Autonomous Car Progress

[mongo]

Just guessing but I think it will probably work similar to how Tesla staggers software updates across the fleet now. Tesla will release a version of FSD to a portion of the fleet, see what the disengagement rate it, make changes, release a new version, rinse and repeat eventually, they accumulate say 1 billion miles of autonomous driving on the same version and the disengagement rate hits the right number and then Tesla can say that their FSD is good enough to remove driver supervision.

I'm talking about after it goes unsupervised. How does anyone handle new releases without reintroducing monitoring/ disengagements?

 

[SandiaGrunt]

in that video is a good sample of deep learning and how it can be used to learn "body language" of cars and pedestrians on the road (at parking lot, stop signs, unmarked lanes, pedestrian crossing etc.). At this moment this is considered the final missing link of self driving implementation. Again no one else with Waymo type Lidar + mapping will be able to solve this.

Everyone — literally everyone — uses deep learning for vision. Zoox, Cruise, Yandex, everyone. Literally everyone.

Waymo can even read police officer hand gestures.

 

[diplomat33]

I'm talking about after it goes unsupervised. How does anyone handle new releases without reintroducing monitoring/ disengagements?

I guess, I just assumed, that the unsupervised stage will be the very last stage of development and therefore Tesla would not make any significant changes to the FSD software at that point. The FSD software would be done and not beta anymore.

 

[DanCar]

That's a weird thing to say. So Tesla is more willing to let people die? That would not be a good thing in my book.

It puts Tesla ahead because they can actually launch a product. 40K people die in the U.S. per year because of traffic deaths. If everyone was using a future Tesla driverless product and only half died, would that be a good thing? That would save 20K lives per year or kill 20K people per year, depending on how you look at it.

 

[mongo]

I guess, I just assumed, that the unsupervised stage will be the very last stage of development and therefore Tesla would not make any significant changes to the FSD software at that point. The FSD software would be done and not beta anymore.

I suppose, but poor Karpathy would be out of a job.
I can't imaging there would not be improvements to made (even more 9's). If it were to have a final release, how do you pick that level at which that happens? (Rhetorical)

 

[diplomat33]

It puts Tesla ahead because they can actually launch a product. 40K people die in the U.S. per year because of traffic deaths. If everyone was using a future Tesla driverless product and only half died, would that be a good thing? That would save 20K lives per year.

Well, hopefully, Tesla's FSD will be a lot better than that. If Tesla's FSD caused 20k deaths per year, that would be absolutely horrific and unacceptable.

 

[DanCar]

Well, hopefully, Tesla's FSD will be a lot better than that. If Tesla's FSD caused 20k deaths per year, that would be absolutely horrific and unacceptable.

It wouldn't be great because it saved 20K people's lives per year? More dead people is better than less dead people? More people injured is better than less people injured?

 

[mongo]

Well, hopefully, Tesla's FSD will be a lot better than that. If Tesla's FSD caused 20k deaths per year, that would be absolutely horrific and unacceptable.

Be clear on what you are comparing. 20k additional deaths is bad, reducing 40k fatalities to 20k is good.

 

[diplomat33]

It wouldn't be great because it saved 20K people's lives per year? More dead people is better than less dead people?

Be clear on what you are comparing. 20k additional deaths is bad, reducing 40k fatalities to 20k is good.

I'm looking at it in absolute terms. I'm looking at the total deaths per year, not the change in deaths. Yes, 20k less deaths than before is good. But 20k total deaths per year is still bad.

 

[Bladerskb]

Um, you just confirmed it

The US has 50 states. 50 -6 = 44 states that Waymo does not operate in, I'd call that multiple (or a majority).
Want to compare how many cities they do vs do not operate in?

Teslas are summoning in as real world a parking environment as it gets. Not specific or test parking lots, any parking lot.
Do Waymo routes currently use parking lots? I only see articles from August saying it's hard.

First of all SS is not an autonomous feature. It doesn't handle all the things a L4 parking system would handle. Secondly Its also about the disengagement rate. Waymo actually has L4 cars without drivers and with drivers handling parking lots inside its 100 SQ mile geofence.

SS is literally the equivalent of Lane keeping and adaptive cruise control. Just because it can be activated anywhere is mean-less.
Do you understand your logic? based on your logic, Waymo is in absolute last place compared to other systems out there. Because their system works everywhere but Waymo's system is geofenced. Therefore Waymo is behind GM, Honda, Hyundai, Nissan, VW, BMW, Mercedes, Audi and every other automaker with an ADAS system

 

[mongo]

I'm looking at it in absolute terms. I'm looking at the total deaths per year, not the change in deaths. Yes, 20k less deaths than before is good. But 20k total deaths per year is still bad.

One would expect each of those events would be analyzed for ways to improve the software (supporting continuous releases).

 

[diplomat33]

You have to decide which is better or worse. You don't get to choose both in this example. You get 20K total deaths which is a savings of 20K lives.

Which is why I fall on the side that saving 20k lives per year is not good enough because it would still leave 20k total deaths per year.

 

[mongo]

First of all SS is not an autonomous feature. It doesn't handle all the things a L4 parking system would handle. Secondly Its also about the disengagement rate. Waymo actually has L4 cars without drivers and with drivers handling parking lots inside its 100 SQ mile geofence.

SS is literally the equivalent of Lane keeping and adaptive cruise control. Just because it can be activated anywhere is mean-less.
Do you understand your logic? based on your logic, Waymo is in absolute last place compared to other systems out there. Because their system works everywhere but Waymo's system is geofenced. Therefore Waymo is behind GM, Honda, Hyundai, Nissan, VW, BMW, Mercedes, Audi, etc

In context: I was addressing the implication of Waymo having less disengagements than Tesla.

Note how Tesla only list the number of accidents, not the number of disengagements.

If one is wanting to compare disengagement counts, would you not agree it is appropriate to also consider the regions of operation?
Kuro did not state whether SS disengagement were in their count or not.

 

[Bladerskb]

In context: I was addressing the implication of Waymo having less disengagements than Tesla.

If one is wanting to compare disengagement counts, would you not agree it is appropriate to also consider the regions of operation?
Kuro did not state whether SS disengagement were in their count or not.

Are you trying to say that Tesla has less disengagement than Waymo?

 

[DanCar]

Which is why I fall on the side that saving 20k lives per year is not good enough because it would still leave 20k total deaths per year.

Elon said it is immoral not to save 20K lives per year.

 

[RubberToe]

We approach the end of 2019 and autonomous vehicles are already here. The technology is now available.

How Reality Has Exceeded Our Expectations About Self-Driving Vehicles

 

[diplomat33]

Elon said it is immoral not to save 20K lives per year.

Maybe it is, but I am saying it is immoral to still let the remaining 20k also die.

 

[DanCar]

Master Plan, Part Deux

... why Tesla is deploying partial autonomy now, rather than waiting until some point in the future. The most important reason is that, when used correctly, it is already significantly safer than a person driving by themselves and it would therefore be morally reprehensible to delay release simply for fear of bad press or some mercantile calculation of legal liability.

 

[mongo]

Are you trying to say that Tesla has less disengagement than Waymo?

Not at all. I'm saying if one were trying to make a point regarding their relative number of disengagements many other factors would need to be concidered. After which it would still be apples and oranges, but at least the same size apples and oranges.

 

[mongo]

Maybe it is, but I am saying it is immoral to still let the remaining 20k also die.

It's not like you are letting them die, it is not intentional.

Is it more or less immoral to let the 40k die so that you are not responsible for the 20k dieing?