Waymo CEO dismisses Tesla self-driving plan: “This is not how it works” • Hacker News Board

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#531

Post by ehsankia » Mon, Jan 25, 2021, 6:21 PM UTC

xedeon wrote: post #372

sidibe wrote: post #364

Earlier quoted context omitted.

If we're cherry-picking examples this comparison of 2009 Waymo vs 2020 FSD Beta is pretty funny https://twitter.com/Tweetermeyer/status/1324122869702389761

Those are not even remotely the same... That Waymo video has been pre-mapped. To infer that Waymo in 2012-2009 can go from point A to B anywhere in the U.S. like FSD Beta is patently false. Ever wonder why Waymo only have their fleet in a geofenced area of AZ? https://www.reddit.com/r/waymo/comments/c24bcu/i_got_finally...

There's a difference between can't and won't. Yes their operation in Phoenix is heavily mapped and fenced, that doesn't mean that their technology can't and hasn't driven in more improvised environment before, and I highly doubt the lumbar street example was mapped.

Videos recorded by the company will always be cherry picked by definition, since they would only post videos that show the technology in its optimal working state. And there's always a driver behind the wheel checking everything goes right.

The reason Waymo is operating the way it does is they realize that they need to be 120% cautious, as ONE accident is all it takes to throw all their work a decade behind, jst like one accident is all it took to completely kill Uber's self driving effort. So again, it's not that they need a fully mapped geofenced area, it's that they choose to use that for now.

Hell, Tesla themselves posted their first "A to B" video back in 2016, yet here we are 5 years later and from a high level, what they showed in that cherry picked video they posted isn't much different to an average viewer, but no one would claim that they haven't improved at all in the past 5 years.

This is basic anecdotal data vs real statistics. A video tells you nothing about where the technology is. Unless we can see real open data about miles driven and number of disengagements, there's no way to tell if Tesla is doing well or not. So far Waymo's numbers are an order of magnitude better than any competition.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#532

Post by ehsankia » Mon, Jan 25, 2021, 6:32 PM UTC

sixQuarks wrote: post #481

ehsankia wrote: post #153

Earlier quoted context omitted.

What Tesla has now, Waymo had back in 2012. They literally had a FSD program they were testing with certain employees, and once they found that that people quickly get distracted and stop paying attention, they gave up on that approach and went for the new strategy of jumping straight to level 4 and skipping level 3. So either they've been doing absolutely nothing for the past 8 years until getting level 4 driverless…

..and so does this mean you want to bet or not?

The parameters of the bet are not well defined. Both companies probably have "working" L5 technology, it just comes down to how much they're willing to gamble on the reliability and how early they're willing to put it out there. These things will obviously never be 100% reliable, so now the question is how many 9s each company is willing to wait for.

Waymo realizes that one accident is all it takes to completely destroy their operation, just like it killed Uber's self driving division. Tesla on the other hand has shown that they are prepared to be far more reckless and release things early just for headlines. To me it's meaningless who puts it out "first" because that's not a metric of how good the technology actually is. I'm interested in miles driven and disengagement statistics, which Tesla hasn't really been openly putting out like Waymo.

All I know is that right now, Waymo is the only company with an actual publicly available L4 driving car out in the streets, that is fully driverless with no one behind the steering wheel. And they have been for over a year. Meanwhile Tesla's latest private beta still requires someone behind the wheel and paying attention, and therefore is only L3.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#533

Post by xedeon » Mon, Jan 25, 2021, 6:36 PM UTC

ehsankia wrote: post #531

xedeon wrote: post #372

Earlier quoted context omitted.

Those are not even remotely the same... That Waymo video has been pre-mapped. To infer that Waymo in 2012-2009 can go from point A to B anywhere in the U.S. like FSD Beta is patently false. Ever wonder why Waymo only have their fleet in a geofenced area of AZ? https://www.reddit.com/r/waymo/comments/c24bcu/i_got_finally...

There's a difference between can't and won't. Yes their operation in Phoenix is heavily mapped and fenced, that doesn't mean that their technology can't and hasn't driven in more improvised environment before, and I highly doubt the lumbar street example was mapped. Videos recorded by the company will always be cherry picked by definition, since they would only post videos that show the technology in its optimal work…

> Videos recorded by the company will always be cherry picked by definition, since they would only post videos that show the technology in its optimal working state.

Except FSD Beta videos are coming from actual Tesla owners. Like the one I linked above. A pretty glaring difference.

These are just one of many:

https://youtu.be/i8qVLmflQjs

https://youtu.be/sd1NV7lviAo

> A video tells you nothing about where the technology is. Unless we can see real open data about miles driven and number of disengagements, there's no way to tell if Tesla is doing well or not. So far Waymo's numbers are an order of magnitude better than any competition.

You just contradicted yourself. Tesla actually publishes their data here: [1] https://www.tesla.com/VehicleSafetyReport and here [2]. https://electrek.co/2020/04/22/tesla-autopilot-data-3-billio...

Unless of course, you have inside knowledge on Waymo's stats/data. If that's the case. Please feel free to share.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#534

Post by xedeon » Mon, Jan 25, 2021, 6:56 PM UTC

mikejb wrote: post #484

xedeon wrote: post #351

Earlier quoted context omitted.

> Tesla's warnings to put your hands on the wheel are obviously insufficient based on regular videos of drivers asleep at the wheel on the freeway Every Tesla owner will tell you that those videos where people are "falling asleep" are staged. Why? because if you don't acknowledge the warnings, it will turn on the hazards stop and disable autopilot for the rest of the drive. https://www.youtube.com/watch?v=2uw97Gx1lYw

That particular attack had been fixed, but I remember people disabling those warnings with oranges: https://www.youtube.com/watch?v=TYZrehVQouc It's an arms race between Tesla being legally on the safe side, and drivers using the car as Tesla advertises it.

> That particular attack had been fixed

Which one?

> drivers using the car as Tesla advertises it.

I'm confused, How exactly is Tesla advertising the car where you don't have to pay attention? It even explicitly tells you to keep your hands on the wheel and pay attention.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#535

Post by pmorici » Mon, Jan 25, 2021, 7:06 PM UTC

andi999 wrote: post #423

pmorici wrote: post #358

Earlier quoted context omitted.

The hud as you call it is not a safety critical part of the car in Teslas. You can reboot it while driving without effecting the car. The self driving computer is separate and has full redundancy to the point of having two processors running redundant code. There is a reason Tesla’s are consistently rated as the safest cars on the road with low probability of being involved in an accident and lowest probability of in…

2 Processors? What happens with majority vote if they disagree? (or maybe they wanted to avoid a 'minority report' situation :-)). But honestly do you know what they do? Although since it is not flying, probably some red indicator will light up. And maybe a stopping maneuver.

The self driving computer was discussed in some detail at their 2019 Autonomy Day event.

https://youtu.be/Ucp0TTmvqOE?t=4244

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#536

Post by sudosysgen » Mon, Jan 25, 2021, 7:39 PM UTC

pas wrote: post #524

sudosysgen wrote: post #522

Earlier quoted context omitted.

I'm fairly certain L5 automation won't need anywhere near the processing capacity of the visual cortex.

The visual cortex part is what current deep learning stuff handles pretty well. What's not so obvious is how to generate a model that can manage the "usual problems" (unclear road signs, debris on road, pedestrians/cyclists, harm minimization in accidents) without "freaking out" (phantom braking). For example if the car in front did not even slow down while going through some obstacle (the proverbial plastic bag), th…

The visual cortex does a lot more. It can predict future states, it does things like attention, it can classify states that are abnormal and direct attention to it, it can make analogies between different objects and use those to decide what is normal and what isn't. It's a lot more than just object recognition - it can also generate object classes on-the-fly and use them in the future.

We are very, very far from that. And having such a developed visual processing system is a prerequisite for feeding an executive processing system that can do counterfactual reasoning.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#537

Post by mung » Mon, Jan 25, 2021, 7:57 PM UTC

jiofih wrote: post #127

mung wrote: post #85

Earlier quoted context omitted.

There's two levels of risk, Tesla owners taking themselves out, and then there is innocent bystanders being taken out. It's pretty infuriating if you consider that we're all put at risk.

You are at a much higher risk from all kinds of humans, with varying levels of skill, in different states of mind manually driving these vehicles.

You've always been at risk of other road users, but that is well understood when you get in a car and drive somewhere.

What's new, and on top of that risk, and not well understood is the risk due to drivers mis-using a half-baked feature named "autopilot" as if it's actually an autopilot, because Musk insists that it works when it clearly does not always work.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#538

Post by mung » Mon, Jan 25, 2021, 7:58 PM UTC

jiofih wrote: post #441

s17n wrote: post #363

Earlier quoted context omitted.

Humans are much, much better drivers than autopilot.

For now, that’s kind of the goal here...

Road users aren't beta-testers. Especially the ones who didn't sign up for it and buy a Tesla.

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#539

Post by necovek » Mon, Jan 25, 2021, 8:07 PM UTC

Marsymars wrote: post #530

necovek wrote: post #493

Earlier quoted context omitted.

Riding a motorcycle is a tough one for car drivers, and not just because of the issue you mention: bikes can accelerate and brake much more rapidly due to their lower mass, and inattentive drivers can easily be caught by that. Them appearing where it shouldn't be possible for a car to show up also amplifies the issue (you don't need to look over your shoulders in a single lane street, but bikes easily show up there).…

> bikes can accelerate and brake much more rapidly due to their lower mass Cars are typically able to brake faster than motorcycles. One of the reasons why tailgating on a bike is extremely dangerous.

Good point, thanks!

Re: Waymo CEO dismisses Tesla self-driving plan: “This is not how it works”

#540

Post by onaboat » Mon, Jan 25, 2021, 8:19 PM UTC

Tesla has been regressing with so many easy-to-fix bugs like Spotify playlists being cut off at 100 songs (broken for two months now). I don't ever see myself trusting their FSD when their developer and QA teams are causing or missing so many obvious regressions.