Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

youtube.com

Post reply on HN
Print view View original

31–40 of 214 posts

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#31

Post by marricks »

dogma1138 wrote: post #12
sam_goody wrote: post #2

tldr: Tesla uses vision alone, and has dropped radar and the other sensor. He makes a very decent argument why. (Surprisingly, he basically ignores night driving.)

Ironically this is what Tesla criticized Mobileye for. I still think that this is far the best demonstration of autonomous driving to date https://youtu.be/A1qNdHPyHu4

Wow that demo has it all.

- car stalled in it's lane - complicated intersections - people exiting cars in it's lane - car going over into it's lane

If I had an hour of driving that I'd be stressed.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#32

Post by Fricken »

RosanaAnaDana wrote: post #20
andyxor wrote: post #8

I like Andrej from his PhD research days and awesome blog posts but this is a series of disasters in the making, that is until FTC steps in after more people die from “self-driving” accidents under interesting and unexpected circumstances. The whole vision vs. LIDAR stuff is a distraction as long as Tesla “AI” doesn’t have common sense. It literally doesn’t know what it’s doing, and the tail of edge cases to "fit" th…

>common sense What a ridiculously useless term.

Try using common sense to figure out what he meant by that.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#33

Post by osipov »

Robotbeat wrote: post #17

He made a very good argument for vision-only, but it seems like training actually uses radar data to help calibrate vision measurements, so it seems to me there’s value in making some vehicles still contain radar (say, one out of 10) even if it’s not used for controlling the vehicle directly at drive time. Also, the sensor resolution issue he mentioned could be addressed by using a higher resolution radar sensor. I f…

> so it seems to me there’s value in making some vehicles still contain radar (say, one out of 10) even if it’s not used for controlling the vehicle directly at drive time.

That's not how neural networks work. You start by training them with a radar, then you deploy them without the radar. Neural nets make the radar irrelevant post-training. This is the entire point of Karpathy's pitch.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#34

Post by bluepanda928752 »

Tesla's decision not to use the LIDAR as a safety feature (i.e. having reliable high-resolution data about things the car can collide with) is so incredibly indefensible, since solving the last 1% of this using only vision likely requires a general artificial intelligence

Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy. Time interval between when Tesla L5 FSD is finally available and when humanity is destroyed by the general AI it runs on will be very awesome and also very short

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#35

Post by SheinhardtWigCo »

bluepanda928752 wrote: post #34

Tesla's decision not to use the LIDAR as a safety feature (i.e. having reliable high-resolution data about things the car can collide with) is so incredibly indefensible, since solving the last 1% of this using only vision likely requires a general artificial intelligence Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy. Time interval between when Tesla L5 FSD is finally availa…

They’re betting that they can use a massive feedback loop to train a set of neural networks to the point where they are as accurate as LiDAR without actually firing any lasers.

Even if you believe this goal is possible to achieve at some point in the future, I think the argument falls apart when you consider that it will take years, probably decades, for a pure vision approach to catch up to where Waymo is today in terms of safety. (They have cameras too.)

That Tesla can’t afford to fit expensive LiDAR sensors to all of the cars it sells is Tesla’s problem. Regulators won’t give a shit that pure vision is “better” in theory. They will simply compare Tesla’s crash rate in autonomous mode with that of Waymo and other AV operators, and act accordingly.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#36

Post by DSingularity »

SheinhardtWigCo wrote: post #35
bluepanda928752 wrote: post #34

Tesla's decision not to use the LIDAR as a safety feature (i.e. having reliable high-resolution data about things the car can collide with) is so incredibly indefensible, since solving the last 1% of this using only vision likely requires a general artificial intelligence Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy. Time interval between when Tesla L5 FSD is finally availa…

They’re betting that they can use a massive feedback loop to train a set of neural networks to the point where they are as accurate as LiDAR without actually firing any lasers. Even if you believe this goal is possible to achieve at some point in the future, I think the argument falls apart when you consider that it will take years, probably decades, for a pure vision approach to catch up to where Waymo is today in t…

Definitely. Especially with car companies like NIO strapping in LIDAR to their upcoming models.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#37

Post by ra7 »

bluepanda928752 wrote: post #34

Tesla's decision not to use the LIDAR as a safety feature (i.e. having reliable high-resolution data about things the car can collide with) is so incredibly indefensible, since solving the last 1% of this using only vision likely requires a general artificial intelligence Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy. Time interval between when Tesla L5 FSD is finally availa…

> Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy.

Tesla is also the only company to claim to target L5 autonomy. Everyone else, including Waymo, is strictly targeting L4 and say L5 autonomy is not possible or realistic. L5 is a pipe dream.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#38

Post by halotrope »

marricks wrote: post #31
dogma1138 wrote: post #12

Earlier quoted context omitted.

Ironically this is what Tesla criticized Mobileye for. I still think that this is far the best demonstration of autonomous driving to date https://youtu.be/A1qNdHPyHu4

Wow that demo has it all. - car stalled in it's lane - complicated intersections - people exiting cars in it's lane - car going over into it's lane If I had an hour of driving that I'd be stressed.

Wait until you see the Jerusalem 40 minute video. Munich traffic is tame in comparison: https://youtu.be/kJD5R_yQ9aw I really don’t understand why mobileye gets so little recognition. They might be quietly winning the self driving race.

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#39

Post by bluepanda928752 »

SheinhardtWigCo wrote: post #35
bluepanda928752 wrote: post #34

Tesla's decision not to use the LIDAR as a safety feature (i.e. having reliable high-resolution data about things the car can collide with) is so incredibly indefensible, since solving the last 1% of this using only vision likely requires a general artificial intelligence Prediction: Tesla will be the last of all major auto manufacturers to get to L5 autonomy. Time interval between when Tesla L5 FSD is finally availa…

They’re betting that they can use a massive feedback loop to train a set of neural networks to the point where they are as accurate as LiDAR without actually firing any lasers. Even if you believe this goal is possible to achieve at some point in the future, I think the argument falls apart when you consider that it will take years, probably decades, for a pure vision approach to catch up to where Waymo is today in t…

I understand why they made the "no-LIDAR" bet early when the LIDARs were completely unpractical for a production consumer car

However, nowadays it starts to look that 100% reliable depth estimation from cameras might actually require a human-level AI to work and also solid-state LIDAR technology is becoming cheap enough and integrateable into normal cars, but Tesla can't really change their stance on this without admitting that FSD options they already sold would not actually become FSD within the lifetimes of these vehicles. I suspect this might also be the reason why Karpathy looks more and more nervous with each new talk

Re: Andrej Karpathy (Tesla): CVPR 2021 Workshop on Autonomous Vehicles [video]

#40

Post by gameswithgo »

RosanaAnaDana wrote: post #20
andyxor wrote: post #8

I like Andrej from his PhD research days and awesome blog posts but this is a series of disasters in the making, that is until FTC steps in after more people die from “self-driving” accidents under interesting and unexpected circumstances. The whole vision vs. LIDAR stuff is a distraction as long as Tesla “AI” doesn’t have common sense. It literally doesn’t know what it’s doing, and the tail of edge cases to "fit" th…

>common sense What a ridiculously useless term.

a better phrase might be “ability to do higher order reasoning to decide what to do in a novel situation”
Post reply on HN