First driverless shuttle in Las Vegas crashes on first day while shuttling passengers

Originally published at:

1 Like

Interesting, I guess the AI is not smart enough to blow the horn?


So the onboard systems couldn’t adapt to an emerging accident situation? And they had an attendant in the vehicle, who one assumes has manual override capabilities? Pretty lame…


It just… stopped? I thought it was supposed to either crash into a wall, killing the passengers but saving a crowd of innocent bystanders, or crash into and kill a crowd of innocent bystanders, thus saving the passengers.


AI has got to account for human error in other vehicles. Like maybe honking the horn, at least.


Next on the chopping block as machines take our jobs: intro to philosophy classes.


Wow, that’s pretty rare, to crash while completely stationary.

Pardon me for personifying a fucking self-driven shuttle, but this headline and most commentary REEKS of victim-blaming.


Yeah, the vehicle didn’t crash; it was stopped, and the truck that was pulling out crashed into it. There’s this disturbing trend, even when talking about regular human drivers, that attempts to distribute blame from where it belongs–the vehicle moving forward, onto all parties in the road. You hear it all the time from people who want to turn the speed limit into the speed minimum, you also hear it from people who try to make “distracted walking” a thing. It’s BS.


“First driverless shuttle in Las Vegas is crashed into while stationary on first day while shuttling passengers” is apparently not enough to draw those extra clicks. It also doesn’t exactly roll off the tongue.


I wonder if Vegas casinos will start taking bets on whether the driverless shuttle crashes that day.

It really ought to be no surprise that the shuttle is a risk.

(EDIT: because it’s in a gambling city [it appears the joke wasn’t obvious enough]).

1 Like

The bus may not have been at “fault” but this is still an example of a situation where a human driver likely would have prevented an accident by honking the horn. Sometimes you just need to field test this stuff before you find out where all the vulnerabilities are.


Yeah, definitely.

If there had been injuries, and a human driver, who would be liable? Not the shuttle. The lack of a human driver doesn’t change that, even if a human could theoretically have prevented the crash using the horn, or by backing up itself.


Nope, distracted walking is totally real. I see people all the time crossing streets with their noses buried in their phones, whether they have the light or not. And half the time when you stop for them at a crosswalk they stand there at the corner still looking at their phones.

As for the OP, I’d venture to guess there’s probably more vehicle contacts with only one vehicle moving than any other kind. Think of all the parking dings, which this basically was.


Misleading title is misleading.


Surprised it isn’t another BB contributor…


No one’s saying it isn’t real. We’re saying that it doesn’t at all abrogate the driver’s responsibility to not run people over.


As CGP grey noted, when we get rid of the monkeys behind the wheels, we will have a giant reduction in accidental deaths.


The burden of responsibility lies on the person piloting a multi-ton vehicle, at all times. That wise old dude Stan Lee put it pretty succinctly: With power comes responsibility. Any other formulation of responsibility leads to an unsustainable escalation that justifies gas-guzzling SUVs as the “safest” vehicles on the road; our public spaces cannot possibly bear the continual externalisation of risk from individual drivers onto the public at large.


Qualified agreement, as with most CGP Grey videos…
Automation works great, particularly in spaces that we’ve already defined as car-only, like freeways, but in that video that he had, didn’t it strike you as weird that the “ideal” intersection he rendered didn’t have any pedestrians in it? Cars were continually moving, which is great for their energy-efficiency, but that left absolutely no room for pedestrians or bike-riders to make their way anywhere. You can imagine some sort of urban-planning workaround that institutes superblocks or some other parallel way for pedestrians to make their way around these rivers of cars, but that also radically increases the overhead for realizing those efficiencies.


I’ve driven a lot more than 2 hours and have yet to had an equivalent accident.

I think there’s a very good chance that a human would have done something to avoid this accident, either by honking the horn, taking evasive manoeuvres, or simply not stopping in a place where a truck was likely to back into.

This is the big challenge for self-driving cars, safe operation of a vehicle is much more complicated than simply obeying the rules of the road.