise them in city centers. So much property could be regained to improve the daily lives of millions of people in various ways. There is a multitude of projects which have shown how successful this approach is.

One such project is the ban of cars from the Jungfernstieg area in Hamburg. 5 years ago you would walk along Jungernstieg and hundreds of cars would pass by, especially young men showing off their shitty, loud, tuned Mercedes CLAs to everyone. Then the city reduced access for cars there upon popular request and repurposed some lanes for plants and bicycles.

It took a while until people understood [1] but eventually they did. All sources in German.

Before (4-lane road): https://www.mopo.de/hamburg/verkehr-in-hamburg-fahrverbot-au...

After (intermediate car-reduced): https://www.hamburg.de/pressearchiv-fhh/14369764/2020-09-30-...

Future (cars banned): https://www.hamburg.de/bvm/medien/17154924/2023-05-30-bvm-ju...

[1] https://www.mopo.de/hamburg/autoverkehr-am-jungfernstieg-das...

The housing stock of a city rebuilds itself over a 30-60 year timeframe though. So starting now doesn't preclude us getting ahead of things for the future. We need some gutsy Robert Moses type figures, utilizing modern knowledge and less racism.

Throwing up your hands whining "Its to haaaaarrd" is very un-tech-bro of you.

You should look at Dutch cities, for example Utrecht got rid of a huge chunk of hideous highway in the center of the city and in its stead is now a beautiful canal with one the busiest bike-traffic streets in all of Europe.

https://bicycledutch.wordpress.com/2016/01/05/motorway-remov...

Yanks could do the same if they wanted to, but unfortunately they've been brainwashed into thinking walking is the devil

Just off the cuff saying things like this makes it seem you are the one that doesn’t understand consequences. The American economy is unfortunately driven by car ownership. So unless you also have a way to change almost every facet of American life in the next 20 years, you’re just blowing smoke.

That being said, I would love it if car reliance was drastically reduced. But it’s not going to happen.

If the US wanted to become less car dependent and a world leader in public transport, it could and would. It won’t be easy, it will take decades, but it can certainly happen.

And without a doubt the economy would be just fine in the long run.

It’s survived a depression, WW2, bailing out Europe, Cold War, Vietnam, Iraq, 9/11, Afghanistan and the 2008 crash after all.

Sure, not arguing that, but it would take an overhaul of almost everything, that will not happen in a span of 20 years. But at the end of the day, the US doesn’t want this, at either the policy or the civilian level. We like going into 60k of debt to drive our big SUVs just a little too much.

Well sure. We can’t get high speed rail in the country or even in California. I’ve heard the arguments & it is all weak sauce.

The major industrial countries have it. Here is is automakers who fight this, nowadays it is Elon Musk, who distracts with BS single file tunnels (Boring Company pushing Teslas in a tunnel) or hyperloop nonsense.

This needs to be a federal infrastructure along the lines of the federal highway system.

This country lacks the national drive and will, which is impeded by Corporate & affiliates like NADA your local 1% in every county in the US.

https://slate.com/news-and-politics/2023/05/rich-republicans...

> I want public transportation infrastructure and far, far, FAR fewer cars. I think the US should strive to reduce car ownership by 75% over the next 20 years, at least in cities over a certain population size.

Do you want to limit the cars people own or do you want to limit how much cars are used?

I could reduce the number of my cars to 1 and wouldn't be driving any less. Or I could buy more cars and still would not be driving more than I do today (but might be able to choose a smaller more efficient one for a trip).

Both, hopefully? Manufacturing cars comes with a lot of environmental impact, too.

I'm not sure you really do.

I live in the UK and public transport is actually very unreliable. Cancellations and delays are frequent, strikes happen a dozen times a year, many services have an hour between trains... public transport will take you much linger to get anywhere and you'll be able to transport much less. The car has been a blessing when shopping.

This seems more a problem in how this specific public transport is run, than with public transport in general.

I think it's more general, trains have lots of single points of failure while there are normally many roads that can be used to get to a destination.

My experience in Germany is that the trains there are even worse. Every single one was delayed by up to half an hour.

This will never, and I mean absolutely never, happen in the United States, for three reasons:

(1) Weak Federal Government. The US is more a close collection of small nation states than a single country. The kind of infrastructure you’re talking about takes the kind of planning and funding that only a strong federal government can demand.

(2) Strong respect for property rights. In the US, it takes a lot to take away someone’s private property in order re-allocate it to a public good. By now, all of the cities have vested interests speaking for every square inch of viable property.

(3) Weak labor protections. This is a big one. Labor protections are so weak in the private sector that no one in their right mind would depend on public transport if they could help it. Most Americans are one illness away from being bankrupt, fired, and foreclosed on. To make it in America, by any definition, for a vast majority of Americans (and that includes people not traditionally called Americans, like undocumented immigrants) you must have access to a private vehicle because your work place could change in a heartbeat. You could have to move because your town no longer has any jobs for you. You could have to move into your car because of divorce/eviction/foreclosure etc…

People make fun of the F150 drivers with spotless beds because they’re used as glorified, cramped minivans but the truth is every one of those drivers has made that investment because they’re acutely aware of how precariously close they are having to use that truck to do hard labor.

> I want public transportation infrastructure and far, far, FAR fewer cars.

I also want this, but I grew up in LA and thus realize that there are places where this won't happen.

> I don't want any more money invested in self-driving cars.

I would take money invested in self-driving cars over money "invested" in public transit in wildly inefficient ways that's less investment and more waste.

The problem with the position that money should go to transit instead of self-driving cars is that transit is a largely intractable problem because it's a political one, where self-driving cars are very much not an intractable problem because they're a technical one.

40,000+ people die every year in the US in car accidents. Self-driving cars have the potential to stop many of those deaths (and the injuries not counted in that). Investment in public transit, when you consider the practical realities, largely doesn't. I understand why folks prefer public transit as a solution, but self-driving cars will be a lifesaving technology, and on that basis they seem clearly worthy of investment.

And this is all doubly true since this isn't government money - it's VCs and private companies putting their cash towards solving a problem that's beneficial towards society. Why hold up public transit as the alternative to self-driving investment when there is absolutely no chance that these dollars would ever go to public transit?

This makes little sense to me?

Investing in public transportation is a much more means tested way of reducing traffic accidents. Countries with good public transportation infrastructure have much less traffic deaths per 100k than car focused countries.

Why? Well it makes sense. Public transportation is much more safe all around as you are greatly reducing the amount of people in charge of machines capable of easily killing others.

So when you say "investing in self driving cars is investing in saving lives" you are right. It's just a much more inefficient way of doing it. Funny that you would consider public transportation a waste of money when looking at the actual data shows it's the most efficient way of moving people.

> Investing in public transportation is a much more means tested way of reducing traffic accidents.

Selection bias. By definition, you're looking at public transportation that exists. How many traffic accidents have been reduced by the bullet train in CA that has had $80 billion in investment so far? Zero.

> So when you say "investing in self driving cars is investing in saving lives" you are right. It's just a much more inefficient way of doing it.

In the cases in which money can be efficiently deployed, you are correct. But has the $80 billion (let's say that again - $80,000,000,000) spent on the CA bullet been more efficient than $80 billion of spending on self-driving cars? Clearly not.

The problem is that there are limits to how much money you can deploy towards public transit at any reasonable level of efficiency. This is true because of geography, and it's true because of politics. In those places where you can use money well on transit, I fully support it, but once you've exhausted those, you're still going to be left with five figures of traffic deaths a year that can be saved by self driving cars.

Also, you ignored my last paragraph - the comparison of spending money on transit vs. self-driving is not one that makes sense, because the money comes from two different places. VCs don't invest in public transit. Google doesn't invest in public transit.

Self-driving cars are something that will provide a huge societal benefit at no cost to taxpayers. If you're a member of the general public, spending money on public transit is infinitely more inefficient than spending money on self-driving cars, because that transit money could be used to benefit you in other ways, whereas the private funding to self-driving cars would probably just be returned to shareholders or invested in other tech companies if it weren't spent there.

Personally, I don't even think the two are in conflict. Many of the comments on this article vastly underestimate self-driving's impact. Self-driving creates motives to use transit because it allows the economics of automobile fleets to consolidate around shared use instead of ownership - the pull of "the robotaxi service is cheap", which can be pulled off with a regular, modest YoY cost reduction, will ultimately make people give up the car, and if they give up the car, they will then save a few more pennies by using the bus on occasion - the bus will be more frequent if it goes self-driving too, because it's the labor cost of drivers that determines a lot of the operational economics of transit fleets. Take out that cost and you can run a lot of tiny vehicles as well as huge trains, meaning transit, deliveries and taxis will blur together as modalities, no longer needing to provision for what pencils out with a human driver or to adopt an ownership model where the user has to buy for their largest use-case and eat the associated cost. They can go smaller without much issue and rent larger as needed. It's a different vocabulary of "what transit is" once you add that dynamic.

The political element follows directly from that: if you aren't a car owner, but you are a bus and taxi user, you stop caring about parking and the specifics of your commute route. The robocar service will lobby the city for their own efficiency in your place - and what benefits their efficiency also benefits the users. So in one stroke you end up with an urban population that is aligned to empty out huge swathes of road space and garages for other uses such as bus lanes, bike parking, etc. People will feel safer on a bike because the cars will drive safely, and this is found in studies to be the main bottleneck to bicycle usage, so investment in dedicated bike infra could flatline in a self-driving environment, and bike usage will still rise!

It's hard to find an angle where there is truly a downside - the doomer response is kneejerk.

I want tell it where I want to go and take a nap.

Unrealistic maybe, but it colors my interest or lack of interest in the middle of the road options.

Half paying attention systems strikes me as dangerous for people.

> I want tell it where I want to go and take a nap.

So, Waymo/Cruise (at this point in time) basically.

Modern cruise control, that keeps my car a constant distance form the car in front (unless it roars off at a speed higher that what I am comfortable with) is great. Helpful, especially in city driving.
I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

Stopping at a parking spot, push a button, and park my car for me, less useful but still worth it.

You already have all of this in a Volvo

I wish I had a Volvo!

Good news.

>Naturally I wand a car I can summon from the pub that will drive me the 20km, over country roads, to my house while I mix cocktails or have a nap.

You're kind of joking, but this is more practical/safe than it sounds.

Most people drive to eat and drink out. You know how they all get home? By driving back. That's driving after drinking however many bottles of beer and glasses of wine. Every night they go out to eat and drink.

To put it simply, everyone breaks the law any time we feel like going out. The police everywhere know this, and the only reason they don't camp out at restaurants to arrest everyone is because they themselves are also guilty of the same.

> I am not interested in anything that lets me take my hands off the when whilst driving on any road. It seems too dangerous.

Hyundai has a lane-keeping assist that centers your car in the lane. It is _very_ good, although of course I am nearly always ready to take over if needed.

IMO, smart cruise control + lane assist gets you a great deal of convenience for both day-to-day city traffic, and long-distance trips. The automation is good enough that it is helpful and makes driving much easier, but not so good that your mind switches completely off, so that you're always ready to take over if needed.

"That may never happen, current technology does not seem up to it."

Any computerized system is only going to advance over time, with its theoretical limits far beyond the current state of the technology.

So full self-driving is clearly inevitable.

> Any computerized system is only going to advance over time, with its theoretical limits far beyond the current state of the technology.

That is not true

Name one class of software, or more pertinently, one class of AI, where that is not true.

I just need a car that will let me catch up on emails or reading the morning paper while it commutes me to work through rush hour traffic

>I just need a car that will let me catch up on emails or reading the morning paper while it commutes me to work through rush hour traffic

Bus? Train?

> What do we want from self driving cars

Frankly? I want another 10-15 years of independence for my dad. His eyesight is failing due to a mix of wonky genetics, he's fine driving in daylight, but already has started being much more cautious driving around dusk.

I want an AR display that will highlight the road and objects in it. I was LiDAR sensors hitting the brakes if it detects catastrophe. I want the car to augment and extend his ability to use it.

That's the big sell of self driving cars. Not "push a button and be whisked away," but having your own abilities enhanced by machine input that would otherwise be inaccessible to you.

I was impressed with my grandma's foresight when she moved house to within walking distance of a station, so a few years later she could give up the car and maintain her independence and health (from exercise).

That currently seems the safer bet than hoping for AI driving/aids, but isn't an option in much of the USA.

The decision making technology is definitely possible, not this year or next but in a decade? Sure, plausible.

However removing all the sensors, reducing redundancy, from a high speed moving car to save manufacturing costs and increase profits is in direct conflict with safe driving for the passenger and other people on the road.

Many more sensors, much more redundancy, roads that give feedback, networked cars that give feedback, satellites for near realtime external updates, all will make cars eventually drive themselves better than humans.

But certainly not "just cameras" that cannot even handle weather without severely compromised guessing. You don't do that in a physical world requiring safety.

I want the technology to prevent others from driving dangerously.

And myself, by extension. We all have human moments.

Running red lights, mounting pavements to cut corners, brake-checking, dangerous overtaking, inadequate stopping distances, distractions, micro sleeps, accidently accelerating into shop fronts, and just generally phasing out because you are human.

I personally hate anti-social driving behaviour.

Brake Checking is the stupidest thing I have ever witnesses. It feels like it's becoming a trend here in Australia. Did it originate as aggressive behaviour in the US, or Russia maybe for insurance?

It's certainly a recent trend here in Australia.

I drive every day in Australia, thankfully not in peak hour but hundreds of km per week and I have never seen brake checking, didn't know it was a thing, had to Google the term to find out what it means.

I have noticed a lot of people failing to keep a safe distance from the car in front (tailgating), though, which is far stupider to me. Maybe the two are related?

I don't think I've ever once been "break-checked" in 15 years of driving. Is that something that only happens to people who habitually drive too close to the car in front of them? (Tailgating as we call it here.)

My 2019 bmw has the automated ruble strips thing, which is pretty nice for long drives.

That’s exactly right. Partial self driving is not only useless but dangerous

> What do we want from self driving cars?

Security, by not making human mistakes. But this won't happen for a long time and probably never with people like Musk leading the way.

Autonomous vehicles are a poor substitute for long-promised flying cars anyway.

Flying cars already exist, they are called helicopters.

What you really want is the TCO of a car in a helicopter.

Helicopters are outside of reach of an everyday person. You can't drive them when you feel like it nor start your journey on your suburban driveway.

And so were cars at one point. Helicopters are still "flying cars", airplanes are "flying busses".

> That may never happen, current technology does not seem up to it.

This already exists, provided the route is extensively mapped beforehand, in two separate products (Cruise, Waymo)

I reckon a "drive 1-2kms on a designed route and park" capability alone would be pretty useful for city planning. Imagine if instead of needing to put parking spaces everywhere, planners could just plan robust routes that any car could use to go and park itself. You could could achieve it today by telling people to just walk from a car garage, but the uproar whenever there isn't enough parking right next to the exact place they want to stop is too huge for it to ever happen

> Stopping at a parking spot, push a button, and park my car for me, less useful but still worth it.

Car drops me off in front of my destination, then goes off to a parking garage 15 minutes away and parks.

> What do we want from self driving cars?

I want a car that will drive the hours of boring ride on the Autobahn. It should take over once I am on the Autobahn, then drive by itself while I am sleeping/working/playing, and then drive to a special "cool-down" parking lot, where it will take a few minutes mandatory break to make sure the driver has time to wake up. From there I will take over again.

Why buy "self-drive", when you can already pay a human to drive for a tiny fraction of the price?

I haven't driven a car for over ten years, and I don't miss a thing. No traffic misery. No parking misery. No government misery (tickets etc). No spending on never-ending this and that. No getting angry about roadworks or other people cutting me off and other nonsense.

It's absolute bliss.

I live in the country

I havetwo modes

1 home to beech with dog. Daily 10k total

2 town for supplies. Weekly 55k total

Not really suitable for taxies or Uber.

> Then what? Modern cruise control, that keeps my car a constant distance form the car in front (unless it roars off at a speed higher that what I am comfortable with) is great. Helpful, especially in city driving.

> I would appreciate "automated rumble strips" that warns me gently if I veer out of my lane.

Honda already does both of these, since 2020 at least iirc

> What do we want from self driving cars?

I would say, drive me long distances on the high way. The rest (last mile, city country side) I can do myself. It also aligns with the fact that it is easier to make autonomous driving work for high ways than inside cities etc.

This would make taking an airplane less tempting and as my car can fit the whole family we do not need to fill up 4-5 seats in an airplane and we use much less CO2 per passenger.

I'd be happy with a self driving car that could just safely stay in a single lane on a freeway for an hour in good weather while I did something else.

I’ve been very impressed with some openpilot forks. In some instances I managed to drive 70 miles on the highway without having to make a single correction. It’s about what I want from self driving - keep me in my lane, and keep me from rear-ending the car in front of me. I would never trust tech to let me take a nap while driving.

Friendly reminder that this system is HEAVILY limited, with the following restrictions:

  - Must be under 40 mph
  - Only during the daylight and only on certain highways
  - CANNOT be operated on city/country streets
  - CANNOT be operated in construction zones
  - CANNOT be during heavy rain or fog or flood roads

And for comparison: https://www.youtube.com/watch?v=DjZSZTKYEU4

Tesla FSD navigating the complex city streets of LA for 60 minutes with zero human intervention.

This seems like a marketing gimmick by Mercedes at best; the two technologies aren't even in the same league. Any comparison is laughable. They are outmatched, outclassed, and absolutely outdone. Karpathy (now @ OpenAI) and the Tesla FSD team have really done an incredible job.

In the end, Mercedes assumes liability for its system, where Tesla doesn’t. Tells you all you need to know.

How exactly is Mercedes accepting liability? I mean to say, how does this work in practice? Who absorbs the demerit points if your car is accused of going 40 in a 30 zone?

> How exactly is Mercedes accepting liability?

You’re indemnified.

Mercedes can’t indemnify you. Local police and prosecutors make those decisions.

Nope. Local police and prosecutors have no authority to make decisions on liability. Their authority is limited to criminal matters. However, police reports can generally be used as evidence in civil trials where judges and juries make decisions about liability.

There may be some circumstances where a driver operating a Mercedes-Benz vehicle using Drive Pilot could be committing a criminal offense. For example, I think sitting in the driver's seat while intoxicated would still be illegal even if you're not actually driving. But that is separate from liability.

> Mercedes can’t indemnify you. Local police and prosecutors make those decisions

Are you thinking of immunity? Indemnity is regularly contracted between private parties. (Also, police can’t give you immunity in America.)

Is there evidence for this?

https://insideevs.com/news/575160/mercedes-accepts-legal-res...

"once a driver turns on Drive Pilot, they are "no longer legally liable for the car's operation until it disengages." The publication goes so far as to say that a driver could actually pay zero attention to the road ahead, play on their mobile phones, and even watch a movie. If the car were to crash, Mercedes would be 100 percent responsible."

Yes, but is there any evidence this is actually happening? Not that they announced it might a year ago.

Folks, don't downvote me, show me where Mercedes actually says they are taking liability for actual cars that are actually on the road with customers. What people cite on this is a vague marketing puff piece from a year ago. If they are doing it, it should be pretty easy to find out how that is structured from their website or whatever the car owner has signed, or just like literally any evidence whatsoever.

This explains the 40mph and other extreme restrictions

for now, that is. in German they call it "traffic jam pilot" Staupilot. they start L3 autonomy with the boring and tedious scenario first.

it's obvious they'll expand it to more and more freeway/highway driving scenarios, and from there grow into any out of town driving.

meanwhile Waymo and Tesla can take their bruises with downtown traffic and pedestrians and children and hand drawn road signs and dirty signs and poorly placed ones and whichever more crazy real life reality show surprise guests appear on stage...

I wouldn't (and I don't think you did but several others in such discussions do) snicker too much on Mercedes. they have a knack on getting a couple of car related things pretty right.

This doesn't mean the system is actually better. It just means that Mercedes thinks that the cost of covering the liability won't exceed the boost in sales that come from this decision. It reminds me of what an automotive innovator once said[1]:

>Here's the way I see it, Ted. Guy puts a fancy guarantee on a box 'cause he wants you to feel all warm and toasty inside... Because they know all they sold ya was a guaranteed piece of shit. That's all it is, isn't it? Hey, if you want me to take a dump in a box and mark it guaranteed, I will. I got spare time. But for now, for your customer's sake, for your daughter's sake, ya might wanna think about buying a quality product from me.

Not that I'm implying Autopilot is necessarily a "quality product" in comparison. It is just that a guarantee or liability coverage is nothing but a marketing expense for the company issuing it. It doesn't actually mean the product is higher quality.

[1] - https://www.youtube.com/watch?v=mEB7WbTTlu4

I can understand financial liability, but what if somebody gets seriously hurt and there’s criminal charges? Is there already a legal framework in US for transferring this kind of liabilities from the driver to manufacturer?

>Tells you all you need to know.

Thanks for the subtle reminder that your statement is most definitely not "all I need to know."

Would you bet your life on a Tesla Twitter promise? They are not even betting one cent on their car safety.

> They are not even betting one cent in their car safety.

Do you mean strictly in the sense of accepting liability for crashes while AutoPilot is engaged?

Because Tesla makes safety a primary selling point of their vehicles. Objectively Tesla invests hugely in the safety of their vehicles, and scores the absolute highest marks for safety in many government tests.

Tesla makes an L2 system where the driver must remain engaged. And part of their FSD system includes the most sophisticated driver gaze and attention tracking system on the market. This has made their FSD system on predominantly non-highway roads safer than the average human driver without FSD.

> And part of their FSD system includes the most sophisticated driver gaze and attention tracking system on the market.

That’s a just blunt lie. Their driver monitoring system is very deficient. They lack, for example, industry standard of infrared camera that allows you to see through the sunglasses.

And all older Model S/X, that have FSD, lack any camera at all.

You’re right, their older cars don’t have their most advanced system. When I said “on the market” I mean the cars they are selling now, not historically.

I’ve found the system extremely adept at gaze tracking and alerting. The cabin camera is infrared in their latest models (at least for the last 2 years).

Please don’t call me a liar. I am happy to be called wrong and corrected.

I should have said “one of the most advanced” because this is in truth a subjective measure and I don’t think agencies like NHTSA rank and qualify attention tracking?

No, I mean actually bet your life, say put your little children in the back and send them to some destination because you seen Tesla videos on YouTube.

if you won\t do it for a randoms destination, on what would you risk your children life, on some white listed high-way? which one?

In minor traffic a week or so ago, I ended up next to a Tesla for a few minutes where the driver had zero hands on the wheel, and her eyes were buried in her phone, with her head angled downward. Whatever system was running seemed to be totally fine with that situation; if that's the most advanced driver attention monitoring system available, we're in a lot of trouble. Tesla caring so much about safety is so obviously a bad joke.

I don’t believe you. Teslas have a cabin camera that monitors your gaze and quickly emits warnings if you look at your phone while autopilot is on. If you ignore the warnings, the car puts it emergency flashers on, pulls over, and disables autonomous driving until your next trip. If you do the five times with the FSD beta, you are banned from using FSD.

> are banned from using FSD

Do you get a refund?

It’s a one week ban I believe.

It is not. The FSD Beta enables an aggressive attention monitor which is not active with the basic AutoPilot system.

The nag when FSD is enabled is actually quite annoying. Even glancing over at the screen for more than a second will trigger it. If it triggers more than a couple times you get locked out for the drive. If it triggers more than 5 times in total across any number of drives, you get locked out entirely for a full week.

Took me less than a few minutes, to find in the same channel, FSD failing and the human having to intervene. Here it is at the proper time and for a video uploaded 10 hours ago... - https://youtu.be/KG4yABOlNjk?t=995

And even worst failures...looking at all these I simply do not believe the videos posted are not edited or a selection of a success out of many failed ones.

https://youtu.be/KG4yABOlNjk?t=1252

https://youtu.be/WowhH_Xry9s?t=77

https://youtu.be/WowhH_Xry9s?t=196

Since the video was live on YouTube no editing was done...

These videos are only posting the successful events, we need to see them all. That is why probably Tesla is not putting their money where their mouth is.

This is a great example! I hope people will take the time to click and watch it drive.

I think it’s a tough call in this case. The Tesla is trying not to block an intersection where a light is green but there’s no room to clear the intersection on the other side due to traffic ahead. In fact an oncoming car is able to turn left in front of the Tesla because it waited.

Probably the law is that you should NOT enter the intersection in such a state, but the human nature would be to make a more nuanced judgement of “how bad would it be” to continue thru and possibly get stuck sticking out into the intersection for a bit until traffic moves again.

I would think - how long might I be stuck? Would it actual impede other traffic? Also factors like, am I’m late for work? Am I frustrated because traffic has been horrible or am I enjoying a Sunday drive?

Ego (Tesla’s name for the driving software) doesn’t get impatient. It can maximally obey all traffic regulations at all times if they code it to do so, even if human drivers behind might be laying on the horn.

This little clip really shows how much nuance is involved in designing these systems, and how doing the technically right thing can either be wrong or at least ambiguously right.

That is not the failure. The failure is that would not move anymore and the driver had to intervene and accelerate.

At 16:55 thru 17:15? Where he manually accelerates the car and blocks the intersection?

https://youtu.be/KG4yABOlNjk?t=999

16:40 to 17:12

Seemed to be completely OK, a driver who is on the safe side would have just done this exactly

If you stay there, blocking traffic for no reason, and do not proceed when you are clear to go, as it did until the driver accelerated, you will fail a driving license exam in most countries.

Maybe it would have moved a second later...

"Most countries" would fail a driver for not crossing an intersection until there is space on the other side? I think you're being a little absurd here.

The driver in this case could see that the cars beyond the blue car were starting to move, and thus predict that he could cross a little bit early (betting on the space being available by the end of the maneuver, but risking being the ass who ends up blocking the crosswalk after miscalculating).

Here in the UK you'd be failed for blocking an intersection.

Totally, you are not supposed to block an intersection, and it is forbidden to stay on a pedestrian crossing.

Again...that was not the failure. Actually there are two failures:

- First one, to stop too far away both from the road intersection and the pedestrian crossing. And of course you should not stop on top the pedestrian crossing. The car behind the Tesla, noticed that right away, and went over the Tesla. Even the Tesla driver in the video commented on that. I wonder what FSD version that driver has :-)

- Second failure, the car ahead, the one already over the road crossing started moving but the Tesla would not move at all. Only when the video author accelerated as he mentions in the video.

> The Tesla is trying not to block an intersection where a light is green

What light? There is none if you're talking about t=995.

LA has an unwritten law where 2-3 cars make unprotected lefts after oncoming traffic has cleared on yellow/red lights. Letting FSD drive, it can’t honor this “cultural” (technically illegal) behavior. If I am operating it, off goes FSD in that moment.

Tesla has been specifically reprimanded by NHSTA for allowing FSD to drive like human rather than to the letter of the law. The rolling stops is one I remember, but basically it'll apply to anything.

Dude it's totally following the law. You're allowed to enter the intersection for an unprotected left even if it's not clear to turn (possibly not if the end of the intersection you're going to is full of cars). If you are in the intersection, you're allowed to clear when possible regardless of the color of your light.

Most intersections with unprotected left turns let you fit at least a whole car, sometimes two in the intersection to wait, and the third car has its front wheels on the paint, so it's totally in the intersection.

While you are allowed to enter the intersection during a yellow, you are also considered legally to have been warned and would be consider liable if doing so causes an accident.

This extra liability would seem like making self driving cars error on the side of stopping to be in the interests of auto makers who would be exposed to that liability.

Incorrect. As long as your front wheels pass the stop line prior to the light turning red, you are legally driving through the intersection and should continue and clear the intersection on the other side (not stop in the middle).

Imagine how crazy the law would be if this were the case: - Light changes from green to yellow 1ms before your front wheels pass the stop line - Other direction traffic runs a red light and hits you from the side - You're now somehow liable because your wheels entered the intersection with zero ability to react quickly enough (no human can react in 1ms and no car can stop that fast) and the other driver that clearly blew a red is not?

That would be pants on head crazy, tbh...

Mostly the cars will have entered during the green, not the yellow, and were waiting for a chance to turn. When there was no opportunity to turn while the light was green, they must turn while the light is yellow or red, because they are already in the intersection and must clear the intersection.

How would entering on yellow cause an accident?

Correct. The law states that as long as your front wheels are in the intersection prior to the light turning red, you should proceed through the intersection. Inexperienced drivers that either stay in the intersection or try to reverse into traffic behind them are breaking the law and create a huge hazard for others.

Even if the light for opposing traffic turns green while the turning car is still clearing the intersection, opposing traffic is legally required to wait and not enter the intersection until opposing traffic has cleared the intersection.

It’s legal to turn left in such a case as long as you entered the intersection already. So at least the first 1-2 cars are not doing anything illegal

You can't legally enter an intersection (advance past the stopping line if present) until you are clear to perform your transit of the intersection. Left on red is never legal and marginal on yellow.

> Left on red is never legal and marginal on yellow.

Not parituclarly relevant to the gridlock violation at issue here, but left on red from a one-way street to a one-way street is legal in the same conditions as right-on-red generally in a large majority of US states, a few of which also allow left-on-red from a two-way street to a one-way street. Left-on-red is only completely illegal in a few states.

The law here is clear. There has to be "sufficient space on the other side" before you enter the intersection. There is no need to make the entire maneuver at once.

> You can't legally enter an intersection (advance past the stopping line if present) until you are clear to perform your transit of the intersection.

Yeah, I've seen tickets written for this in Emeryville, CA for cars that entered on green and weren't going to clear the intersection in the next ten minutes. (I don't know if the intersection is still that bad.) Bicycle cop, because there was no way they'd get a car in there.

> Left on red is never legal

I recall that this was legal in Michigan, where I grew up, but it either had to be onto or off of a one way street. I could never remember which, so I didn't try it.

But I did see someone make a left on red on market street in SF from the rightmost lane, which is very illegal.

> I recall that this was legal in Michigan, where I grew up, but it either had to be onto or off of a one way street. I could never remember which, so I didn't try it.

I believe it requires two one-way streets.

Yeah, this covers the case when turning left doesn't cross any oncoming lanes of traffic - so it's basically the same as turning right on red at a normal light.

Although searching for this now, it does appear to also be allowed from a two-way,[0] which sounds vaguely familiar but isn't really the same as turning right on red.

[0] https://www.michigan.gov/msp/services/traffic-safety/traffic...

Yeah, but sometimes the second and usually the third (and fourth - I have seen it!) are not in the intersection.

I'm not sure exactly your point. The Tesla does sometimes require intervention, that's why it's Level 2. But it's still attempting to drive in significantly more complicated situations than this Drive Pilot thing. Does Drive Pilot stop at stoplights or make turns? I don't think so.

Regarding deceptive editing, plenty of people post their Teslas doing squirrely things and them intervening. So it's not like a secret that sometimes you have to intervene.

Look at is happening at this point, another additional failure mode, not seen in the links I posted before

https://youtu.be/WowhH_Xry9s?t=272

See how is confusing the speed limit for cars and trucks, and doing a sudden break. More precisely, at 5:12.

So if you have car driving behind you, not expecting the sudden break, this could cause an accident. They are joking with people lives.

Ethics matter in Software Engineering.

FYI: Not unique to Tesla, I get plenty of sudden slowdowns when riding Cruise.

Attempting and failing is clearly, clearly worse for the general public and in my opinion Tesla should be strictly liable.

It's not necessarily worse, since there is a person driving the car who can prevent the car from behaving badly. What's the safety difference between this and a regular cruise control, which will happily plow you into a wall or car if you don't intervene?

And, empirically, there's no evidence that these cars are less safe when driving this way. Tesla claims people driving with FSD enabled have 4x fewer accidents than those driving without it, and nobody has presented any data that disputes that.

"Tesla Again Paints A Crash Data Story That Misleads Many Readers" - https://www.forbes.com/sites/bradtempleton/2023/04/26/tesla-...

"...Several attempts have been made to reach out to Tesla for comment over the years since these numbers first started coming out, however, Tesla closed its press relations office and no longer responds to press inquiries..."

This critique of their impact report (I was referring to a more recent statement) only goes as far as saying FSD beta is equally safe to humans driving, not worse, which seems perfectly acceptable?

Depends on the average of human driver. Especially if the average includes motorbikes.

Saying fsd on tela has the same statistic than the general driver population prints a grim picture, as it puts it in a strictly worse performance than peers vehicles (SUV or saloons depending on the model)

You do not get to assume safety in a safety-critical system, period. The burden of proof lies with the entity trying to release a dangerous product to prove that it is safe, not demand everybody else to prove that it is unsafe. The entire argument that there is no evidence that it is unsafe and thus it is okay is wrong to its very foundation. It is such a bad argument that a licensed engineer would stand a good chance of literally losing their license if they advocated such a safety-regressive position. Whoever you heard that from who is actively actually promulgating such inane logic is completely unqualified to talk about safety-critical systems engineering and should be completely ignored.

First, thanks for being an asshole. Second, the product has been released, I did not engineer it, and people online are arguing about it. I’m interested in whether the criticism I’m seeing is valid. Is your idea that anyone who claims something is unsafe, for any reason, should be immediately trusted? If not, how do we have such a discussion? How do we weed out poor arguments? I gave a first-principles argument for why the system might be about as safe as a commonly accepted feature, and also mentioned corroborating evidence of it’s safety. You are welcome to disagree with it, but so far, like most people, you’ve just come in hotheaded and not provided any substantive argument.

It is assumed unsafe. It must be proven safe. This can be done via a engineering, statistical, or other appropriate analysis done by entitys with full access to the design specifications, usage information, etc. who are competent to do the analysis and who have no conflict of interest/unbiased. This has not been done for Tesla FSD therefore it must be assumed to be unsafe. As Tesla deliberately misclassifies their Full Self Driving Beta testing program to avoid government reporting requirements and does not release raw data to any unbiased non-government entity or research organization, it is literally impossible for any external analysis to prove that Tesla FSD is safe. The most a third party can do is affirmatively prove that it is unsafe as you need millions of times less information to affirmatively prove it is dangerous relative to existing systems.

I am not joking when I say millions of times less information. The aggregate motor vehicle fatality rate in the US is ~1.3 per 100,000,000 miles. To even have a chance of proving safety you would need to analyze on the order of 1,000,000,000 miles exhaustively. In contrast, to demonstrate it is not safe you would only need a list of like 20 fatalitys over those same billion miles to affirmatively prove it is unsafe. A list of 20 names versus the raw data for 1,000,000,000 miles is at least a factor of millions in information required and qualitatively different to achieve. Proving safety is enormously difficult and basically requires direct access to the raw information. Proving unsafety is extremely easy and can be relatively easily demonstrated by third partys when dealing with safety critical systems.

Safety engineering is not some sort of new idea that we need to reinvent by the seat of our pants from first principles, it is a mature methodology regularly deployed in aerospace, medical, civil, and automotive engineering applications. Assumptions of safety are fundamentally incompatible with safety critical engineering. This is not a fringe opinion, it is literally a core concept at the foundations of safety critical engineering and not understanding that concept demonstrates a fundamental misunderstanding of modern safety. It is like a geologist arguing the Earth is flat; anybody who suggests such a idea is so fundamentally divorced from the ideas of the field that their knowledge not only fails to reach the level of a expert, they fail to reach the level of even a basic practitioner.

I don’t think we are talking about the same thing and I’m not interested in discussing it further with you.

We know Tesla cannot match Mercedes. We don't know whether or not Mercedes can match Tesla. Mercedes isn't reckless enough to let untrained fanboys play with defective software in two-ton vehicles on public roads.

"We know Tesla cannot match Mercedes" - how? You know this?

"reckless" "untrained fanboys" "defective software" - what is this tone? Why is it reckless? Why do the fanboys need training? Why do you think the software is defective? These are significant unjustified claims!

To me, it seems each company has a different strategy for self-driving, which aren't directly comparable. Beta testing with relatively untrained people on public roads seems necessary for either strategy though.

>how? You know this?

California seems to think so.

>Beta testing with relatively untrained people on public roads seems necessary for either strategy though.

Then why is Tesla the only company doing it?

Mercedes has an L3 system shipping today and they didn't see any need to endanger my life to build it.

Mercedes' system does not do most of the things Tesla's does, right? Such as stop at stoplights or make turns, or do anything at all off-highway. It's a significantly different product, and since they didn't try to do many of the things Tesla is trying to do, it's pretty difficult to claim that those things aren't necessary because Mercedes didn't do them, when they haven't even attempted to deliver the same feature.

Anecdotally, in LA, I turn off FSD basically every minute while trying to use it, due to it doing something slightly “inhuman” or not ideal/too to the letter of the law, signaling incorrectly while staying in a lane that curves slightly, etc.

I can’t imagine letting it go for a full hour without causing some road rage or traffic blockage.

To be clear there is a definite driving “culture” in LA that is very permissive and aggressive (out of necessity). FSD doesn’t follow this culture.

Now imagine it trying to manage a Dutch Roundabout - https://youtu.be/41XBzAOmmIU or driving in a city like OPorto - https://youtu.be/VIWikUUl5YQ

Or Naples https://www.youtube.com/watch?v=dCE1aUkeMe0

Well I was trying to still give a chance to FSD, but as you are raising the stakes...Let's agree to have Marrakech as the baseline - https://youtu.be/SsZlduEIyPQ

And we can also test it in Ho Chi Minh. Tesla says it has superhuman abilities... - https://youtu.be/1ZupwFOhjl4

> signaling incorrectly while staying in a lane that curves slightly

This is super annoying, but the most recent FSD update seems to have mostly fixed that.

It's still pretty crappy in a lot of other ways, but at least that improved.

> but the most recent FSD update seems to have mostly fixed that.

I have read this comment before.

I only started noticing it since the Autopilot/FSD fusion update a month or two ago. I haven’t been driving much the past few weeks so maybe it has been fixed.

> mostly fixed that

The problem is that issues may be fixed one week but re-appear later.

Self driving is not an exercise in fixing bugs despite what many people think.

Every time you disengage it invites you to leave immediate voice feedback as to why, and presumably they are using all this feedback in conjunction with camera and data feeds from cars that are opted in (which includes all FSD beta cars I believe).

So, they are getting what they need to make it better.

> Every time you disengage it invites you to leave immediate voice feedback as to why

lmao do people do this for free?

You're still driving, so it doesn't take any time. And not only can it lead to bug fixes, it can be cathartic to complain.

Worse, they actually think it matters and that Tesla looks at the reports or uses them for anything.

The problem with Tesla is the lack of LiDAR not training data.

You need to be able to accurately do bounding box detection in order to determine whether that billboard of a person is real or not or if that dog with a hat should be avoided.

Research has conclusively shown that vision only systems simply can't match LiDAR for this task.

Don't humans drive with vision only?

We have the world's most advanced supercomputer behind them.

And we also can move our eyes around in three dimensions to infer depth.

You are right that one problem is possibly humans don't drive with their heads fixed in place, they move their neck and constantly adjust viewing angles.

w.r.t. the super computer point, AI systems have been able to out perform humans for specialized tasks for a while.

Vision feeding into trained general intelligence.

We usually use our ears, too.

Now, lets see Mercedes ones...

I’m left wondering what this comment proves? The four situations you posted were the most minor of minor annoyances when the car was being too cautious, and you’ve decided they must be cherry picked without any evidence given.

It's shows the system does not have a semantic understanding of what is happening. That is why he stops so far away from the gate, it does not know it is a gate.

This is another one, definitely one of my favorite examples ( at correct time) - https://youtu.be/2u6AgLuwVqI?t=86

But the Mercedes system has a semantic idea of what is happening?

They don't need. They are not trying to achieve Level 5 and are not relying only on cameras.

Notice also on this example:

https://youtu.be/gzcR8RaC3-g?t=397

They have a driver attention detection system, that works even with dark glasses, as the driver in this case. Also even if you do not take over in the 10 seconds, like in the case for example of the driver fainting, or having a heart attack, the car will stop safely on the side of the road, alert emergency services, and unlock the doors for first responders.

Tesla not only isn't putting their money where their mouth is, they have a lond-standing history of using illegal DMCA takedowns to remove these sorts of videos from Twitter, reddit, Facebook, etc.

Incorrect

Why don't you find a video of FSD failing on a 40mph highway with no lane changes for an apples to apples comparison...

Oh wait that is an unreasonably niche scenario to be worthy of consideration in the context of full autonomy.

Will almost accelerating straight into the guard rail at 30 mph do it? - https://youtu.be/sNBEHumIHJI?t=14

https://www.carscoops.com/2022/05/tesla-driver-claims-model-...

That road is not within the operating requirements of the Mercedes L3 which is only allowed on highways where such turns are never found.

Plus it's a U-turn situation on ancient version of FSD Beta from two years ago (see the dotted lines in the display, or the YouTube upload date for that matter) that did not support U-turns. And it's not necessarily clear it wasn't going to brake more abruptly if the driver had not intervened.

2 years ago is ancient for a feature that was supposed to be delivered in 2016?

They are trying to build something that has never been built before, and your criticism is that they are late?

Yes, since they've been selling cars claiming it would be ready for the last 7 years.

Nobody else is doing that, and I resent their beta-testing features with the rest of us there as unwilling participants.

In that 3 year old build, you can see the driver disengage out of pure fear before the turn started. You can also see the route planner plotting the corner maneuver. You can then see the same software completing the corner twice when the driver does the exact same thing but merely refrains from overriding.

"Watch the Self-Driving Struggle" - https://youtu.be/2u6AgLuwVqI?t=88

Very serious limitation of Tesla:

The human has to keep the hands on the wheel at all times and eyes on the road.

Edit: I looked at these FSD videos before and am very sure it will not work that well in an average European city. Unfortunately, because I really would love to buy a real self driving car that legally allows me to watch a movie while driving (highways is enough).

They are all heavily limited.

But FSD isn't even necessary to help, an autonomous acceleration and follow mode for all cars on traffic lights and in traffic jams would already have huge benefits.

It's very different goals, the Tesla approach is more of a hack, which is to release things, without any liability or guarantee "we quickly hacked this together, good luck, if you die or get injured this is your problem!", and Mercedes is delivering a product that only support a few features but do it well, and they put their responsibility on it.

What features does Tesla FSD offer that Mercedes doesn’t?

Liability.

Not sure how that relates to the point OP was making. Unless I misunderstood his point?

> Mercedes is delivering a product that only support a few features but do it well

This naturally implies that Tesla experimented with a broader set of self-driving features in an throw-it-at-the-wall experimental fashion. Which is more of a technical/product management question than a financial/legal/marketing one. Unless OP simply means Mercedes released the same set of features just extremely restricted in what you can do with them?

If so the hacker/experimental vs limited focus on "small set of features" dichotomy is not super relevant. It's just business risk aversion or gov regulatory strategy, not product/technology development strategy.

> "we quickly hacked this together, good luck, if you die or get injured this is your problem!"

Or kill someone who didn't ask for any of this

So, in the absolute best-case scenario cherry-picked from thousands of videos, Tesla's system is capable of going twenty-five miles without making dangerous mistakes.

I agree the comparison is laughable.

Even considering its heavy limitations, the Mercedes system is miles better than the Tesla, because it is actually useful. You can actually turn on the system and do your email, or take a nap. Yes, the 40 mph limitation on highways seems to make it useless, but many highways get congested to the point where you are not going over 40 mph anyways, and those are the times when it is most frustrating to drive.

And regarding the FSD, most youtube videos I have seen run into human intervention within 20 minutes or so. Going 60 minutes without human intervention is certainly possible if you get a bit lucky but that does not mean you should take your eyes off the road for a second. FSD still has to be constantly supervised, which makes it of very limited utility. At this stage it is still a fun experiment at most.

"do your email, or take a nap" <- The definition of an SAE Level 3 system is that you must remain in the seat, prepared to take control immediately when the system requests you to. Taking a nap or otherwise not paying attention is not what such a system supports.

> prepared to take control immediately

I don't think that's correct. I've seen "in a certain amount of time" but you make it sound like it's a safety issue, which it's not, that being a key differentiation from a L2 system. When it can't drive it stops driving and you have to drive. If it's not a safety issue you could conceivably bee sleeping, as long as you can wake up in a timely manner.

The SAE makes it clear that the car is driving at L3, and on that basis you would expect the transition to another driver would be graceful, just like with two human drivers.

That "certain amount of time" is variable, but in practice is on the level of 10-15 seconds for the Mercedes system - at least that's what it was when it was first certified in Germany some time ago. It is designed to let you take your hands off the wheel and look at the scenery, but anything more than that is too much distraction for when it requires you to take over. And it will, because it's really not very capable at all - it's basically an advanced traffic jam autopilot that can follow other cars in a straight well marked line, that's it.

What would be great is the SAE tightened the definitions so there were standards of performance, maybe with different quality sublevels within each.

For person like me that takes very light naps when sitting, 10 seconds is plenty to wake up and take control. Also if I am using a tablet doing email or browsing or even playing games, 10 seconds is plenty of time to put the tablet away and grab the wheel.

>For person like me that takes very light naps when sitting, 10 seconds is plenty to wake up and take control.

I'd argue that isn't remotely enough time to safely take control. You're betting your brain wakes up sufficiently in that time, and if it doesn't the consequences are potentially deadly.

>Also if I am using a tablet doing email or browsing or even playing games, 10 seconds is plenty of time to put the tablet away and grab the wheel.

That's a bit better from an alertness standpoint, but not from a situational awareness one. You're deprived of both context and situational awareness on the road.

Some key details extend far beyond a 10-15 seconds, such as: another car driving erratically, lanes ending, line of sight on pedestrians or cyclists visible prior that are now occluded by traffic. The list goes on.

>>Also if I am using a tablet doing email or browsing or even playing games, 10 seconds is plenty of time

> That's a bit better from an alertness standpoint

How do you know?

Subjective measurement of time is very inaccurate. I have experienced this, when I timed some regular activities I undertook every day, total habit. I would have said they took me ten seconds. When I timed them, it was forty five seconds.

I have no faith in anybodies' perception of how long they can switch context like this unless they have objectively measured it

I have heard that this has been studied, a quick web search turned up just this: https://www.sciencedirect.com/science/article/abs/pii/000169...

>How do you know?

That's a good point. On the face of it, I'd assumed an awake person already engaging their brain would perform better than a person just waking up in that context, but perahps not.

>I have no faith in anybodies' perception of how long they can switch context like this unless they have objectively measured it

Agreed. Moreover, I'd say that the risk for faulty information processing or incorrect preconceptions is higher if the person is distracted but awake.

I'm not worried that it will actually take anyone 45 seconds to put down a tablet when the car is doing the handover.

I guess it is not ok to be asleep as even a loud alarm will take time to wake you up, but not paying attention, like when you are doing your email is fine. In fact, that's the entire point of level 3 autonomy.

You are "on call" rather than "at work", so you must be prepared to act when the car rings. If the car doesn't ring, you are free to do whatever you want as long as you can hear the ring and take back control when it happens.

Afaik the key differentiator of a L3 and a L2 system is that if you don't take control when the system requests you to, the L3 system can safely stop/pull aside while all bets are off with an L2 system.

You remember wrong. That's the difference between L3 and L4.

The L2/L3 difference is that with L2 you have to supervise at all times and you have to decide to take control.

Now try to look up videos of Mercedes’ L3 system in operation and you’ll see how hilarious this claim is. It shuts off immediately without a vehicle to follow in front of you. Good luck taking a nap and typing emails. L3 my ass.

You're not allowed to sleep, you need to take control within 10-15 seconds.

And no, it doesn't turn off if there are no vehicles in front, but it does turn off above 45 MPH. On most highways you'll only be below 45 because of traffic.

Above 45 they have a separate L2 system that requires "normal" attentiveness.

> And regarding the FSD, most youtube videos I have seen run into human intervention within 20 minutes or so. Going 60 minutes without human intervention is certainly possible if you get a bit lucky but that does not mean you should take your eyes off the road for a second. FSD still has to be constantly supervised, which makes it of very limited utility. At this stage it is still a fun experiment at most.

Are there equivalent videos for Mercedes?

Autopilot works fine in highways under 40 miles per hour. You still pay some attention but that’s also true with this, if traffic goes above 40 mph you have to take control in 2 seconds while Autopilot is more gradual

I’m curious what streets are under 40 mph and not “city” or “country” streets?

Congested highways?

Skeptical but genuine question here: if Tesla's Level 3 system is more advanced, why don't they have authorization to release it? And why does the article quote Tesla as saying they only have a level 2 system?

They aren’t pursuing it, and stopped publicly reporting disengagement data. They found out their stans are happy to buy the cars and use FSD without liability protection or safety data, so why bother? Whenever one crashes it’s the driver’s fault by definition.

For comparison, Tesla FSD veering directly at a cyclist, who was saved only by the driver's lightning reflexes: https://www.youtube.com/watch?v=a5wkENwrp_k

For comparison, Tesla FSD cruising right through a crossswalk with pedestrians and the driver is gleeful about it: https://arstechnica.com/cars/2023/05/teslas-full-self-drivin...

For comparison, Tesla FSD veering into oncoming traffic https://www.reddit.com/r/teslamotors/comments/128zncy/fsd_te...

I remember seeing a video where Tesla FSD veered right at an island/telephone pole. Another where it veered at a crowd of pedestrians on the corner of a signalized intersection waiting to cross.

A reminder that the software has been buggy for years and Tesla is somehow being allowed to "beta test" it in public among people who did not consent to the risk.

As a counter point: these Tesla videos are much more common because there are more Teslas driving around with the feature enabled.

We don't know of the Mercedes is as much of a murder machine based on the little material from their FSD system there is.

All self driving car manufacturers have videos of their cars doing the most ridiculous things. It's hard to pick one above the other. At least Mercedes seems to have documented their limitations rather than assumed their system works everywhere, so that's a good sign to me.

Yes, we don’t know if Mercedes is a murder machine.

What we know, is that if Mercedes murders someone, they’ll take liability for it.

Tesla will meanwhile smear you in the media for abuse of the system, that they actively promote as superior.

Because ultimately there hasn’t been a single publicized death or injury resulting from FSD Beta, while it only took a few months for a professional tester of Uber’s SDC effort to kill someone. So maybe it makes sense to test with millions of people for a few minutes a day instead of 8 hours a day with a few hired testers. The media would sure jump on such tesla headlines.

I have a Model 3 in Europe with the so-called FSD, and it’s mostly terrible. I regret paying for it. The car often doesn’t understand even speed limit signs, so it fails at the bare minimum of safety.

Recently I visited LA and a friend gave me a drive in their Model 3. The difference in sensor interpretation quality was massive compared to what I see back home. That gave me hope of my aging Model 3 still seeing improvements… But also a nagging suspicion that Tesla may have spent a lot of manual effort on data labeling and heuristics for LA roads specifically, and the results may not be easily scalable elsewhere.

It will likely take a while because there is a difference. I'm not sure if it's manual effort or just having more data in certain regions.

When I first got into the beta, it wasn't great, but has improved significantly since then.

The difference in FSD behavior is due to the EU’s laws on autonomous vehicles. The EU mandates maximum lateral acceleration, lane change times, and many other details that make for a worse (and less safe) driving experience.

Could you explain why the limits make the experience less safe.

Less lateral acceleration should mean more safety, or I m wrong?

Now why would the EU mandate details that make for a less safe driving? You are implying they are either malicious or stunningly incompetent.

> stunningly incompetent

Not stunningly, just moderately. But generally yes, that's a good description for EU bureaucrats.

>> stunningly incompetent > Not stunningly, just moderately. But generally yes, that's a good description for EU bureaucrats.

I disagree.

Wordy, worthy, self interested, rule fetish... All of that. But not incompetent like this.

Teslas being optimised for L.A. makes more sense in this case.

Unusually it looks to me like conspiracy rather than cock up.

Government stunningly incompetent? Who would have thought..

https://electrek.co/2019/05/17/tesla-nerfs-autopilot-europe-...

> Model 3 in Europe with the so-called FSD

No Teslas in Europe offer the FSD Beta program.

If you have purchased the "FSD capability," your car should be upgraded to include FSD Beta (or "release") at some point in the future. In the meantime, what you have is essentially Enhanced Autopilot with some bonus features like automatic lane changing on freeways.

So it can only be used on highways with a 40mph speed limit??? :D

Can you explain how

  - Must be under 40 mph
  - Only during the daylight and only on certain highways
  - CANNOT be operated on city/country streets

doesn't exclude all roads? Where I live, highways all have a speed limit of 65-75 mph, and all streets with a speed limit of 40 or below are city our county roads. So where can you actually use this?

It does. The system is useless. Not all freeways are even allowed.

In a traffic jam. Which is most of LAs roads at least some of the time.

From what it sounds like, when you are in a traffic jam you use it and then the second that traffic jam ends you immediately have to take control.

Ngl that sounds like it requires more attention than autopilot, if you’re not doing anything and then have to take full control that sounds like it could lead to an absurdly unsafe outcome

Yea you will struggle to find a single video of drive pilot where it’s not driven by a representative from Mercedes.

In those, you can see the system disengage the second it doesn’t have a vehicle to follow at slow speeds right in front of it (like when the vehicle ahead changes lanes).

Useless.

It is sad to see how many people in this thread will assume the absolute worst with respect to Tesla's FSD capabilities and then will accept this Level 3 announcement from Mercedes in the most charitable way possible. It requires little effort to achieve Level 3 in extremely constrained situations. For example, it would be almost trivial for any autonomy company to achieve "Level 3" at 5 miles per hour within a specific geography. 40 MPH with all of the other limitations of this system is simply not very interesting.

It seems many people in this thread are highly motivated by prior bias toward the negative on Tesla; far more than those of us who appreciate what Tesla is doing.

I'm also finding this really sad. Its painfully obvious to anyone who has used either system recently the majority of commenters have no idea what they are talking about either, the amount of nonsense in this thread about both Tesla and Mercedes ADAS implementations. As always with car discussions on this site, people's feelings about brands and their stereotypical owners seem to cloud any reasonable discourse (especially when Tesla or FSD gets mentioned, sigh). That and of course everyone on Hackernews is apparently an automotive industry expert to boot...

Friendly reminder that there are many humans who choose to not operate their vehicle outside of these conditions as well. Not me, but I know people...

> Tesla FSD navigating the complex city streets of LA for 60 minutes with zero human intervention.

17 fatalities, 736 crashes: The shocking toll of Tesla’s Autopilot

https://www.washingtonpost.com/technology/2023/06/10/tesla-a...

You're right, that is shockingly low! 43,000 people die in car crashes in the US every year, and only 17 in Teslas since 2019. That's only 1 in 7500.

These statistics are meaningless.

How many Tesla cars with full self driving are there compared to regular cars?

A Tesla with FSD will only be using FSD a small fraction of the time.

If we had statistics on the number of hours FSD was active compared to the number of hours all other car driving of all cars we might be able to compared these numbers.

Hour wise, I think normal driving is way above 7500:1.

Why are people willing to excuse things because they happen in a car? If an AI power tool were going haywire once in a while and chopping the arms off of bystanders near by, would we find that acceptable because people using non-AI powered tools also chop off their body parts sometimes? 'We can't know if it would have happened if a person were in control' is not an argument that would fly for anything else just as dangerous.

> would we find that acceptable because people using non-AI powered tools also chop off their body parts sometimes?

Yes, if the AI power tools chop off less limbs than the human controlled power tools, I would find that acceptable.

For that to be meaningful, one would have to consider what proportion of miles driven are in teslas, and where.

Just because you crash that doesn't mean you aren't able to navigate without human intervention. Humans are also capable of crashing their car.

He means under their supervisor control. There are "normal" people who do their best to cause a crash.

Without any comparison to humans those numbers are completely meaningless

> Without any comparison to humans those numbers are completely meaningless

I disagree.

What I want to know, is is it safe? Not is it safer than the average human driver, but is it safe in an absolute sense.

When a human driver crashes we can almost always pin the cause down to human error. Errors caused by some human failing, being less than they could be. The promise of machines is that they are consistent. They are never "less than they could be", but are consistently at their best.

Comparing humans and machines is comparing dogs and roses - it is interesting a rose smells better, interesting a dog is more loving (or fierce), but not a valid comparison

Self driving cars stand or fall on their own capabilities.

Sure it's different whether you were hit by a malfunctioning machine that was confused by sunlight or whether you were hit by a driver who wasn't paying attention because they texted. In one instance you have the person genuinely apologize to you while in the autonomous car case, maybe there wasn't a passenger at all.

But on the other hand, if you can identify the most accident prone group of humans, and require them to use AI cars, and with this you could significantly reduce the number of road kills/accidents, wouldn't that be an improvement?

1. It's the adaptive cruise control, not the FSD.

2. What would a good absolute number for that period be? Our threshold can't be zero or we'd have to give up every technology from showers/baths (~60x more deadly) to fresh produce (E. coli, salmonella, listeria).

> 2. What would a good absolute number for that period be? Our threshold can't be zero or we'd have to give up every technology from showers/baths (~60x more deadly) to fresh produce (E. coli, salmonella, listeria).

Zero accidents caused by the failure of the software. (Clearly a good system may have some obscure bug that makes it fail - but failure of a self driving system should cause uproar and consternation - unlike the reaction of Tesla to the deaths caused by their systems, it seems)

If I slip over in the shower it is not because the shower head went rouge and strangled me. If a shower was designed that in the period of one year strangled eleven people (a fair comparison to Tesla's record) imagine that?

Point being your comparisons (food poisoning or household mishaps) are not relevant

- It's not clear why you no longer care that a technology is killing a lot of people if there's a human in the loop that you can blame. If a driver runs over your sister/son/friend, does it matter if you can blame someone?

- About 43,000 American motorists died in 2021. If we have a way to prevent many of their deaths with software, would you not want to unless you could prevent every death?

- Why is a software tool failure different from any other tool failure? Brakes can fail, wheels can fall off) (happened to me once), etc. Listeria in produce is a failure in the production chain. Showers can be made safer. You use your car/shower/spinach and some day, for reasons entirely beyond your control, it might kill you.

- Why isn't the driver to blame in a Tesla? They're supposed to be watching and responsible.

- There's no clear distinction between what's a software failure and what isn't. Collisions have multiple contributing factors. Perfect software will still have collisions. The numbers reported aren't just collisions where the Tesla was at fault. One person was killed by a self-driving vehicle when a person jumped a concrete barrier and ran across the highway at night. The human driver couldn't react in time, and the software didn't see them. Is that a software failure? At what point do we accept that a collision is no longer the car's fault?

Autopilot is just fancy cruise control. How many crashes have there been with cruise control turned on? It's just a completely meaningless thing to even look at. How many crashes would there have been during those miles without autopilot?

Tesla doesn’t sell Autopilot as cruise control though.

Yes they do? Autopilot is standard in all Teslas, which is different from the Tesla FSD (Beta) that Mercedes is being compared to here.

Because even if they did, would still be worst than many others.

"Ford’s BlueCruise Ousts GM’s Super Cruise as CR’s Top-Rated Active Driving Assistance System - Tesla Autopilot falls to midpack as other high-tech systems improve" (Jan 2023) - https://www.consumerreports.org/cars/car-safety/active-drivi...

"...Of the 12 ADA systems we just finished testing, Ford BlueCruise came out on top, followed by Cadillac Super Cruise and Mercedes-Benz Driver Assistance. Tesla, once an innovator in ADA with its Autopilot system, fell from its second-place showing in 2020 to seventh this time around—about the middle of the pack. That’s because Tesla hasn’t changed Autopilot’s basic functionality much since it first came out, instead just adding more features to it, says Fisher...

...“After all this time, Autopilot still doesn’t allow collaborative steering and doesn’t have an effective driver monitoring system. While other automakers have evolved their ACC and LCA systems, Tesla has simply fallen behind.”..."

That's the number of crashes where Autopilot was enabled, not the number where it was at fault. So its a meaningless number.

> This seems like a marketing gimmick by Mercedes at best; the two technologies aren't even in the same league.

You might be right or not but gave zero arguments in your comment. The fact that these safety/regulatory restrictions exist does not mean the system is not more capable.

good god! No autonomous drive should drive through fog or flooded roads or construction zones. I wouldn't trust any system in those conditions regardless of what stans say.

A few weeks ago, i saw a Tesla suddenly go 90 degrees on a small road downtown Atlanta ... It took a while for it to get going again... Of course I don't know if it was FSD but I have never seen a human do something this dumb.

Yeah this isn’t beating Tesla to anything.

If you constrain a system to where it’s effectively useless and declare victory, that’s worse than trying to actually solve the problem and saying you’re not there yet.

It’s lowering the goal precipitously until you can achieve and then pretending you did it. Tesla has flaws, but this is a dumb article.

> Yeah this isn’t beating Tesla to anything.

They beat Tesla to being certified in California as Level 3, with restrictions.

anything that matters (which was implied by the rest of my comment)

They can check a box and make this press release though, I guess that’s worth something.

It means the headline is not wrong.

You're also kind of walking a fine line on the "No True Scotsman" concept...

Mercedes accepts legal liability for collisions if the system is used under these stringent conditions. Tesla doesn't do this.

It also doesn't mean it doesn't have the technical capability to drive through the streets of LA for 60 minutes.

https://www.thatcham.org/mercedes-to-accept-liability-for-ac...

Working in the AV space, it's really frustrating how confidently people who have no idea about what's hard and what isn't go off about Tesla right now.

Mercedes has soundly beaten the last decade of Tesla efforts by reaching L3.

I've personally watched FSD go off the rails and into a crash situation within 60 seconds of being turned on three times this month (I have a friend who loves to try it in San Francisco)

Had it crashed it'd be on my friend, not Tesla. The fact Mercedes is taking responsibility puts it in an entirely different level of effectiveness.

People also don't seem to understand Mercedes has a separate L2 system that works above 45 mph that already scores better than AP by consumer reports

> People also don't seem to understand Mercedes has a separate L2 system that works above 45 mph that already scores better than AP by consumer reports

Isn't Mercedes' L2 system non-upgradable, though? i.e. people with Merc's L2 system today won't be able to upgrade to L3+ later on, they'll need to buy a whole new car - right?

I bring this up only because I bought my X in 2018 with FSD prepaid - it's 5 years old now, and thanks to the gratis retrofitted computer upgrades it has the same level of FSD functionality as a factory-new Model Y purchased today - and I value that level of commitment to their carbuyers - whereas all the other carmakers I've ever dealt-with are the type that gladly charg $lots just for annual updates to the navigation maps SD card.

(I don't want to be seen defending Tesla, let alone Elon Musk here, but Tesla has avoided plenty of the worst practices of traditional automakers (okayokay, excepting the heated-seats thing).

Drive Pilot uses their MB.OS which is updatable so assume it also will be.

Unless you have some source that suggests otherwise.

And it uses LiDAR which in my opinion is mandatory for FSD.

According to Elon HW V2/2.5 was going to be FSD solved in 2016. Now HW V4 will be FSD solved. There've been multiple paid upgrades along that path that depending on exactly how Elon convinced a given buyer to pay for FSD previously.

The fact that you're going to claim that Tesla, the company that famously discounts vehicles wholesale before unannounced upgrades, silently downgraded their CPO program to a used car program, removed ultrasonic sensors and didn't tell buyers if they'd be missing basic features like Summon temporarily (read: 4+ months) until their vehicles were delivered... is showing a commitment to their car buyers... really says something.

People who paid for the FSD package have been given the upgrades for free.

It’s painful to see your matter-of-fact tone regarding things you have absolutely no clue about. You seem to know next to nothing about the auto space in general. Every carmaker raises and lowers their prices according to supply and demand. And no, there is no driver assistance system among any of the big manufacturers that come anywhere near Tesla’s system. And yes, exploring a problem space that hasn’t been explored before will face setbacks and require many iterations. Those iterations are free for the customers, however.

Elon makes a lot of “promises”, you’re right in that. Although, he really doesn’t. He makes educated guesses that you take for promises. Once Tesla’s own website and news outlets deliver those “promises”, should you start taking them for actual ones.

I work on autonomous vehicles, what on earth do I know about LKAS or exploring new problem spaces?

It's not like a basic Corolla has had features like drowsiness detection for close to a decade.

And you seem to have forgotten that in 2016 Elon claimed _all_ Teslas had FSD capable hardware, never contingent on if you had paid for FSD. That means the upgrades to 3 and 4 were paid upgrades for people who thought they were paying for a car that could already do FSD.

Overall you're upset and barking up the wrong tree (hardly even in the right forest)

>I bring this up only because I bought my X in 2018 with FSD prepaid

Mine came as standard equipment.

> I bought my X in 2018 with FSD prepaid - it's 5 years old now, and thanks to the gratis retrofitted computer upgrades it has the same level of FSD functionality as a factory-new Model Y purchased today

i.e. neither your 2018 model X nor a factory Model Y is anywhere close to "Full Self Driving". Remember what Musk promised in 2017: "In November or December of this year, we should be able to go all the way from a parking lot in California to a parking lot in New York with no controls touched in the entire journey."

We're in 2023. You got scammed, and you're still defending the scammer.

This FUD is insane.

You make it sounds like Tesla's with FSD on are crashing all the time. This makes no sense at all, because... they aren't?

FUD is an interesting term. Why do you use it here? Are you implying GP’s message is part of a broader campaign?

Not op but there are several groups that are independent but essentially aligned on anti-Tesla motivation that have reasons to support campaigns against Tesla: TSLAQ, union supporters, oil producers, legacy auto, to name just a few. There are more. And FUD is well known and easy enough that even individuals can practice it, so I wouldn’t read too much into that.

I’m not convinced FUD is a real thing.

Right, I fear there is doubt and uncertainty, I see what you did there.

I can understand anti-Tesla motivation considering that Tesla doesn't act in a responsible way: they call their system "Autopilot" when it really isn't that; and they don't take responsibility in case of a crash. Both of these things should imho be illegal but somehow they get away with it.

There you go. With people believing this stuff of course they face some opposition.

I was thinking hedge funds paying troll-farms to bash anything TSLA related when they're short the stock (yes I actually believe that is a real thing that happens all the time), but sure those too.

I'm not sure what the track record for FSD is so far but Autopilot's isn't great and inspires little confidence: https://www.washingtonpost.com/technology/2023/06/10/tesla-a...

17 fatalities, 736 crashes.

And yet it’s been used on several orders of magnitude more miles than any other offering. Check the numbers for general driver crashes and fatalities.

> And yet it’s been used on several orders of magnitude more miles than any other offering.

Based on what? ACC/LKAS has been a feature since before Tesla even existed as a corporation. You might be confusing their claims about FSD with AP (which they very much would like you to)

Similarly, they love for people to compare AP to all driver crashes and fatalities when AP only works in the situations that are otherwise least dangerous to drivers, in cars that are newer than the average car, driven by a demographic with higher than average safety records, etc... which means AP would need to kill a lot of people to not have better numbers than average

Those are crashes WITH, not crashes BECAUSE.

Without knowing the usual amount of no fault crashes given the number of driving miles the number is meaningless

$TSLA speaking?

I have no financial involvement in Tesla, owning neither shares nor shorts.

And even if I did, my motivation for speaking up does not affect the fact that a crash happening does not mean autopilot was at fault.

If anything, Tesla was found to be deactivating the autopilot mode at the second before a crash [1], embellishing the statistics of miles driven between crashes...

[1] PDF https://static.nhtsa.gov/odi/inv/2022/INOA-EA22002-3184.PDF

Teslas with FSD are disengaging all the time. And even worse, it's not always FSD initiating: often times it's FSD not realizing it's about to do something wrong, the humans take over, and then them brushing it off as "well it's a beta" or "I bet it would have gotten it but I wasn't so sure".

Getting to a point where Mercedes can guarantee their system will correctly initiate disengagements and do so with a 10 seconds of response time while putting the liability on themselves is a massive leap of anything Tesla has ever put out.

Again, this is what happens when you have literally no clue how the space works (and then turn around and accuse people of FUD.)

Yup Mercedes should be applauded for accepting liability.

Tells me what I want to know about which system is safer.

Tesla's with the FSD option (with or without FSD engaged) are crashing less than any other car.

Yet, according to you that is somehow a bad thing (dangerous?) or worse han some other brand. Which again, is factually untrue afaik.

This reasoning still makes no sense at all.

I agree, the Fucked Up Deaths are insane.

For point of comparison, the entire rest of the auto industry, worldwide has fewer advanced driver deaths than Tesla, despite millions more cars on the road with advanced driving functionality.

On an absolute, and relative, basis Tesla advanced driving is the most dangerous advanced driving system in the world.

I'd love to know the numbers on this. How many crashes with advanced drivers enabled are there and how many miles of driving have they done?

They might be mislead by articles like this one: https://www.nytimes.com/2022/06/15/business/self-driving-car...

Where in the first few paragraphs it gives the impression Tesla has 5:1 deaths, but later you learn that they are including simple lane keeping and cruise control in the comparison, and

> None of the cars using the automated systems were involved in fatal accidents

None of the cars using the automated systems were involved in fatal accidents

If you mean Tesla, that is definitely false. Mostly famously, the Bay Area crash where autopilot was conveniently disabled seconds before the car crashed into a barrier, killing the driver. In the aftermath, the driver's family, and other Tesla drivers noted that a recent update to FSD had caused a regression, whereby Teslas would swerve toward the divider instead of away from it.

simple lane keeping and cruise control in the comparison

Some of Tesla AP's worst fatalities are instances where the car was just replicating simple lane keeping functionality and yet the car still crashed into a big rig or emergency vehicle the Tesla didn't see but which any other carmaker's system would have seen.

Notably, in instances where the Tesla was at fault Musk refuses to discuss the accident, even though he otherwise has no issues releasing information about the driver when the driver is at fault. It's telling that Musk rarely (but still) does this, which means that every time he doesn't it's very strong circumstantial evidence that the car was at fault.

At some point you need to stop making excuses for Tesla's atrocious safety practices and abysmal safety record and just except that they're fundamentally dangerous cars.

It's a review of a particular 10-month dataset. I'm making a comment about the framing of the article vs reality which might lead to misleading conclusions. Not claiming that Tesla (or any automated driving system) has never had fatal crashes, which is obviously not the case.

You said that "the entire rest of the auto industry, worldwide has fewer advanced driver deaths than Tesla, despite millions more cars on the road", what are your sources for that? Do you have numbers on non-Tesla L2 cars on the road to compare?

> This FUD is insane.

What are you talking about? The real FUD is Tesla not taking responsibility when "Autopilot" crashes into something.

An there are no official investigations into Tesla crashes either?

https://static.nhtsa.gov/odi/inv/2022/INOA-EA22002-3184.PDF

> go off about Tesla right now.

People don't go off on Tesla because of its lack of technical prowess. People go off on Tesla because their marketing department is a bunch of pathological liars.

It will be poetic justice if the marketing department is the first thing at Tesla to be completely automated.

That is so unkind to the Tesla marketing department...:-) Where do they get their orders from?

https://youtu.be/zhr6fHmCJ6k

https://youtu.be/5cFTlx68itk

Working in the space appears to be misleading you. It is quite obvious that Teslas can already drive themselves - the thread ancestor literally linked a video.

Your comment has complaints about an academic engineering definition/legal standard that, realistically, is probably going to be paid for in lives by delaying the roll-out and development of self driving vehicles. We know from the march of AI in other fields that the potential for cars to drive to superhuman safety standards is just on the cusp of technologically achievable and will likely be achieved with some sort of evolutionary approach of trying a bunch of things and seeing what works.

This is exactly the wrong time to be making automakers liable for experimenting. If anything, we should be lowering the standards automakers are held to to push through the last technical barriers and break into the oasis where less people kill themselves on the road. I've already lost too many tired friends to car accidents; we should be sparing no effort to get humans away from the driver's wheel. In a fair world, people advocating raising standards should be held liable for the lives they are throwing away on net.

Nobody is being forced to drive a car with autopilot if they aren't comfortable. The risks are obvious and understood. Personal liability is fine.

> Nobody is being forced to drive a car with autopilot if they aren't comfortable. The risks are obvious and understood. Personal liability is fine.

What about all the others participants on the road? Should pedestrian and cyclist also take personal responsibility for getting hit by a self driving car?

> Should pedestrian and cyclist also take personal responsibility for getting hit by a self driving car?

No. Personal liability is fine.

Whose? Drivers, cyclists or the person that made the self driving that ran in to the cyclist?

The "driver" should be responsible. They were the one with most control over what brand of car was used, what conditions it was driving in and the choice to let it drive itself. They have the power to choose cars that don't run people over. Hold them liable and the situation will sort itself out as quickly as is efficient.

The situation would be very similar to how it is now, except a lot less people would be dying in car accidents.

Do they use a RealTime OS?

If so, (how) do they run GPU drivers on it?

And if not, how do they guarantee that the system is responsive and safe?

(Of course a third option would be that they don't use an OS at all for the mission critical stuff).

Yowzzaa, but the damm Beta on Tesla can't still think of slowing down while cornering on steep curves, throwing people on to the side and not navigate into a dead end that drops into the ocean. This is not an anecdote, it's my personal experience. No way I'm not gonna trust a system again, that puts human lives (even those who are not driving a Tesla) for a beta test.

Sounds like they don't want people to die from their product.

Tesla FSD is a marketing gimmick

I don't think we're going to agree here.

> This seems like a marketing gimmick by Mercedes at best;

No, this is what a professional roll-out of a feature like this should look like.

> the two technologies aren't even in the same league. > Any comparison is laughable.

Well, at least we agree on these two. But probably not about the direction.

And here is FSD: https://dawnproject.com/drivers-have-reported-thousands-of-c...

By my count, 46 videos of it not staying in lane, 54 videos of it swerving into objects, 23 videos of it ignoring red lights, 22 videos failing at intersections, 12 videos of pulling out into traffic, 15 videos ignoring road signs, 7 videos ignoring pedestrians, 5 videos ignoring speed limits, 23 videos stopping randomly, 5 videos failing in the snow, 3 videos ignoring buses, 5 videos speeding in school zones with children present, and 3 videos ignoring school bus stop signs.

I own a Tesla with the FSD Beta. (Model S Plaid). I don't know why you seem so bitter about Mercedes product. I'm excited about it! The Tesla FSD really isn't useful, except as an enhanced cruise control.

Owning a Tesla with FSD, it no where close to being able to drive autonomous. I have intervened multiple times where it would have caused a serious accident like merging into a concrete barrier or not detecting a car while trying to turn right into a lane. They still can't fix the phantom braking problem has plague my car from the beginning and i have to drive on the freeways with my foot over the accelerator so i can cancel the hard braking.

Tesla FSD is useless though, because you need to constantly monitor the system for it to do something stupid (which it more or less regularly does).

With a level 3 system like the one from mercedes you can check your phone or do anything else and only have to react when it tells you to take control again.

Mercedes plays tic-tac-toe while Tesla plays Chess. It's just not comparable.

Tesla is taking a radically different and open approach to autonomy compared to every other player. That opens them up to intense criticism though as there are thousands of videos online, with successes and failures. When close to a half a million people are using an autonomous system across the full ODD and all of the US and Canada it's going to be very easy to compile a collection of critical failures. But IMO this is the best way to make fast real progress in this space and not just a limited Disney ride.

> Only during the daylight

This makes no sense to me based on the technology used. Cruise actually previously had a permit to ONLY operate at night.

In addition to radar and lidar, they use cameras for something according to the article. That's probably the component that doesn't work as well as night. (If I had to guess what the cameras are for maybe traffic light detection, lane detection, and human review of video after an incident)

"Conditionally automated driving: Mercedes-Benz DRIVE PILOT further expands U.S. availability to the country’s" - https://media.mercedes-benz.com/article/81a29ac5-4d02-4b58-b...

”Mercedes-Benz DRIVE PILOT is the world's only SAE Level 3 system with internationally valid type approval. It builds on a very robust foundation, setting new industry standards. DRIVE PILOT uses a highly sophisticated vehicle architecture based on redundancy with a multitude of sensors enabling comfortable and safe conditionally automated driving. The certification by the authorities in California and in Nevada once again confirms that redundancy is the safe and thus the right approach.”

"...Mercedes-Benz is focusing on SAE Level 3 conditionally automated driving with the ultimate goal of driving at speeds of up to 80 mph (130 km/h) in its final iteration..."

"Mercedes-Benz S-Class DRIVE PILOT Sensors Details" - https://youtu.be/9m-VS55w9HA

>"Mercedes-Benz S-Class DRIVE PILOT Sensors Details" - https://youtu.be/9m-VS55w9HA

- LiDAR

- Radar

- Cameras

- Ultrasonic sensor

- Road moisture sensor

Curious how this collection of sensor compares with Tesla and also with run of the mill cars with ACC / lane keeping.

https://youtu.be/BFdWsJs6z4c

I'm aware of Elon's opinion: "LiDAR is lame. LiDAR is lame. Lame. Losers use LiDAR. LiDAR is expensive."

I can't stomach watching a video that starts out like that.

Do you use LiDAR when you drive?

No, but at the same time Boeings and Airbuses don't flap their wings and birds don't use jet engines to fly.

Just because nature and millions of years of evolution have solved a problem in a way that looks simple on the surface, doesn't mean the same thing can be copied with current tech to solve the same problem.

Elon's "but humans drive with only use two eyes therefore we can do it with cameras" is the most moronic argument ever and it saddens me when I hear others here parrot it.

Our two eyes may be enough to drive, but our eyes can move in their sockets, our heads can move around and parallax for depth perception, our retinas have miles better dynamic range than any commercial sensor that Tesla is using, but most importantly, our brains are orders of magnitude ore intelligent at reasoning with new or unknown situations than current self driving tech which is just basic pattern matching in comparison.

It's moronic because it's motivated reasoning.

Do you use camera when you drive?

I use 2

How much processing power do you use to process the input and control the car? You've got more of it in your head than a whatever Tesla puts in their cars. Your cameras are also better.

It seems a bit suboptimal to place LiDAR in the front and low. More like a lane following kinda setup.

It's also worth pointing out that this is an extremely limited system:

> On suitable freeway sections and where there is high traffic density, DRIVE PILOT can offer to take over the dynamic driving task, up to speeds of 40 mph.

Basically, this will follow the car in front of you in slow freeway traffic. It won't navigate, it won't even change lanes. It won't work (and it's not clear to me how it disengages) when traffic speeds up.

That's not useless. There's a reasonable argument that this is the product the market wants. But the "Germans beat Tesla" framing here is really quite spun. Autonomy and capability aren't the same thing.

You forgot the best "feature": It allows you to take your hands off the wheel and your eyes off the road.

This is significant as it means MB takes on liability and will give you sufficient time (i think 10s in the European version) to take over.

In the world of corporate liability this is huge even if it isn't a technical achievement by itself.

Only people who don’t understand problem domain think it’s not a technical achievement.

I'll give you 2:1 odds that before this reaches market (looks like end of the year for deliveries per the article) Tesla will duplicate the stunt and lift the driver monitoring requirements in similar situations.

(Edit: two replies here playing gotcha games with prior FSD announcements. But I'm serious: Tesla can do what this Mercedes does already, essentially perfectly, and has been for the two years I've operated the vehicle. The cost to them of duplicating this stunt is near zero. I'll bet you anything they do it.)

The Tesla, who claimed they’re full self driving next year, for more than half a decade now?

Cool, let’s wait.

I think it shows how little you understand the problem, to call it a “stunt”.

They claimed in 2018 that in 2019 you will be able to let your Tesla work for you as a Robo-Taxi and make 300.000,- a year. Why would they still wait?

Why did Elon Musk wait so long to say the videos showing "him" saying "We will have full self driving in 201x" are deep fakes?

> I'll give you 2:1 odds that before this reaches market (looks like end of the year for deliveries per the article) Tesla will duplicate the stunt and lift the driver monitoring requirements in similar situations.

I'll put $100 to the charity of your choice.

are u saying current tesla CAN"T do this?

40mph with a lead vehicle in daytime on a highway?

TODAY, my model3 (on FSD beta) can go 40-80mph, without lead vehicle in night anywhere, AND take turns, watch lights, watch for pedestrians, monitor who came first in stop sign...etc etc... TODAY. Not 2 years from now.

where do i signup?

No, I'm saying Tesla won't sell an L3+ product approved for a CA autonomous vehicle deployment permit before Mercedes does.

There are lots of level 2 systems where you can take your hands of the wheel. That not the definition of level 3.

The interesting thing is it's roughly the same system my 2022 Mercedes has, except it requires my hands on the wheel (and does not disengage over 40mph). I mean, this is just adaptive cruise with lane keeping.

Really the big thing here is Mercedes is saying it's good enough (under 40mph) that the driver doesn't have to pay attention at all.

„Beating“ depends on the perspective. German engineering and marketing approaches are most certainly more conservative, which is important in certain technological advancements such as driving assistance / self-driving.

Many people are saying Tesla FSD is far more advanced, and I've seen videos of A Tesla driving around LA for 2 hours completely autonomously so I agree Tesla FSD is the world leader by far and blows what Mercedes has built out of the water.

However Mercedes are taking liability whilst their system is in use which implies Tesla takes no liability whilst their system is in use.

I'm surprised Tesla is scared to take legal responsibility for their system and I am surprised lawmakers are allowing autonomous systems when the manufacturer doesn't believe in it enough to take responsibility whilst it's in use.

How is Tesla getting away with this?

Of course they can beat the competition especially if they do not need to take legal responsibility for any deaths/accidents that occur when it's in use.

If I were Tesla why would I take responsibility for something no one is forcing me to take responsibility for? Lawmakers are responsible for this charade and its embarrassing to me that a company can be so misleading with marketing and get away with it.

Both state and capital can be bad at the same time y’know

I wouldn't do any business with you for example if that is your default level of ethics.

You say that, but it’s how most companies operate unfortunately.

> You say that, but it’s how most companies operate unfortunately.

Thank goodness you are incorrect.

Most people in business, like most people generally, are ethical human beings. Good to deal with, trust worthy and reliable.

That is my experience.

The sharks that rise to the top of billion dollar companies I only see in the news. My impression is that they are generally mentally disturbed (Elon Musk's "jokes" on Twitter about taking Tesla private is a case in point). But I hope that is a selection bias from the news cycle. Still waiting for the evidence....

Maybe "most companies" was the wrong choice of wording. How about "most large companies that actually influence the world?"

>Most people in business, like most people generally, are ethical human beings. Good to deal with, trust worthy and reliable.

It doesn't matter if the people you've anecdotally dealt with are ethical. It doesn't even matter if most of the people "in business" are ethical. The incentive structure to make the right choice for large public companies, or even venture-backed private ones just isn't there.

I wish I was incorrect, but I'm not. If I was, we wouldn't still have companies producing fossil fuels, animal agriculture, or surveillance capitalism.

OK then don't!

That's the magic of the market! People have different values.

It's not about me. This is exactly what Tesla is doing and for some reason no one seems to care and continues to hand Tesla money for subpar quality cars with fake FSD. I don't get it.

Because it means the feature is actually useful.

If the human "driver" is liable for what an autonomous car does, it means you have to watch the car like a hawk. At that point, may as well be driving.

> [...] blows what Mercedes has built out of the water.

Mercedes FSD prototype, 10 years ago: https://youtu.be/G5kJ_8JAp-w

In that video they mention doing localization based upon a prebuilt map of the route by matching images to the model 10 times per second.

That is by definition, not FSD. That is like the system announced today, a limited route autonomy.

For comparison, FSD v3 (they are shipping v4 in every vehicle now) performs localization 2,000 times per second based upon a hybrid model of every road in open street maps and a generalized model of roads. That is why it is FULL. Even if you are on an unmapped brand new road built yesterday, it will know how to drive appropriately.

They aren’t shipping v4 in the model 3, so no on the “every vehicle”

https://www.teslaoracle.com/2023/06/07/tesla-model-3-highlan...

That’s not shipping…

If you buy a model 3 TODAY you get V3

Not on the 3. But they are shipping it with the Model Y.

https://driveteslacanada.ca/news/tesla-now-shipping-model-y-...

So not “every vehicle” then?

Yeah that’s nowhere close. It’s easy to make a prototype that looks good in a marketing video while driving a very tightly mapped route. It’s a whole other thing to let anyone use self driving tech anywhere, especially on routes it has never seen before.

That was teen years ago, remember. All I can say is that these guys are extremely knowledgeable, kind and an absolute joy to work with. Big shout out to Eberhard, Carsten, Christoph, Clemens and Thao, and to the ones not appearing in the video, like Uwe (enjoy your retirement), David and Henning and a lot others from the chair of Christoph Stiller and from Mercedes research.

>Mercedes FSD prototype, 10 years ago:

Mercedes FSD prototype 1986 to 1994 via the 400 Million Euro EU funded Prometheus project https://www.youtube.com/watch?v=I39sxwYKlEE

It's funny that German, and Italian researchers and car makers had the early lead on self driving tech and then lost it by shelving the tech. Oof.

Which reinforces my earlier point I made in another thread here today, that innovation only happens in the EU as long as it's government funded and as soon as the funding stops, work stops and everything gets shelved instead of the private industry picking up the slack, funding it further to commercialize it like in the US. Sad.

“It’s possible that [Germany] threw away its clear vanguard role because research wasn’t consistently continued at the time,” Schmidhuber said. He added that carmakers might have shied away from self-driving technology because it seemed to be in opposition to their marketing, which promoted the idea of a driver in charge of steering a car."

> What do they have now?

> > Mercedes sprinter

I don't know why but that is hysterical.

Ernst Dickmanns. Legend. Sat close to him at a CVPR and he could not resist to rant about "How's that new? We did that in the 90es!" =:-D

>It's funny that German, and Italian researchers and car makers had the early lead on self driving tech and then lost it by shelving the tech. Oof.

Actually a very common occurrance. I don't think FSD on todays level was possible in '94 and the projects failure was inevitable unless it had been continously funded for at least 15 years more.

>innovation only happens in the EU as long as it's government funded and as soon as the funding stops

Seems like a bad example. Funding stopped because the technology didn't work.

The responsibility already isn't on the car occupants, it is on the occupants' insurance carrier. The only way to meaningfully diminish responsibility on the car occupants is to lower insurance premiums.

To that end, Tesla is offering insurance directly to consumers now, offering lower premiums based upon driver safety system utilization. In my case it would cut my insurance rates in half.

> However Mercedes are taking liability whilst their system is in use

Are they? This press release doesn't actually say so. There was an announcement a while back when they deployed this system in Germany, but that's obviously a different legal environment.

FWIW, this is mostly spin anyway. Liability isn't generally something that is granted, it's either there or not. If Tesla FSD causes an accident, they can absolutely be sued for that in the US. And they have been on a handful of occasions (Successfully even, I think? Pretty sure there were settlements in some of the early AP accidents?).

The reason that "Tesla is liable for accidents" doesn't make news is... that there are vanishingly few AP/FSD accidents. The system as deployed is safe, but that doesn't match people's priors so we end up in arguments like this about "accepting liability" instead of "it crashed".

[1] (2) (A) “Autonomous vehicle” means any vehicle equipped with autonomous technology that has been integrated into that vehicle that meets the definition of Level 3, Level 4, or Level 5

[2] (c) The manufacturer has in place and has provided the department with evidence of the manufacturer's ability to respond to a judgment or judgments for damages for personal injury, death, or property damage arising from the operation of autonomous vehicles on public roads

[1] https://leginfo.legislature.ca.gov/faces/codes_displaySectio...

[2] https://www.dmv.ca.gov/portal/file/adopted-regulatory-text-p...

> FWIW, this is mostly spin anyway. Liability isn't generally something that is granted, it's either there or not.

Mercedes is releasing an L3 product in a jurisdiction where operation of L3 products is insured by the manufacturer. That is substantially different than "someone could sue them and maybe maybe maybe win"

> The reason that "Tesla is liable for accidents" doesn't make news is... that there are vanishingly few AP/FSD accidents.

There have been 736 known AP/FSD crashes and 17 deaths. The reason that "Tesla is liable for accidents" doesn't make news is that they aren't legally liable for their level 2 system under existing AV regulation.

https://www.msn.com/en-us/autos/news/theres-been-a-whopping-...

Sorry, where do you get "accept liability" from "ability to respond to a judgement for damages"? That's not a requirement to pay, that just requires that the company have the ability to pay if they are found liable! It's the corporate equivalent of requiring liability insurance.

Show me the contract (or hell, even a press release) where Mercedes acknowledged accepting liability in the US. I really don't think this happened.

There were a ton of articles around 2022-03-20 (e.g., [1]) that had a line like this:

> Once you engage Drive Pilot, you are no longer legally liable for the car's operation until it disengages.

Not quite a press release but given that Mercedes never denied the claims it's pretty close. It'll be interesting to see how this is implemented legally, of course.

[1] https://www.roadandtrack.com/news/a39481699/what-happens-if-...

That was a German product though. The press release about the american rollout is notably missing that language.

Hmm, the latest article [1] specifically about the California authorization says:

> When active, Mercedes takes responsibility for Drive Pilot's actions.

and

> "Mercedes-Benz Drive Pilot is the world's only SAE Level 3 system with internationally valid type approval," Mercedes CTO Markus Schäfer said in a statement.

Not as clear-cut as you'd want it to be but certainly leaning towards the claim. I guess we'll know for sure once the cars actually go on sale in California.

[1] https://www.roadandtrack.com/news/a44139131/mercedes-benz-se...

Why are you quoting "accept liability" as if its something I actually said?

> Show me the contract (or hell, even a press release) where Mercedes acknowledged accepting liability in the US.

I'm not sure why this is the bar to clear, there is no reason Mercedes would want to potentially open themselves to more responsibility than the law requires. That said, it should be obvious to most people that Mercedes is taking some legal exposure when they:

1) call their product SAE L3 when SAE L3 is the legal definition of a vehicle where the operator doesn't have to pay attention under certain circumstances.

2) tell the user they can watch movies while driving! (no US manual available yet but they make a similar statement in the press release) https://moba.i.mercedes-benz.com/baix/cars/223.1_mbux_2021_a...

That very obviously speaks to their level of confidence in their system compared to something like FSD:

> It may do the wrong thing at the worst time, so you must always keep your hands on the wheel and pay extra attention on the road.

some legal analysis of US law on AV liability if you're interested: https://www.nortonrosefulbright.com/en/knowledge/publication...

> Liability isn't generally something that is granted, it's either there or not. If Tesla FSD causes an accident, they can absolutely be sued for that in the US.

That’s not what liability means here. Assuming the story is true, it means Mercedes is responsible for damages caused when the system is engaged and the users know that when they buy the car. Being sued afterwards is not the same thing.

> The reason that "Tesla is liable for accidents" doesn't make news is... that there are vanishingly few AP/FSD accidents.

Or because Tesla actively hides accident data by using suspect methodology and not following regulations about disclosure.

Just a couple of days ago there was a user report of FSD hitting and killing a dog: https://twitter.com/TeslaUberRide/status/1666860361381818384...

Unsurprisingly, it won’t show up in any of the stats Tesla publishes in their two-paragraph “safety report”. That’s because they don’t consider any contact that doesn’t deploy airbags to be an accident. There are plenty of reports like this that are not being counted.

Not to mention, Tesla is openly violating at least CA DMV regulations that require reporting of all disengagements and contact events.

> Not to mention, Tesla is openly violating at least CA DMV regulations that require reporting of all disengagements and contact events.

Tesla letters to CA DMV claim they don't participate because their system isn't a self-driving system. Which is fine, other than they're telling customers it is at the same time.

A word's meaning is a specific legal document can be different than a word's meaning in a product's marketing material.

Can. It can also be the same.

Yes, because an accident is when there are personal injuries, otherwise it’s a collision.

I’m not aware of any definition of accident that says personal injuries have to occur. But sure, you can replace accident here with collision or contact events. Also, airbag deployment doesn’t automatically mean there are injuries either.

Point is these types of events are not being reported by Tesla, while every other company testing self driving technology (specially ones that have CA DMV permit to do so) are reporting them.

Mercedes could also ship their own L2 AutoPilot without having to take legal responsibility. Their customers would love it as much as Tesla drivers love theirs.

The thing is, they just aren’t capable of it.

Or Mercedes has certain reputation to maintain. This is not a dig at Tesla, but Mercedes not shipping L2 AP doesn't imply they aren't capable of it.

Mercedes has been shipping their L2 system since 2013, at least in Europe.

Huh? The L2 system on my 2022 Mercedes works just fine. And I'm sure that's not the first model year where it was present.

I tried to buy the Mercedes EQS with that feature. All the dealers had a $50k markup , you had to wait months for delivery and most of the cool features in their marketing were not available due to parts shortages.

So call me skeptical…

Skeptical? Because it’s so popular that it’s very hard to get?

popular would mean at least 100k-1mil. is this mercedes really popular like this?

Tesla's sell at 100k - 1 mil mark up?

I think they mean 100k to 1 million units sold

Okay so 10%, assuming one million unit roughly, of what the likes of Toyota, VW and others build every single year across their models.

When it's an "extremely sold out" kind of popular, that can definitely be a reason for skepticism. It's easy to sell out if you barely make any, and when it comes to halo products with significant flaws you want it to be sold out.

We’ve actually had nearly self driving vehicles for over a century. They’re called trains. We should invest in some. The navigation problem gets way easier

Driverless trains are actually a pretty new thing, being on tracks is not as set/forget as people assume.

If you mean "transportation where I'm not driving" then sure but taxis also fit that bill.

Sky train in Vancouver bc has been driverless for near 40 years now

Subway trains have been able to be autonomous since at least a decade.

Paris Metro introduced driverless on all their MP 55 train sets in 1967.

Proportionally, they are far, far closer to self driving even with multiple people driving the train. A ratio of just 1:100 drivers:passengers is a vast improvement on personal cars which would probably be what, 1:1? Maybe 1:2 on average?

Paris had it's first fully automated subway line in 1998.

Today there are 2 fully automated ones and one other mixing manual and automated.

The sofware was developped by Matra transport international (now Siemens Mobility).

It can even dynamically add/remove trains in service to best match the demand !

It also apparently uses a 48 bit cpu architecture.

[0] https://fr.m.wikipedia.org/wiki/Syst%C3%A8me_d%27automatisat...

Do you have a train station in front of your house and any target destination you can imagine wanting to go to? Because most people don't.

Maybe if we built more of those, more people would be within walking distance of them.

Trains for local transportation make no sense. Trams for high-throughput and buses for everything else is the only sensible thing for local public transport.

Works well in Japan.

When I was in Japan a decade ago trains were only used between cities. I doubt that's changed.

I can't fathom how you got this impression. It would be like a visitor to the US saying that in their experience, cars are only used for travel between cities.

> Trains for local transportation make no sense.

Plenty of people take trains for local transportation every day.

Trams and metros however are quite reasonable.

I’ve kind of just considered trams, metros, and trains to be the same thing. Their just different scales of wheels following tracks/

> Trains for local transportation make no sense

Depends what you mean by "local".

For the :last mile", well duh.

But for journeys over 20km there is nothing like trains for efficiently moving vast numbers of people.

Yes. And yes.

I live in the Netherlands and we have this. Because it is what people want and need. And because it's been invested in for decades. And because it has a way higher ROI than car infra.

Well the road infra has also been invested in heavily. Dutch highways are the best

Yup. It's not an either/or as many seem to think. Public Transport works best in addition and next to cars, bikes, and such. It's mixing and combining that offers the true value.

> I live in the Netherlands and we have this. Because it is what people want and need. And because it's been invested in for decades. And because it has a way higher ROI than car infra.

That would be those incompetent Eurocrats....

Do you have a self driving car? Because no one does.

Maybe the all the money invested in self driving and EV insentives and insane highways and stroads maybe you could afford decent public transit.

> Do you have a train station in front of your house

There is one 100 metres from my house. I cannot catch trains at it because some reason I do not understand the powers that be will spend zillions on roads and SFA on rail tracks.

I also know that the companies that run the rail system have not updated their logistics as if it is the twenty first century. I cannot pt a box on the railways, and pick it up in another city, faster than driving it be truck no matter how well I plan it.

I can track my goods on road transport close to real time.

Most trains are far from self driving; training to operate one is easily 6-12 months work, depending on the line, with strict licensing as well.

I seem to remember there being a number of drain derailments lately. Hardly a great safety record compared to either Mercedes' or Tesla's systems.

Let’s see the fatalities per billion passenger km:

Rail: 0.09

Car: 2.52

https://international-railway-safety-council.com/safety-stat... (eu data)

How many fatalities have the aforementioned car systems had? Also European trains seem to run much, much safer than American ones (the derailments I was talking about were in the US), so doesn't seem like the best comparison either.

Indeed, it seems the ratio of fatalities for rail/cars is around 1:25 in EU (see above) and 1:10 in USA [0].

Comparing those data, driving is also much safer in the EU: death rate per 100M passenger for cars in 2021: 0.57. Multiply par 10 (100M to 1B) and divide by 1.6 (miles to km) = 3.5625. Rate in EU = 2.52.

Your right to point out those numbers aren’t split by driving assist options. Until that (still imaginary) future with +95% AV it seems hard to beat that 1:10 ratio, even more with vehicles weight steadily increase [1]

[0] https://injuryfacts.nsc.org/home-and-community/safety-topics...

[1] https://www.epa.gov/automotive-trends/explore-automotive-tre... (select « weight » then « cars » for regulatory class and vehicle type)

Clickbait title to say the least !

While Tesla did amazing things for EVs adoption, and self driving (even if one may say the main contribution is hyping it) it’s hard to see them as leaders anymore.

Fact is, Mercedes is taking responsibility for its system while Tesla is not. Tesla claims (and especially it’s fanboys) starts to look like Theranos. Yeah it almost work and it will be a game changer. Yeah well…

I’m curious about those Tesla videos that are everywhere - is there some kind of dataset somewhere with videos of similar situations, annotated with which version it is etc, so one can make a kind of historical evaluation of its progress ?

(Would also like the history of Elon tweets claiming each version fix this or that to go along in the dataset)

Is there a database of videos of Mercedes? I know “they take liability” but if I’m dead, that still sucks.

What has Mercedes ceo commented about it?

I think you’ll find Mercedes just don’t publish anything, where as Tesla pretty much develop in the open. Regardless of the sausage, some people just don’t like knowing how it’s made.

I don't think there are enough users of this Mercedes thing - and even it there were, they are most certainly not the type of people who will post their videos on youtube...

Tesla publishes very limited and extremely biased numbers - they don't publish raw numbers allowing anybody to do the work.

"Tesla pretty much develop in the open" - that's the issue though - should something like car auto-pilot be allowed to be developed in the open ?

My request was not a shot at Tesla, I just think it would be a very interesting to see - say you are able to isolate and filter all situations of some type (say left turn) from all youtube FSD videos - and compare across the different versions in time. I would love that !

Agree, so that’s at least possible because Tesla allow people to share. Literally no other company has anything like that kind of data in the public domain. Except that stupid “collision self reporting” register California keeps for self driving systems.

> I think you’ll find Mercedes just don’t publish anything, where as Tesla pretty much develop in the open. Regardless of the sausage, some people just don’t like knowing how it’s made.

Can you make this statement with a straight face? That a company develops machines that can kill you any second, up in the open, and have untrained people expose general public to those tests?

How would you feel if I’d be developing magical drug, that cures cancer just by drinking it, with no side effects, by just throwing each iteration into your municipal water supply and put it in your kids milk?

But they don’t. People ask to opt in to it. And that is exactly how human trials work in the medical test. So your analogy is apt.

Are you saying Mercedes process is better? Where they haven’t done extensive testing and have a brand new drug that has actually never been properly tested in the wild?

I’m confused what your point is.

Do you know you’re not the only user of the road? And there’s tons of videos, with predictable abuse, where people allow cars to do illegal/dangerous things to see “if it’ll figure it out”?

If Tesla influencer his me with his car, I didn’t consent to that. Or when erratic behavior (like FSD driving against the traffic, which you have tons of videos of) will cause others to crash - they didn’t consent to that.

Tesla has full “test in prod” approach to their FSD. You can see that from erratic regressions and hot fixes.

But we already let learner drivers try things out on the road, and their supervisors can’t even take control. Human drivers crash and make mistakes too… where is the uproar for allowing them to drive?

Where as Tesla only allows fully licensed drivers and until recently only pretty good ones at that.

I think you’re just regurgitating the standard anti-Tesla talking points… and avoided the questions on the fact Mercedes hasn’t tested this in actual streets and has just rolled it out and while it’s admitting liability isn’t that just as bad as what Tesla is doing? Worse in fact as there no bar for getting this feature, and it doesn’t get shut off for misuse (which FSD beta does).

Given all the approaches (with a hope of working) are using AI - how do you get the data to train the AI? This isn’t a Tesla problem, this is an AI training problem. Tesla already makes heavy use of synthetic data, and using real world human drivers to collect training material. How would you do it?

I don't know the US, but in Europe supervisors definitely can take control - at least for acceleration / breaking.

I also think Tesla's approach is the most efficient one, however I do have a problem with their advertising, which is misleading to say the least.

I think we kind of all agree it's a beta version, and they 'develop it in the open'. But it's been at least 5 years they say it's pretty much FSD - their first video where they say 'the driver is there only for legal reasons' was in 2016 !

Same goes for accidents data - they don't publish raw data, that would allow some fair comparisons. Instead they publish bullshit charts which are extremely biased.

When I took my sons out to practice driving I most certainly could not take control.

But no one publishes detailed data. Every time I hear anti Tesla/Elon chatter, Tesla is always being held to a higher standard than other companies.

FSD naming, nvidia call their Chauffeur, which is way more misleading to me. You have Blue Cruise which means nothing. Audi don’t call theirs anything and just talk about the features which I think is your preference. I don’t believe anyone has bought a Tesla not understanding exactly what FSD does and don’t do, are people being misled?

Uh, you are joking here right ?

Tesla uses extremely deceiving marketing (and same goes for investor relations) regarding a lot of their products lately - FSD being one of them. It's called Full Self Driving... Is that misleading ? How do you interpret this name ?

Are people being misled ? I don't know. But, last year California did pass a law explicitly aimed at Tesla bullshit.

Again, I do think, from society perspective, it's rational thing to do - to 'test in the open' - especially that there are not that much accidents (although the numbers are much worse that Tesla would like people to believe - when you compare apples to apples, there are more accidents per mile with Tesla FSD than with human drivers). What bothers me is their aggressive and misleading marketing, and the fact that they get away with something other companies do not.

> But we already let learner drivers try things out on the road, and their supervisors can’t even take control. Human drivers crash and make mistakes too… where is the uproar for allowing them to drive?

Drinking water supply also gets contaminated all the time. So I assume you won’t have any issues if I start pouring my experimental medicine into your municipal water supply, right? How else can I get large scale data to test it and develop?

Remember, cancer kills millions of people per year! I hope you are not against progress to fight it!

> I think you’re just regurgitating the standard anti-Tesla talking points… and avoided the questions on the fact Mercedes hasn’t tested this in actual streets and has just rolled it out and while it’s admitting liability isn’t that just as bad as what Tesla is doing?

What makes you think that Mercedes didn’t test on the streets? Just because there were no army of YouTubers, not trained in any way, with no relation to the company, making videos?

> Given all the approaches (with a hope of working) are using AI - how do you get the data to train the AI? This isn’t a Tesla problem, this is an AI training problem.

Yeah, if you think it’s an AI problem, not how to build and test safety critical systems that’s all I need to know about your understanding of the problem domain.

You are clearly trolling rather than trying to understand different points of view. I know it can be fun to be in the pro or anti Tesla camp and run around responding to anyone pro Tesla with your small set of points. But it didn’t progress anyone’s understanding of things. You don’t learn anything and everyone dismisses your views as zealotry.

Why not try to engage with the points? Has Mercedes tested this? What evidence do you have for that? If I’m wrong on the AI training problem, how would you collect and validate data without real world feedback? (Ie as you put it “testing in prod). Do other companies do this right?

I’m using sarcasm, sure. But do you think it’s unwarranted when I’m against arguments how only person driving the car could be affected by FSD or that self driving is an AI problem?

I assume you either don’t argue in good faith, or have understanding of the problem area from following Musk’s and his influencer’s tweets.

But I agree with you, it’s not productive. I’ll take my “small set of points” about such an irrelevant things like safety engineering (and no, it’s not a just noun, it’s an engineering discipline) and disengage here.

Just to address one thing tho - YouTubers making videos is and Tesla releasing it to more people to recognize deferred revenue is not testing. And you can see from constant regressions that it doesn’t work as testing.

And before you throw “well, that’s just how AI works”, maybe you’ll make a connection on how self driving isn’t an AI problem. It’s just one of the many many tools used to solve it.

Note you still didn't answer or provide any evidence of "Safety Engineering" from Mercedes. Just the usual Tesla trolling.

AI is the approach that Tesla, Waymo, Nvidia, Cruise, Comma, and even Mercedes are taking. So to ask the question one last time (third time lucky) - How is Mercedes validating the data they use for training without "Testing in Prod" as you criticise Tesla for.

Either: They aren't, in which case your whole rant on the lack of "Safety Engineering" at Tesla is moot, because Mercedes is rolling out essentially untested AI systems.

They are validating their data in real life just like Tesla do, in which case your whole rant on the lack of "Safety Engineering" at Tesla is moot, because Mercedes is just as bad and supposedly have "Safety Engineering" in your opinion.

This is why I keep asking, and this is why you keep not answering.

considering the circumstances the system currently supports I can't imagine a scenario that a human driver could have avoided in which the car could cause your death.

I searched and it looks like the Mercedes L3 is available for 7430 EUR on top of 149900 EUR for the base EQS 450+ model (link below, in french).

Anyone looked if there's a cheaper way to get this Mercedes L3?

https://www.lesnumeriques.com/voiture/drive-pilot-mercedes-c...

https://www.lesnumeriques.com/voiture/essai-mercedes-eqs-450...

"Slow and steady" better than "move fast and break things"? :-)

break people

Define better? A Tesla will do exactly the same thing[1] and just nag you every few minutes to tug on the steering wheel (those nags are getting rarer every release as the driver monitoring camera improves, FWIW). The Mercedes driver can have their hands doing other things, but still needs their eyes out of the cockpit because the second traffic speeds up to 40mph they need to take over.

In practice, this is "Level 3", but only within an ephemeral subset of driving conditions that aren't going to be consistent enough for you actually exploit that autonomy.

[1] Actually much more, obviously, but it at least does this.

People forget the status quo is millions of deaths and trillions in property and economic damages. Moving slow condemns millions more to die.

Those millions (especially in developed countries where the per capita number of accidents is highest) won't be able to buy Teslas anyway.

Statically you could decrease the number of deaths by 20-30% by just banning cars older than 10 years. There are some other (cheaper) things that can be done as well.

40k+ people die per year in car accidents in the US alone. I would say "move fast and break less things" is better.

We can get that to 60k!

You know how you could REALLY fix that?

1. Actual car checks everywhere. I've seen some crazy cars in the US that would be impounded ASAP in most of Europe.

2. Add pedestrian safety requirements for new cars, which would practically ban tall cars (trucks and SUVs) except for professional use.

3. Something for smartphones, not really sure about the best attack angle.

This "might not ever happen" self driving shtick is just moving the goalpost into the middle of nowhere.

You're fooling yourself to think that Tesla is the solution. First, most of the fatalities are in cars that cost much much less than Tesla. Their owners will never afford a Tesla. Second, the only realistic way to prevent some of these deaths is to empower public transport, the way it's done in Europe. Third, although that's the least important point, Tesla statistics of miles driven between crashes are biased, because they turn off autopilot before crashes [1].

Please be careful not to conflate the goals of capitalism with the goals of technology and science.

[1] PDF https://static.nhtsa.gov/odi/inv/2022/INOA-EA22002-3184.PDF

The actual title is:

> Germans beat Tesla to autonomous L3 driving in the Golden State

Which mashed a lot more sense since Waymo, Cruise and Zoox all have L4 autonomous cars on the road today in California operating with no human inside at all.

> Waymo, Cruise and Zoox all have L4 autonomous cars on the road today

Wonder why you can’t buy these L4 cars. No Level 4 cars are commercially available in the US. [1]

What you are describing are pilot programs propped up with a huge, expensive, unprofitable infrastructure of remote monitoring, support vehicles, and backup drivers. Waymo’s technology might be promising, but today these companies are bleeding money.

In contrast, Mercedes is selling retail cars with Level 3 features to consumers and taking responsibility for them.

[1] https://www.autopilotreview.com/self-driving-cars-sae-levels...

What exactly do people think it means when Mercedes says they will “take liability”? Decisions about whether to prosecute a driver when something bad happens are local and state level decisions. Mercedes can’t protect you when a local prosecutor decides to say “yeah level 3, whatever, you were behind the wheel.”

I assume they can only take over civil, not criminal liability.

As another user posted at least in Germany the laws were changed to account for this.

That was handled in Germany by changing existing laws. I would imagine equivalent laws were made or are being prepared for California and Nevada.

"Germany will be the world leader in autonomous driving" - https://bmdv.bund.de/SharedDocs/EN/Articles/DG/act-on-autono...

"...The Autonomous Driving Act took effect in Germany on July 28, 2021, introducing significant and comprehensive amendments to the German Road Traffic Law..." - https://www.ippi.org.il/germany-autonomous-driving-act/

Law here (German): https://bmdv.bund.de/SharedDocs/DE/Anlage/Gesetze/Gesetze-19...

Well, they can provide a rigorous criminal defense, perhaps.

But I don't think there's a lot of criminal prosecutions for driving on highways at speeds under 40 mph.

... only when driving under 40 mph on highways. So it is pretty useless.

The main thing is that Mercedes Benz will take liability for anything that goes wrong. But the system as a whole is far less capable than Tesla FSD.

> The main thing is that Mercedes Benz will take liability for anything that goes wrong. But the system as a whole is far less capable than Tesla FSD.

"It's less capable except for the minor fact that if we <<kill>> someone, it's <<our fault, not yours>>".

You know, that minor factor called "life or death".

> ... only when driving under 40 mph on highways. So it is pretty useless.

If Waymo/Tesla/whatever would do this, they'd be the darlings of the internet.

If Mercedes does this, no one thinks: "ah, it's the first iteration, they can you know, <<increase>> the speed later on".

Also known as "draw the rest of the owl"

At least they started with an owl shape. Tesla is starting with a square before "draw the rest of the owl"

Mercedes has drawn a lot more of the owl than anyone else. Accepting liability is huge.

It's putting their money where their mouth is.

Musk is putting his mouth on a lot of things, I guess that makes him a baby?

You aren't going to kill people while going under 40mph on a highway (cars only, no pedestrians) anyway. Under 40mph is like stop and go traffic.

This is a marketing stunt. They've taken a well-solved problem (stop and go) and taken responsibility for it, and getting headlines by screaming something about J3016 level 3.

A highway is any public road per legal definition. https://law.justia.com/codes/california/2022/code-veh/divisi...

40mph is more than enough to kill someone.

“Limited access highway” is a different thing, and that’s the roads this system will engage on. Zero pedestrians, unless you know, a climate protest is shutting it down.

Mercedes' system only works on specific listed limited-access highways, not all public roads.

> If Waymo/Tesla/whatever would do this, they'd be the darlings of the internet.

How exactly do you think Waymo works that they wouldn't take responsibility?

Waymo takes responsibility for all accidents and Tesla is able to drive autonomously at 40km/h on highways, but neither does both. Mercedes does. That's –I believe– the point that parent is making.

Now, I'd argue that Waymo has probably somewhat solved highways already as they're much simpler than city driving, they're just not offering it to the public yet.

Pure speculation but I bet Waymo doesn’t do highways because they can’t safely reconcile the speed limit with the speed of traffic. In CA and AZ (where Waymo is currently operating) typical highway limits are 55 or 65 mph while traffic speeds in light traffic are easily 70+. Speed differential is dangerous; having a Waymo going 55 while everyone else goes 70+ sounds like a terrible idea but they can’t exactly teach the robot to speed the way a human does either. I’d love to be a fly on the wall during that discussion, ethics, human factors, law, you’d have to tackle the whole shebang.

If the local DA decides to prosecute you after your car kills someone, there is nothing Mercedes can do for you. This liability they are assuming must be civil only.

Changing who holds liability does not make the system more or less capable. Think about it. Was Mercedes system any less capable the day before they declared it L3 than the day before? Clearly not. It was the same system.

If you want to convince me that their system is way more capable than Tesla, you would have to show me evidence that Tesla FSD fails in the same limited environment as Mercedes is certified L3 in, and I have not seen any.

The system was of course the same, but we as the public have new information about Mercedes' level of confidence and are updating our views.

If a magician bets you $1,000,000 that the card he is holding is the one you chose, are you not going to change your position on how likely it is that he's right?

> but the system as a whole is far less capable than Tesla FSD, the renowned “exaggerator” told me so

We actually have no idea what Benz et al.’s systems are capable of if they were to deploy them with the same recklessness as Tesla.

This statement makes no logical sense. Their systems are “capable of” what they can be used to achieve in actual operation.

It’s not like it’s this massively sophisticated system for FSD with most of the code commented out.

These are purpose-built and heavily tested to a specification - that’s how engineering works. The Mercedes system specification is extremely limited in practice, but gives them the marketing win of being able to claim they do L3.

Systems that are required to have ~100% performance within an operational design domain generally have more than 0% performance beyond that domain.

For example, Tesla FSD seems like a pretty good highway cruise control, but it can be dangerously deployed in city driving too. Of course this is a little difficult to argue about since AFAIK Tesla has never defined its ODD and their lead autopilot software guy said in his deposition that he didn’t even know what an ODD was?

I suppose one man’s “marketing win” to “claim” L3 is another man’s “actually deploying L3.”

I maintain that the Mercedes system is “capable of” what it can functionally be deployed to achieve by a user. That’s what “capable of” means. If Mercedes wants to try its hand at a universal L2 system, I wish them the best. This system is in truth not capable of L2 or L3 autonomous driving under almost all typical driving conditions. There is a small geofenced and speed constrained window under which the car will lane-keep and TACC without requiring hands-on-wheel. To me, that’s what we call a “gimmick”.

An autonomous system which cannot even change lanes should not - in my opinion - be classified as L3 Autonomy. Not only is it a gimmick, it’s basically a loophole in the spec.

As an aside, since you mentioned it, I think that describing FSD as “pretty good highway cruise control” is gaslighting.

Describing FSD as dangerously deployed in city driving is merely misinformed. In fact FSD (including the human driver and FSD’s comprehensive attention tracking and alerting) is safer than the average driver in city driving.

https://www.tesla.com/impact

I’m not sure what you’re even trying to argue here. You disagree with SAE’s designations? Interesting trivia I guess.

Anyway I’m sure Tesla will say it’s Level 3 whenever it’s Level 3. Would’ve thought that was a while back when they marketed, sold, and rolled out “Full Self Driving” on public roads.

Well Mercedes does supposedly offer a L2 system. Why do you say it doesn't?

Tesla's system can be used to achieve nothing in actual operation. It makes constant mistakes and requires an ever-vigilant human to take over in a fraction of a second when it does. There is no advantage over manual driving, let alone manual driving with modern driver assistance.

Only on CERTAIN highways if it's daytime and it's not raining.

That's all enforced by the government, it's not a limitation of the system. They might be comfortable with liability for more than this but are not allowed to by regulators - which is wise

It literally is a “limitation of the system”.

The implied system is Benz's autodriving. You seem to refer to "society" as a system. Clever, but not really fruitful.

I wasn’t trying to be cheeky. I must have misunderstood your comment?

As I understand it the L3 system is limited to limited access highway roads, < 40mph, and no lane changes. Those are its limitations, and I believe they are self-imposed by Mercedes, not any US regulations.

"Regardless of those capabilities, the California DMV is placing some serious restrictions on the Merc system. It can only operate during the day on certain limited roads, and only at speeds of up to 40 miles per hour. The system will be usable on California highways in the Bay Area, Central Valley, Los Angeles, Sacramento and San Diego, as well as on Interstate 15 between southern California and Nevada, where the autopilot was approved for use in January."

This indicates to me the limitations come from regulators and not the manufacturer but I might be wrong.

People spend years of their life going under 40 on the highway in LA. This seems huge in socal.

HN just like every other forum. every time a Tesla/Elon related thread pop up. there are so many emotional bags come with it. you can't even have a fair discussion without someone's blind hate

Troll article commented on by forum trolls

All these comments are a perfect example and a study of confirmation bias.

Level 2 FULL self driving is way harder of a problem than location limited, condition limited, maneuver limited, speed limited level 3 so to me it is hilarious to call Tesla beat in any sense of the word.

40mph single lane travel during clear skies and daytime down certain highways can be done by any number of naive driver assistants from many manufacturers. Tesla's autopilot from 2015 could certainly accomplish that, let alone the 2023 FSD versions. The only thing Mercedes has accomplished here is ambitious paperwork.

You really think there’s no engineering difference between system that’s designed to actively kill you, and others if you don’t pay attention vs system that’s designed not to do it, and company putting their money behind that statement?

You really think Mercedes legal filings guarantee a safer system? It is more important to focus on the technical capabilities of the systems than who can add more asterisks to their crash statistics.

You are comparing a mature system with billions of miles of testing to a system which you will struggle to keep engaged for 2 mile intervals.

Tesla FSD is not mature, it’s just old

Ok, let’s focus on the technical aspects.

Tesla has no safety life cycle process and as a result FSD is inherently unsafe, from the engineering point of view.

You can try to hand wave over that, but if you are an engineer working on the safety critical systems (it’s an actual engineering field), that’s all you need to know about FSD.

How will you discretely identify every problem to be solved in the domain of live real drivers in real world conditions? Applying a model of an industrial plant where machines interface with machines, and occasionally humans in prescribed and orderly functions, will not get you any closer to safety.

What will get you closer to safety is insanely large corpus of real world data being trained on in a continuous feedback loop.

Ask Boeing how ignoring it works out.

Which Boeing execs ended their career in disgrace after killing hundreds of people in pursuit of lower training costs ?

Dennis Muilenburg.

189 people: https://en.wikipedia.org/wiki/Lion_Air_Flight_610

157 people: https://en.m.wikipedia.org/wiki/Ethiopian_Airlines_Flight_30...

That's 346 people killed.

Here's an article about his being fired in disgrace after this: https://www.nytimes.com/2019/12/23/business/Boeing-ceo-muile...

Wow, totally missed anyone at all was held accountable for that. It’s not much but something at least.

You are not actually focusing on technical aspects. You are focusing on regulatory credentials and bureaucratic process. Focusing on technical aspects would mean looking at actual data from on road performance.

Design process and lifecycle of the product and development is not a technical aspect only at startups building apps to share cat photos.

By the way, Mercedes tested a self-driving van in 1986[1].

[1]: https://www.politico.eu/article/delf-driving-car-born-1986-e...

Seems like a mistaken title. The limitations are so significant, it can't be called autonomous driving.

Comments are surprisingly negative - I would have thought hn would celebrate such an advance

Only Elon is allowed to make advances in the automotive, space, tunneling, and tweeting spaces.

sadly I can only do a +1 once :)

they probably hold tesla stock since its all time high :)

I think with a different headline the comments would have been positive, but this one is misleading at best

Over the last handful of years, it does seem like HN commenters have become much more negative about tech advances.

One good use-case this is traffic jams. Bosch developed the "traffic-jam pilot" a few years ago with help of DARPA grand challenge participants.

I'm not sure which system Mercedes is using here. Could also be NVIDIA hard- & software (Mercedes and NVIDIA teamed up).

A useful public transportation system build out in a America is more ambitions than fully autonomous driving. Human-less driving is more of a technical problem than can probably be overcome with research, 10s or 100s of billions worth of development and municipal cooperation on autonomous driving infrastructure.

Public transport is a cultural problem as well as a technical problem. I live in a city with decent public transportation system options. I use it frequently and I'm grateful for it. But as soon as I go even a few miles outside the city, the public transportation options are slim to non-existent. I wouldn't even pretend for a second it's a practical option. You better have a car.

Limited to 40 miles an hour and only allowed on freeways seem like incompatible constraints...

It’s a car following system for traffic jams on pre-approved limited access highways. That’s why it also requires other cars nearby.

I have been riding around San Francisco in a Waymo all afternoon with no driver.

Yeah, and seems Cruise is also doing well, but their cars are not for sale.

Headlines like this really speed up the decline of media. While it is factual it really exploits what a common person would interpret the headline to mean. Tesla is exploring autonomous driving on most roads and conditions. L3 autonomous driving is not binary, you can get it at certain times, in certain places, under certain weather conditions yet the comparison is anchoring it to a system that isn't aiming to operate under those constraints. So the headline is taking a decent accomplishment for Mercedes and turning into catnip for Tesla haters.

Mercedes achieves Level 3, while putting their money and responsibility, behind what is a life and death scenario. Tesla fans call it "decent". It's worst than a religion..it's a cult.

Address nothing about the misleading headline. Call the person that points it out as part of a cult. Maybe you should take a harder look at yourself.

Mercedes shipped a system that’s legally self driving, under some set of conditions.

Tesla sells system that’s not legally self driving under any set of conditions. You’re always driving, and Tesla will fight you in court if you want them to take liability for any action it took.

There’s nothing misleading about this headline.

All of that is technically correct but isn’t the point the OP was making. These are products with wildly different aspirations and at different stages of achieving those aspirations.

Yes. One product works and other doesn’t.

And one that doesn’t work is better and for sure will improve and work in the future? And one that does work won’t improve?

I don’t know who you’re replying to, because that’s certainly nothing to do with what I said.

It’s bizarre and absurd to claim a system that’s driven billions of miles “doesn’t work”.

I really like how Simondotau put it. These are wildly different systems. Benchmarks can always be rigged and I see this move by Mercedes no differently.

Having failed to compete with Tesla on actual self driving technology, they concocted a minimally useful system that could check a box in their marketing material. Mercedes, here, is an ostrich.

My primary benchmark for how useful the system is would be how many passenger-miles its driven. We’ll be waiting a long time for Mercedes to be bragging about that metric.

Because it’s very very simple. One system is self driving and other isn’t, no matter how you try to spin it.

One is a universal L2 autonomous driving system and one highly limited L3 autonomous driving system. They are both considered forms of autonomous driving.

[1] - https://www.synopsys.com/automotive/autonomous-driving-level...

One legally drives the car for you. Other doesn’t.

Actually they are both legally autonomously driving according to SAE. You should read up on it!

“SAE International is the world’s leading authority in mobility standards development.” Forgive me if I’ll take their internationally recognized standards over your characterization.

Certainly only one of the systems accepts liability for accidents while it’s enabled, but liability for accidents is just one component of autonomous driving, and not the most important one.

The most important one is clearly the percentage availability of the system across varying roads and conditions.

An L3 or L4 system that is geofenced to work on a few miles of select private roads might be a “world first” while simultaneously being of little practical value.

In essence, a pissing contest or vanity metric.

Ok, you’ve convinced me. I’ll go drive in FSD Tesla and be certain that you and Musk will personally guarantee safety of it and pickup the bill if something happens.

Also, please actually go read the standard, and talk about design intent, and which main way to differentiate the systems.

Your sarcastic retort is not consistent with anything anyone is arguing. Nobody has said FSD is ready to be relied upon, least of all Tesla, who still calls it a "beta". This isn't what I said. This isn't what zaroth said. Nobody said it. So why did you say it?

I realise this isn't a simple argument, but let me try again:

DRIVE PILOT is ~100% capable of limited self-driving on ~0.01% of roads (a curated set of slow highways in California, unless rainy, foggy or dark).

FSD Beta is ~90% capable of full self-driving on ~99% of roads.

Different ambitions at different levels of completion. The comparisons you've been trying to make are, I'm sorry to say, fanboy-esque. "N64 Mario would definitely beat 16-bit Sonic in a game of tennis."

I’m sarcastic as it’s hard to assume people argue here in either good faith or have any idea about the problem.

Being capable of 90% self driving means it’s not self driving, period. And going to ~100% is extremely, extremely difficult. Anyone comparing those numbers to show Tesla’s superiority (or even being in the same league) just doesn’t understand anything about safety critical systems.

What’s better - Airbus 320, which has limited range, or a hypothetical plane, that can fly around the world, and to Mars, but just 90% of the time?

And arguments here are especially funny, as Tesla has been openly lying about their abilities and progress for almost 7 years now.

And talking about ambitions - design intent defines your ambitions and capabilities of self driving systems on the SAE scale. Tesla openly claims to CA DMV, that they’re developing level 2 driver assist system, not a self driving system.

What Tesla has deployed and continues to develop is absolutely an L2 DAS in its current incarnation.

One day, if they are successful, you will see them rollout L3 operation. I would expect this to happen on highways first, and then off-highway. It would be a truly crowning achievement of software engineering and AI, a moonshot realized. I hope they get there.

I think their approach is sound and has proven merit. In the worst case they will have an L2 system that makes commutes more relaxing, and saves lives both inside and outside the vehicle.

But they are uniquely positioned with the massive amount of operational data they collect to solve this problem in a generic / worldwide fashion, and their iterative approach has put them years ahead of the competition.

Their solution is truly scalable, in that it doesn’t require geofencing nor extensive and constant local mapping of the world state in order for the vehicle to safely navigate through it.

Now, Elon openly and readily admits that the same optimism that leads him to try to tackle these moonshots in the first place is what also leads him to over-estimate how quickly they will surmount the challenges they face.

But, he has an extraordinary amount of capital, and an equal level of determination to reach his goals, and frankly no one else on the planet has a better track record for problem solving at this level.

> it’s hard to assume people argue here in either good faith or have any idea about the problem.

Funny, I had the same feeling.

I'm not a fan of tesla, but not a fan of self driving over all. This headline just tells me that benz is the first to be selling legal automatic driving cars in California, something I assumed Tesla was doing. I found it enlightening.

Non technical people would probably assume that Tesla already has this, since we hear about it all the time, and that this is Benz getting up to speed - while it's not the case at all and Benz is beating the largest competitor in the market.

I think this is a fair title.

I see this as a win for everyone. Yes, Mercedes has been faster to get a legal approval but the fact that they got it is a huge win for the industry. Same as the rate of electrification, I think Tesla will not be hurt as they position themselves as enablers. Smart marketing move by the way.

And as « there is no such thing as bad publicity », it keeps people talking about the concept of autonomous vehicles. Which seems to work like a charm according to the number of comments I see for this post !

L3 is profoundly unwise. L3 can disengage at any moment. The attention required of the human in order to drive safely exceeds that required in L2, and is much more difficult to maintain.

It will give a 10 second warning - https://youtu.be/1gjweWq8qAc

I have verrrry limited connectivity RN so can’t watch video. If they really do give 10 seconds, then does the system remain engaged for the “dog/kid darting out between parked cars” scenario? And engage a collision avoidance trajectory planner? 10 seconds is an eternity, even in 35 mph zones. (60 kph)

Yes L3 must handle the dog running into the road situation, this Mercedes will even do evasive driving when possible to avoid a collision.

Mercedes' system will always give the driver a 10 seconds warning before disengaging. Enough to put away your phone and assess the situation.

How is that different from L2? SAE Levels 0-2 demand constant supervision and the ability to take over at any moment. Level 3 is actually better, as it only requires you to be able to take over after a short (e.g. 10 second) notice. Levels 4 and 5 do not require the driver to be able to take over (e.g. drunk, sleeping, no license).

https://www.sae.org/blog/sae-j3016-update

How short is short? How fast is your OODA loop? L3 plays into known human factor weaknesses.

I don't know that it's "unwise"; that depends on failure rate.

But it's absolutely true that the practical difference between "human must supervise" and "human must be ready to take over" is MUCH smaller than people want it to be. Mostly everyone wants to yell about Elon Musk and "2 vs. 3" is ammunition.

L3 requires zero attention, that’s what makes it L3 vs L2. The driver needs to stay sober, awake, and in the drivers seat but are supposed to be able to read a book or something.

A L3 car is required to be able to handle all short term situations and only do a handoff with 5-10 seconds of warning. The idea is it’s ok for the car to come to a complete stop and say I have no idea what to do, but it isn’t ok to simply fail on a freeway at 70 MPH.

Failing safely is a huge difference, as mentioned in the article: if a driver doesn't take over when prompted the car activates the hazard lights and slowly comes to a stop before making an emergency system call to alert first responders to a potential problem.

I love reading the mental gymnastics the anti-Elon folks in here are doing to actually believe Mercedes has beat Tesla in autonomous driving.

because is not like that?

Anyone who is knowledgeable about the subject knows it is not. They’re not even in the same league, it’s really laughable.

Than why is Mercedes able to sell me a system where I can legally watch youtube, etc. while in traffic on a highway without being liable for anything happening and at least 10s until I have to take over control of the car while Tesla can't?

Because it's a party trick that requires a vehicle in front for tracking and can only be used under 40 mph on the highway. It's a joke.

I can do the same thing. Why can Tesla offer FSD beta which can drive around on any roads while Mercedes can't?

It is not a party trick. It literally allows you to do anything while stuck in traffic, which is definitely not a rare occurrence. Meanwhile Tesla's FSD beta doesn't allow me anything I can't do with another car, as I still need to pay the same amount of attention and hold the same liability. Mercedes could offer a similar service, but won't as their assistants already offer a similar experience without the dangers of Tesla's FSD.

You've got that exactly backwards buddy. Tesla could easily offer the same thing, but why would they waste time with that BS? Mercedes 100% cannot offer full self driving on any street. What world are you living in?

Does anybody know how it compares with comma.ai? I've been using it for years and seldom use the steering wheel on the highway.

I think it's a shame that Cruise https://getcruise.com/ isn't mentioned thus far.

They've been fully autonomous in San Francisco for something like a year, and are piloting in Austin and Phoenix(?) ...

No need to own a car if it can be doing dozens of trips instead of paying for parking.

I see. This car is good enough for a morning commute in Bay Area traffic. The question is the disengagement mode. 40 mph on the highway means it only makes sense in traffic, but when the conditions are no longer met how does it decide to hand over control?

If it can reliably wake me up or pull over at the off ramp then it's good enough.

It can try to wake you up, but you only have 10 seconds. You should probably only use it for reading books/using computers and that sort of thing.

I just came here to say that we already have autonomous "someone-else's" vehicles out there.

If you're in a bar and want to go back to your place, you call one if such autonomous vehicles. Or you want to go from home to work without needing to care about parking. They are called taxis.

It doesn’t take a lot to retro-fit roads with autonomous guardrail markers when we repave. It’s going to take a federal-level effort to accomplish this (which should have been apart of pandemic infrastructure stimulus).

I strongly believe this would improve safety for autonomous vehicles by 1000x.

Detecting lane is mostly solved problem. Detecting AND predicting what other things will do on the road (not only cars, but people, animals, trash, etc) is not.

> predicting what other things will do on the road… [like] trash

At least we do have a model for this one (physics). I wonder what additional signals the car would need to accurately simulate trash movement? Wind?

Wind, yes. But even without it - thing get tricky, when you’re traveling at 80mph. You need to detect enough of movements to be able to predict future path. It not easy, even with very advanced sensors.

And you cannot “play it safe and assume it’ll collide with us”. You get phantom breaking, which is very dangerous.

I don’t think you could ever accurately simulate trash movement.

Trash is irregularly shaped, even with the highest resolution cameras, it’d be impossible to know what the far side of the trash is shaped like. Then on top of that there’s no way to predict highly localized gusts of wind.

The best way to deal with trash is to just have an ML model deal with lots of it so it can make predictions about what it is likely to do (which is basically what Tesla is already doing).

Have you seen the video where a tesla tries to hard left into the guard rail on a freeway because of construction/redirect lines going to the left

Lane keeping is not the hard part, that’s been solved for a while. It’s dealing with pedestrians and other drivers doing unpredictable things. Slow moving infrastructure projects will not help that at all.

I used to think that V2X road infrastructure was a major necessary component for self driving. I no longer hold that opinion.

The “self-driving” infrastructure would not be magically any better than the “human-driving” infrastructure, and how to handle conflicting data from the infra in terms of sensor fusion is not at all clear.

In short, it creates as many problems as it solves, and doesn’t really solve the problems that it sets out to in the real-world.

Even simpler things like supplemental signaling that could be used in special circumstances like road work and emergency vehicles, if they are not used properly and consistently in all cases, it doesn’t change the fact that the system needs to correctly handle these cases internally.

I definitely agree state of the art is good enough for 97% of driving scenarios but they’re also not fail-safe and robust. In order for autonomous to be truly acceptable, they need to be 100x more safe. No, nearly flawless. In order to accomplish that you need physical and hardcoded methods as well. In addition, and equally as important, is that road infrastructure gives you improved ability to coordinate cars on the road without relying on individual and disparate compute units.

There’s just no need for this. Tesla can detect even the most poorly defined edges of a dirt road without a problem.

Maybe they should start at detecting children? Luminar Tech can... - https://youtu.be/3mnG_Gbxf_w

Does Mercedes accept liability for any crashes that occur while its system is in control?

> only at speeds of up to 40 miles per hour.

> on Interstate 15

These two things are not compatible.

such clickbait

Ugh, clickbait. Dont bother.

Sigh.

> though the driver must remain behind the wheel to take over when prompted

That is a modal interface and the system may also decides to switch modes at will. Anytime a human is expected to wakeup from a mode and takeover from an automated system on short notice, we have failure modes that are unique compared to a modeless system (including Full Human Drive )

These articles are idiotic clickbait that try to manufacture some kind of drama by throwing the word Tesla in the title.

Tesla is smoking the competition in AVs. Go watch any of WholeMarsBlog's videos on YouTube, of FSD Beta perfectly navigating rush hour San Francisco streets, or driving him from SF to LA without intervention. This is already happening regularly, and has been for months.

Reply on news.ycombinator.com