Uber’s self-driving car could not detect pedestrians outside of a crosswalk

Re:Uber’s self-driving car could not detect pedestrians outs... by mcphage on Wednesday November 06, 2019 @02:29PM (#21466131)

> If they added the suppression because there were too many false positives, that would just have resulted in the car honking at apparently arbitrary times. It's just converting the garbage signal from one form into another. It's still too noisy to be reliable.

I love how they went from "our vision system is too unreliable to have warning signals every time it doesn't know what's in front of it" to "okay let's do it anyway but just not have warning signals". Like it didn't make them stop and think "well maybe basing a self-driving car off of this isn't a good idea".

[ Parent | on HN ]

Re:Uber’s self-driving car could not detect pedestrians outs... by EpicEng on Wednesday November 06, 2019 @02:42PM (#21466307)

Oh, that's not being fair; they checked in a fix after all!

  /* issue 84646282 */
  sleep(60 * 1000)

[ Parent | on HN ]

Re:Uber’s self-driving car could not detect pedestrians outs... by pkolaczk on Thursday November 07, 2019 @02:28AM (#21471181)

I don't know why you're downvoted, because I find your comment funny. Reminds me of a similar real fix for a race condition I found somewhere in one of the companies I worked before:

    Thread.sleep(Random.nextInt(1000));

[ Parent | on HN ]