“Thought of the day, and I wish there were a way to get this to legislators:
Come the next Big One earthquake, all of San Francisco’s emergency services will be blocked by Waymos.”
I’m AMAZED they’re not designed to handle this better. This does indeed seem like a massive problem. “Oops we give up” right when things get the worst? How is this OK?
I’ve been very impressed by Waymo’s more cautious approach. Perhaps they haven’t fully thought through the ramifications of it though.
Because it’s a power outage. If we instead learned about this during a real disaster people could have died because these things were let on the road without planning what they should do in abnormal circumstances.
> If we instead learned about this during a real disaster people could have died
This is universally true. The question is how bad could it have been, and in which cases would it have been the worst?
> We’re lucky it’s not a disaster
This is always true. Again, the question is how lucky?
We have an opportunity to count blocked emergency vehicles and calculate a hypothetical body count. This lets us characterize the damage. But it also permits constraining hysteria.
It can do that with a normal driver too. Doesn’t make it ok for there to ever be a situation where they need to when the target vehicle/driver is just fine and capable of doing it themselves.
Yes. OP is inferring Waymo's internal processes from this meltdown. ("Makes me think there are likely other obvious use cases they haven’t thought about proactively either.")
If Waymo literally didn't foresee a blackout, that's a systemic problem. If, on the other hand, there was some weird power and cellular meltdown that coïncided with something else, that's a fixable edge case.
>If Waymo literally didn't foresee a blackout, that's a systemic problem.
I agree with this bit
> If, on the other hand, there was some weird power and cellular meltdown that coïncided with something else, that's a fixable edge case.
This is what I have a problem with. That’s not an edge case. There will always be a weird thing no one programmed for.
Remember a few years ago when a semi truck overturned somewhere and poured slimy eels all over the highway? No one‘s ever gonna program for that.
It doesn’t matter. There has to be an absolute minimum fail safe that can always work if the car is capable of moving safely. The fact that a human driver couldn’t be reached to press a button to say to execute that is not acceptable. Not having the human available is a totally foreseeable problem. It’s Google. They know networks fail.
It also means that their claims of "autonomy" are fraudulent, like most "self driving" cars. A car which depends on powered infrastructure outside the car to drive is not autonomous.
I think the 100 or so miles I can generally drive in my car while it still has gas is different from my car just stopping suddenly because of a power outage.
Yeah, I was shocked by this. Blackouts in California aren’t some sort of rare event. I’m primed to expect rolling brownouts/blackouts yearly in the summer.
Neither a lack of traffic lights nor cell service should cause the Waymos to stop in the middle of the road, that’s really troubling. I can understand the system deciding to pull over at the first safe opportunity, but outright stopping is ridiculous.
Waymos rely on remote operators to take over when the vehicle doesn't know what to do, and obviously if the remote connection is gone then this is no longer available, and one might speculate that the cars then "fail safe" by not proceeding if they are in a situation where remote help is called for and inaccessible.
Perhaps traffic lights being out is what caused the cars to stop operating autonomously and try to phone home for help, or perhaps losing the connection home is itself enough to trigger a fail safe shutdown mode ?
It reminds a bit of the recent TeslaBot video, another of their teleoperated stunts, where we see the bot appearing to remove a headset with both hands that it wasn't wearing (but that it's remote operator was), then fall over backwards "dead" as the remote operator evidentially clocked off his shift or went for a bathroom break.
That’s clearly unacceptable. It needs to gracefully handle not having that fallback. That is an incredibly obvious possible failure.
Things go wrong -> get human help
Human not available -> just block the road???
How is there not a very basic “pull over and wait” final fallback.
I can get staying put if the car thinks it hit someone or ran over something. But in a situation like this where the problem is fully external it should fall back to “park myself” mode.
> How is there not a very basic “pull over and wait” final fallback
Barring everything else, the proper failsafe for any vehicle should be to stop moving and tell the humans inside to evacuate. This is true for autonomous vehicles as well as manned ones–if you can't figure out how to pull over during a disaster, ditching is absolutely a valid move.
If the alternative is that the vehicle explodes, sure. And since GP did say "final fallback", I suppose you're right. But if the cars are actually reaching that point, they probably shouldn't be on the road in the first place.
The not-quite-final fallback should be to pull over.
I think it’s clear that both use cases are a must have during an emergency. Even more, rescue services and stranded people would need all the bandwidth and reception they can get, Waymos shouldn’t be online during such times at all.
FirstNet is AT&T. Verizon and T-Mobile have their own alternatives (FrontLine and T-Priority, respectively). QCI is a fun rabbit hole, because some of these networks share the first responder QCIs with certain business use cases.
I lost cell during the whole outage on Verizon, came back immediately when power was restored. There seemed to be some towers up, if i walked down the street I could find one, but plenty were down.
It was predicted by many, including me. It'll be a lot worse in an earthquake where power and cell service are out and there's debris and road damage. Good luck to our first responders.
Coincidentally we were on the Robotaxi during the black out (didn’t know about it, we were going to Japan town from the Mission). Noticed that it navigated through the non-working traffic lights fine, treated it like a stop sign junction. One advantage of building unsupervised system from public version that had to deal with these edge cases all around the country.
Though the safety driver disengaged twice to let emergency vehicles pass safely.
I'd default to assuming it's the respective roadmaps for Waymo and Tesla differed on which things to implement when, not training data, that results in the two behaving different.
50/50 bet it would either go right through or treat it as a stop.
Don’t think I have had a totally inactive light. I have had the power is out but emergency battery turned to blinking red light, and it correctly treats as a stop sign.
FSD is level 2. Level 3 doesn't require the human driver to monitor the outside environment, only take over when requested. Tesla also doesn't report data from FSD under L3 reporting requirements anywhere in the US.
SF Muni & BART both stopped service in many areas. Though most of the trains still had electricity, many sensors and control systems were inoperable. Also underground stations had no lighting, so it would be hazardous to allow people to board or exit there.
Waymo's problem is obvious in hindsight, and quite embarrassing for them, but it can be solved with software improvements. Tesla's FSD already treats dark traffic lights as stop signs, so I would bet on Waymo fixing this as soon as they can.
But transportation that depends on infrastructure along the whole route (such as trains and busses powered by overhead lines) are always going to fail in these situations. I think that's acceptable considering how rare these events are.
Read more carefully, Waymo didn’t boast about their numbers and that part of the story was unrelated to the issue except as a measure of impact their shutdown could have.
“Thought of the day, and I wish there were a way to get this to legislators:
Come the next Big One earthquake, all of San Francisco’s emergency services will be blocked by Waymos.”
I’m AMAZED they’re not designed to handle this better. This does indeed seem like a massive problem. “Oops we give up” right when things get the worst? How is this OK?
I’ve been very impressed by Waymo’s more cautious approach. Perhaps they haven’t fully thought through the ramifications of it though.
https://mastodon.social/@jripley/115758725115731454
Were any emergency vehicles actually blocked?
We have an actual failure here–step one is identifying actual failures so we can distinguish what really happened from what hypothetically could.
They need to drive or pull over. Never just stop there in the road and wait.
They're not. But it's also not a disaster. Pretending it is on Twitter is pandering, not policymaking.
> They need to drive or pull over. Never just stop there in the road and wait
Agreed. Waymo has a lesson to learn from. Sacramento, and the NHTSA, similarly, need to draw up emergency minimums for self-driving cars.
There are productive responses to this episode. None of them involve flipping out on X.
Because it’s a power outage. If we instead learned about this during a real disaster people could have died because these things were let on the road without planning what they should do in abnormal circumstances.
We’re lucky it’s not a disaster.
This is universally true. The question is how bad could it have been, and in which cases would it have been the worst?
> We’re lucky it’s not a disaster
This is always true. Again, the question is how lucky?
We have an opportunity to count blocked emergency vehicles and calculate a hypothetical body count. This lets us characterize the damage. But it also permits constraining hysteria.
Sure, but it would be notable if one had to. If none had to, we have a problem to solve, not a catastrophe.
Makes me think there are likely other obvious use cases they haven’t thought about proactively either.
We have zero evidence a power outage wasn't foreseen. This looks like a more complex multi-system failure.
Once you’re on public roads, you need to ALWAYS fail-safe. And that means not blocking the road/intersections when something unexpected happens.
If you can physically get out of the way, you need to. Period.
Yes. OP is inferring Waymo's internal processes from this meltdown. ("Makes me think there are likely other obvious use cases they haven’t thought about proactively either.")
If Waymo literally didn't foresee a blackout, that's a systemic problem. If, on the other hand, there was some weird power and cellular meltdown that coïncided with something else, that's a fixable edge case.
I agree with this bit
> If, on the other hand, there was some weird power and cellular meltdown that coïncided with something else, that's a fixable edge case.
This is what I have a problem with. That’s not an edge case. There will always be a weird thing no one programmed for.
Remember a few years ago when a semi truck overturned somewhere and poured slimy eels all over the highway? No one‘s ever gonna program for that.
It doesn’t matter. There has to be an absolute minimum fail safe that can always work if the car is capable of moving safely. The fact that a human driver couldn’t be reached to press a button to say to execute that is not acceptable. Not having the human available is a totally foreseeable problem. It’s Google. They know networks fail.
Perhaps traffic lights being out is what caused the cars to stop operating autonomously and try to phone home for help, or perhaps losing the connection home is itself enough to trigger a fail safe shutdown mode ?
It reminds a bit of the recent TeslaBot video, another of their teleoperated stunts, where we see the bot appearing to remove a headset with both hands that it wasn't wearing (but that it's remote operator was), then fall over backwards "dead" as the remote operator evidentially clocked off his shift or went for a bathroom break.
Things go wrong -> get human help
Human not available -> just block the road???
How is there not a very basic “pull over and wait” final fallback.
I can get staying put if the car thinks it hit someone or ran over something. But in a situation like this where the problem is fully external it should fall back to “park myself” mode.
Barring everything else, the proper failsafe for any vehicle should be to stop moving and tell the humans inside to evacuate. This is true for autonomous vehicles as well as manned ones–if you can't figure out how to pull over during a disaster, ditching is absolutely a valid move.
The not-quite-final fallback should be to pull over.
Sure, I go out and drive around on the roads for no reason all the time. I'll avoid doing that during the crisis.
Some folks enjoy driving.
https://news.ycombinator.com/item?id=41688847
Though the safety driver disengaged twice to let emergency vehicles pass safely.
https://www.tesla.com/ownersmanual/modely/en_eu/GUID-A701F7D...
It might work, if other Tesla drivers regularly drive in that area. Also, it might not work, and you should assume that it won't work.
Looks like it treats it as a 4-way stop. Is this because Tesla has more training data?
Don’t think I have had a totally inactive light. I have had the power is out but emergency battery turned to blinking red light, and it correctly treats as a stop sign.
Its human takes over. FSD is still Level 3.
(Robotaxi, Tesla's Level 4 product, is still in beta. Based on reports, its humans had to intervene.)
Waymo's problem is obvious in hindsight, and quite embarrassing for them, but it can be solved with software improvements. Tesla's FSD already treats dark traffic lights as stop signs, so I would bet on Waymo fixing this as soon as they can.
But transportation that depends on infrastructure along the whole route (such as trains and busses powered by overhead lines) are always going to fail in these situations. I think that's acceptable considering how rare these events are.