Driverless (Self Driving) car pulled over by San Francisco police

[hridge2020]

Source

 

[Bill906]

1. Awesome video. Thanks for sharing
2. This is proof the Level 4 autonomy exists, right? Am I missing something?

 

[firsttruck]

My question is shouldn't the GM Cruise autonomous software be able to tell that the headlights were not working?

Is the GM Cruise autonomous image recognition so bad it can not tell it's own headlight projections aren't there?


---------------------------

GM Cruise autonomous taxi pulled over by police in San Francisco without humans, ‘bolts’ off (U: Cruise responds)
Seth Weintraub
Apr. 10th 2022
https://electrek.co/2022/04/10/gm-c...o-without-humans-bolts-off-u-cruise-responds/

 

[JBee]

Lol on April 1st?

 

[John K]

guy with hat:

1. Accurate description of autonomous vehicles?
2. Pulling answers out of an orifice?

Hmmm…

 

[Crissa]

2. This is proof the Level 4 autonomy exists, right? Am I missing something?

Yes. The different ones are limited by different speed limits and maps, tho. Each one is also building their training data differently. Lowering human markup is keen, as is massive data, which is Tesla's advantage, despite looking clumsy now.

My question is shouldn't the GM Cruise autonomous software be able to tell that the headlights were not working?

Is the GM Cruise autonomous image recognition so bad it can not tell it's own headlight projections aren't there?

It's probably a failure case they had not anticipated more than anything. That's why the testing. Now that it's happened, tho, you can bet others will pay attention (but that it'll happen again to someone who didn't). LED headlights rarely fail, especially both at once.

-Crissa