You’re arguing something I didn’t state or anyone has stated. No one said they went all out with their equipment. What I’m flagging is that you see the bad dashcam footage as an intent to mislead or perhaps even intentional doctoring.
That’s not “giving them the benefit of the doubt”. Rather, it’s about not making unsubstantiated assumptions.
I saw the footage as unusually dark and it gave the public an impression that that street is darker than it really is so you have a bunch of people, not necessarily here, running around saying hey you can’t see her until the last 1.5 seconds when headlight beam hits her directly.
Well “it” is the car, which is supposed to see in the dark. The light is well-lit, which is not really obvious in the dashcam, and the human driver in that car should have seen a lot better than that.
I find the video misleading, intentional or otherwise, and I’ve seen a few videos taken of that exact road and spot since the accident posted. The difference between that foot and what Uber has is pretty… large. If you’re saying a Synth says the cheaper cams would show more light and maybe Uber used a more expensive one but you say hey it’s not even part of the sensor so maybe it’s not top… even those two statements don’t mash.
The thing is though… I still think the answers too these questions should have been publicly, openly, and available before the crash. They’re testing in the public.
Nobody has said Uber has doctored anything but I find it very disturbing that the camera footage released is so poor. If they do have better footage or failed to effectively post process it, shame on them for releasing this. If they don’t have better footage then that’s even crazier. I just assumed these vehicles would have sensors and cameras out the wazoo for learning and after fact analysis and they would be high quality.
I am an enormous fan of the advancements being made by the various companies but this just seems crazy to me. Not that a mistake was made, but the effort that went into capturing that mistake seems so low. Again, they might not be sharing all the data with us, but either way it’s not a positive reflection on Uber, as if we needed even more reasons to distrust them.
I think any sort of post-processing, done for whatever reason, would have been a terrible, terrible idea. They would be accused of all kinds of doctoring, no matter which way they went with that. Releasing the best, unadulterated video that they had is the only defensible position, other than not releasing what they had. Not releasing comes with its own problems.
On why they (presumably) don’t have better video, I assume that straight up optical cameras that present a relevant view of all the action are not a part of their sensor suite.
I agree with you that this accident shows some serious fault, of some sort, with the system or that particular vehicle.
I’ll argue this point just a little bit. I want to stipulated that I don’t think that the cam footage released by Uber is representative of what a human might see… but that said:
Some nights ARE darker than others. If it’s low-overcast, you can get reflected light from nearby city lights to provide more ambient light. High overcast doesn’t do that and can prevent the moon (if present) from providing direct lighting. If it’s a clear night, the absence, phase or ascension of the moon makes a radical difference. Humidity can make a big difference in how far light travels and how sharp or deep the shadows appear to the naked eye.
On a suburban city street like the one in question, the TIME of night can make a big difference. Many modern streetlights dim themselves after a certain time to save energy and/or reduce light pollution. Office buildings have fewer lights on, and the overall city-light spillover can be much less at 1 in the morning compared to 7PM (the Uber incident happened around 10PM in the winter). Weeknight vs. weekend night could make a difference in some areas. Other cars on the road can make a very big difference as they contribute their own headlights to the overall lighting situation. Many of the youtube videos shown as counter-evidence seem to be taken fairly early in the evening with lots of pedestrians and other vehicles about.
The presence or absence of streetlights too don’t entirely abrogate the environmental effects – I’m sure everyone has driven on a night where the streetlights don’t seem to be cutting through the overall gloom on a road that’s normally much brighter.
And although I believe that the low-detail cam footage that Uber released is not indicative of human vision, you also have to take the videos posted by people with high-end cameras with a grain of salt – many phones and especially dash-cams have excellent low-light modes, which might actually be seeing more than what an average human could detect.
So it is a conspiracy and that is why we should all be up in arms. You believe that the footage is doctored.
Meanwhile 102 people every day die in traffic accidents. So we should be in an uproar over this one death and demand perfection from computers while 35,000 people die each year in the US alone (over a million worldwide). Seems logical.
What the actual fuck?
If that is true, whomever designed that system to work that way should be on trial for manslaughter. I can think of few worse ways to design such a system.
Inspector: “Does this car have a braking system?”
Engineer: “Yes. Of course.”
Inspector: “Does it work?”
Engineer: “The braking system? No, it’s disconnected.”
Inspector: “Why?!”
Engineer: “Because it would slow the car down!”
It feels like a bunch of different systems that were designed by different teams and no one was talking to any of the other teams about them.
Team A: “Hey, we need a complex emergency braking protocol.”
Team B: “Even a strong braking machine learning algorithm is going to have false positives that will directly impact the passenger experience. Best to simply note when an emergency braking protocol is called for.”
Team C: “If the system is going to have false positives, then we need to not distract the driver with them, so let’s not go overboard with alerts.”
Team D: (To drivers) “NO! It’s a self-driving car. You don’t need to pay attention to the road.”
Trust Uber to always be shitty in everything. Fuck them.
Move fast and don’t brake things.
Is it not standard practice to do drink/drug tests in any fatal accident? I note they tested the victim.
Here in Burlington, VT, there are two types of crosswalks. One is the usual white lines at the traffic light, where you have walk/don’t walk signals. The other is a strip of paint across the road where, legally, pedestrians always have the right of way; you are supposed to watch and stop if a pedestrian is in the crosswalk or about to enter it. There are no lights for these, and they “activate” as soon as the pedestrian enters them. Many are up around the University of Vermont area, which is also where the region’s major medical center is. You can imagine how much foot traffic there is, especially when you also consider it’s at the top of the hill sloping down to the downtown area and the lake.
People stroll into the crosswalks without looking all the time, secure in the knowledge that it’s the law that they have the right of way. How effective that knowledge is in stopping a dump truck is something I leave to the interested reader.
It is 100% effective.
Until it’s not.
Effectiveness drops sharply at that point to 0%.
You and I read about tragedies like this every day. They must read about them too. But there is a disconnect that I don’t fully understand. Possibly total faith in the competence of every driver. After all, they are licensed by The State.
I imagine them silently thinking: “Don’t you dare hit me or I’ll sue you if I’m still alive, you SOB!”