4 ms·

Good summary of technical evaluation from equipments. I think the other thread https://news.ycombinator.com/item?id=16643056 https://news.ycombinator.com/item?i

Good summary of technical evaluation from equipments. I think the other thread https://news.ycombinator.com/item?id=16643056 https://news.ycombinator.com/item?id=16643056 delved pretty deep in various subject scenario too. It would be the best for Uber to release the raw sensor data during the event, and it might be worthwhile for the community to see exactly what happened.

dnomad 9y ago
This blog post, like much of the HN discussion, is full of baseless speculation. There's really nothing here except "LIDAR should've seen it." Duh. The idea that Uber should release more data which won't be analyzed Donnelly but will only fuel even more speculation doesn't make much sense. The NTSB is investigating. They are extremely through. Most importantly once all the data has been carefully analyzed and the failure mode had been understood the NTSB will be able to devise guidelines and protocols to prevent it from happening again. Until then I don't see how all this speculation helps except to fulfill some sort of misguided anti-Uber fantasy.
- buvanshak 9y ago
  > is full of baseless speculation... Can you list all the "baseless speculations" in the discussions/blog post?
- Twisell 9y ago
  If LIDAR can’t do, car shouldn’t speed. Either way is a fail.
  - raverbashing 9y ago
    Car wasn't speeding
    - JustSomeNobody 9y ago
      Yes, it was. Posted speed limits are for optimal conditions. Night driving is not optimal, so if the driver (robot or otherwise) outruns their visibility, then they're speeding.
      - TomMarius 9y ago
        It was going 20 km/h below the speed limit
        umanwizard 9y ago
        Who cares about the posted limit? If you can’t see in front of you well enough to avoid obstacles in time, you are going too fast by definition.
        TomMarius 9y ago
        The word "speeding" refers to breaking the speed limit[1]. I agree with you completely, but you can't blame me for using words like they were supposed to. You might argue that this includes your point, but I've never seen it used like that. [1] - https://www.thefreedictionary.com/speeding https://www.thefreedictionary.com/speeding - "The act or instance of operating a motor vehicle or motorboat faster than allowed by law."
        JustSomeNobody 9y ago
        And by "... allowed by law", they do not mean the posted speed limit. They mean driving too fast for conditions to allow. https://www.nhtsa.gov/risky-driving/speeding https://www.nhtsa.gov/risky-driving/speeding > Speed also affects your safety even when you are driving at the speed limit but too fast for road conditions, such as during bad weather, when a road is under repair, or in an area at night that isn’t well lit.
CobrastanJorji 9y ago
> It would be the best for Uber to release the raw sensor data during the event. I agree with that statement only if we define "best" as "for the common good." The best thing in Uber's own interest is very likely to lock that shit down and hold onto it until a judge makes them give it up. Somebody died here, and it very well may have been Uber's fault. Uber could very well be on the hook for a tremendous amount of punitive damage.
px1999 9y ago
They'll probably fight that as much as they can, because it's not in _their_ best interests. Raw data will likely make what happened appear easy to stop, particularly if they were capturing LIDAR at the time and (assuming no negligence), at best will make Uber look incompetent. People won't accept the nature of how image recognition tends to fail - ie seemingly randomly and catastrophically. Assuming a driver-less future, we'll definitely see more of these random tragedies.

dnomad 9y ago

This blog post, like much of the HN discussion, is full of baseless speculation. There's really nothing here except "LIDAR should've seen it." Duh. The idea that Uber should release more data which won't be analyzed Donnelly but will only fuel even more speculation doesn't make much sense. The NTSB is investigating. They are extremely through. Most importantly once all the data has been carefully analyzed and the failure mode had been understood the NTSB will be able to devise guidelines and protocols to prevent it from happening again. Until then I don't see how all this speculation helps except to fulfill some sort of misguided anti-Uber fantasy.

buvanshak 9y ago
> is full of baseless speculation... Can you list all the "baseless speculations" in the discussions/blog post?
Twisell 9y ago
If LIDAR can’t do, car shouldn’t speed. Either way is a fail.
- raverbashing 9y ago
  Car wasn't speeding
  - JustSomeNobody 9y ago
    Yes, it was. Posted speed limits are for optimal conditions. Night driving is not optimal, so if the driver (robot or otherwise) outruns their visibility, then they're speeding.
    - TomMarius 9y ago
      It was going 20 km/h below the speed limit
      - umanwizard 9y ago
        Who cares about the posted limit? If you can’t see in front of you well enough to avoid obstacles in time, you are going too fast by definition.
        TomMarius 9y ago
        The word "speeding" refers to breaking the speed limit[1]. I agree with you completely, but you can't blame me for using words like they were supposed to. You might argue that this includes your point, but I've never seen it used like that. [1] - https://www.thefreedictionary.com/speeding https://www.thefreedictionary.com/speeding - "The act or instance of operating a motor vehicle or motorboat faster than allowed by law."
        JustSomeNobody 9y ago
        And by "... allowed by law", they do not mean the posted speed limit. They mean driving too fast for conditions to allow. https://www.nhtsa.gov/risky-driving/speeding https://www.nhtsa.gov/risky-driving/speeding > Speed also affects your safety even when you are driving at the speed limit but too fast for road conditions, such as during bad weather, when a road is under repair, or in an area at night that isn’t well lit.

CobrastanJorji 9y ago

> It would be the best for Uber to release the raw sensor data during the event. I agree with that statement only if we define "best" as "for the common good." The best thing in Uber's own interest is very likely to lock that shit down and hold onto it until a judge makes them give it up. Somebody died here, and it very well may have been Uber's fault. Uber could very well be on the hook for a tremendous amount of punitive damage.

px1999 9y ago

They'll probably fight that as much as they can, because it's not in _their_ best interests. Raw data will likely make what happened appear easy to stop, particularly if they were capturing LIDAR at the time and (assuming no negligence), at best will make Uber look incompetent. People won't accept the nature of how image recognition tends to fail - ie seemingly randomly and catastrophically. Assuming a driver-less future, we'll definitely see more of these random tragedies.