There's a huge amount of work to be done, between releasing promising videos and having customer ready self driving product.

For video, you don't need any 9s of reliability. For a product, you need to have a lot of 9s, and each nine is getting exponentially harder.

How many 9s do you need? I have long said you just need more than humans: one is probably enough for that... Getting to 1 9 is pretty hard though.

I used to think that getting just to human level of 9s is enough, but I've changed my mind recently.

I think we can get to human level, by pure fact that we computers don't get distracted (they only crash ;-) ) and can also pickup signals, we cannot, due to 360° sensors. But computers are still bad at dealing with cases that are obvious to humans, due to lack of context and awareness.

I don't think that public will accept technology like that. I think there'll be a lot of public backslash once videos showing crashes in cases that are extremely obvious to humans. Videos from avoided crashes, that wouldn't be that obvious to humans won't be enough to change the sentiment.

Distribution of headlines that say "car crashed" vs "car avoided a crash" will be the same as "someone does a bad thing" vs "someone does a good thing".

Statistics can tell the difference.

It can, but when was last time that average joe understood and cared about statistic, compared to flashy headline?

The average american has about five nines on a per mile basis in the US. That's in all weather conditions and including all types of accidents / road rage issues / car maintenance levels. Under good conditions, I wouldn't be surprised if another few nines could be added to that.

What are 9s?

I'm not 100% sure, but I think they're referring to the nines of reliability. Something, such as web hosting or server time, is typically deemed as "reliable" when it is 99.999% available as intended. I think what they're trying to say is that a video doesn't need to be reliable at all. As long as you can execute something once or fake it realistically, the video is good enough. For real-world application, though, you need at least 99.9% reliability and, arguably, you'd need 99.999+% for people to accept it.

I might be totally off, though, since I've never heard of the nines being used in the context of autonomy but it does make sense in context of the conversation being had.

I understood 9s in the parent context to mine as referring to accidents per mile. Humans have 5 9s reliability in the sense that the probability of having an accident in any given mile is <0.001%

That's not how it works. Autonomous cars would need to be a couple of orders of magnitudes safer than humans. Human killing human is one thing and machine killing humans is quite another. But more importantly, the blame currently gets distributed over many people's insurance policies. For autonomous cars, all blame would end up at car manufacturer. Imagine you getting blamed for every single accident that happens ever. Even if you have achieved human-level safe driving and chance of an accident is 0.001%, multiplying that with millions of vehicles is a huge number. A dominant autonomous car manufacturer will be blamed for 4 human deaths each hour even if they have achieved human-level safety.

I do agree in principle but there will be nations like china who dont care and just adopt it. After some years the benefits will be obvious and the rest will follow.

Humans have higher expectations for technology. Better than humans is not enough.

>Better than humans is not enough.

By that logic, perfection is not enough because perfect is better than humans.

Saying “x condition is not sufficient” does not mean “x is always insufficient”.

Because every automotive company already has a self driving division and they have realized it's not a realistic dream to work on for at least a decade?

If you are an automotive company you need a self driving plan (not always a division) as risk control. If your competitor comes out with a self driving car that is statically safer than humans you might find your manual driven cars banned from the road. Having some IP in self driving cars is thus a good way to ensure that should that happen you can get good license terms for their tech (play nice with us, or we will sue your for patent violation). Note that good license terms can be millions of dollars.

A plan means either you have a division, or you have an agreement that you can use someone else's if it works out. Either works.

As car makers discover self driving cars are hard they worry less about the risk and put less into development.

How is it "not realistic to work on"?

Tesla has a self-driving functionality in their cars right now on the market, and a quick youtube research will show you that it blows the competition right off, they are better by an order of a magnitude.

Now, it is nowhere near full (level 5 or whatever) self-driving, but noone says that it has to be. The point is that you bring up all other car manufacturers, and they all have obviously inferior products compared to Tesla's (if we are looking at commercially available general personal vechicles). The self-driving as it is right now (even though not level 5) is obviously very useful to many people, it is a good product, the market is right there. And yet these Every Automotive Companies have failed to produce something that Tesla already has on the market. So the problem is not that it is not an "unrealistic dream",(if it exists already), the problem is that other automotive companies are basically behind on this.

My guess is that they are not software companies, they have never had to develop really complicated software, and while having expertise on the mechanical side of things, they are ways behind on the software.

> Tesla has a self-driving functionality in their cars right now on the market

It’s semi-autonomous, requiring constant attention from the driver and operating in an extremely limited set of conditions.

Truly autonomous cars are decades away.

Which is exactly what I wrote if you read the complete message. The point is, how does a new company produce this good product (which many people find very useful, even if it sounds like you don't) - when all the other major manufacturers with their experience and budgets are failing in this regard?

You called it self driving, which it most definitely is not. Unless you’re going to argue that old school cruise control also constituted self driving tech.

And Tesla’s assisted driving tech isn’t that much better than other manufacturers.

Other manufacturers are not failing, you’re just blinded by Tesla’s flashy marketing.

For example, Volvo’s current breed of cars has everything that Tesla has to offer. However, their approach is “no person killed by Volvo by 2020”, so their tech is rightly called assisted driving. Other manufacturers have similar tech and approaches.

A few years back European truck manufacturers even had a competition that involved autonomous driving on general roads across multiple state borders in Europe.

However, Tesla’s marketing keeps insisting that their tech is full autonomous driving, and their lawyers correctly insist that its not.

Now, back to the “not realistic to work on”. Because it’s not realistic. Real autonomous driving is decades away. And “state-of-the-art” is barely semi-autonomous (requires driver’s attention at all times), barely works under ideal conditions (read: sunny California weather with good roads, signs and road markings), can barely handle a very restricted set of road and weather conditions etc. etc.

> For example, Volvo’s current breed of cars has everything that Tesla has to offer.

Have you actually tried it or did you just read it off some anti-tesla rant on the Internet? I desperately want to believe that statement because Teslas are very overpriced for the amount of comfort they provide, but every time I test any other car, the "lane assist" and similar things are nowhere near what Tesla offers. They are all "just as good" on paper, and then you activate lane assist and smallest problem like one small segment of the road line being off makes a Volvo or something else scream in panic and force you to take the wheel while Tesla gracefully doesn't even notices at as a problem and keeps driving. Just because the other manufacturers say on paper "we also have lane assist" does not mean it's the same function.

Look at this latest Audi e-tron. https://www.youtube.com/watch?v=vjsfMmJ4AiM

When you say that other manufacturers are not doing as much marketing for this, you are right. But you got it the other way around: it's not that they are behind on the marketing, and that's why they are failing. They are failing to implement these features that Tesla's already have in great quality on the mass-market, - and that's why they don't have the audacity to market those features, because they know that if they forced people to actually take their "lane assist" seriously, they would be scrutinized by the market even more and all those flaws would be brought to light.

Go actually test a Tesla autopilot, on a long drive, and then go actually test a Volvo, VW, BMW, etc lane assist -and then come back and review your statement. Your mind will be changed.

> barely works under ideal conditions (read: sunny California weather with good roads, signs and road markings),

Watch more of videos on that channel. Many rain situations are fine, and barely some visible road markings is all that is needed. Yes level 5 might be a decade away, but even as it is - that's like 80% of driving time covered. What part of it is "unrealistic" when it already exists?

> Have you actually tried it or did you just read it off some anti-tesla rant on the Internet?

I have.

> They are failing to implement these features that Tesla's already have in great quality on the mass-market

> Go actually test a Tesla autopilot, on a long drive

You keep buying into Tesla's marketing.

Here's a riddle for you: every time a Tesla fails and kills a person, Tesla quickly blames the driver. They say that the driver has to always keep watch and grab the wheel at the first sign of trouble.

Here's an official statement by Tesla in such an accident: "We also ask our customers to exercise safe behavior when using our vehicles, including following the car’s instructions for remaining alert and present when using Autopilot and to be prepared to take control at all times" [1]

Why is that, do you think?

Other companies are just less careless about human lives. They don't pretend it's a full autopilot. They don't pretend that the system can handle 100% of situations. They don't pretend that you can safely fall asleep behind the wheel, and nothing will happen. They call it what it is and introduce the features gradually because they are fully aware of the current capabilities of autonomous systems.

Tesla is fully aware of that as well, but their marketing is willfully engaged in deception tactics.

> What part of it is "unrealistic" when it already exists?

The part where it doesn't exist. In all cases it requires the driver to be fully engaged and alert at all times.

[1] https://abcnews.go.com/US/teslas-autopilot-blamed-driver-acc...

It killed five people, Erlich.

Probably investors just not "feeling" like it. Real advancements in tech need years of R&D funding before they can come to fruition; most of the venture capital industry is hell-bent on short term numbers and PowerPoint hockey stick curves and not hard promising science.