Comment by TulliusCicero
20 hours ago
It's fascinating seeing all the comments elsewhere anytime Waymo starts testing in another city along the lines of, "ah, but how will they handle X, Y, and Z here?? Checkmate, robots!" despite having already launched service in several other cities.
Granted, NYC is the biggest city in the US, so maybe that sort of reaction is more reasonable there than when people in Dallas or Boston do it.
Since Waymo is very reliable in LA and SF, you will be just fine in NYC.
Your grid system is far less of a challenge than the amount of hills, twists, narrow streets and low visibility back streets in California.
I genuinely think the most complicated challenge for Waymo in NYC will be…winter snow and ice.
NYC is a new set of challenges. As you already mentioned snow and ice is new. But also missing the high density of people and cars per square area. Behavior of drivers and pedestrians are different and much less polite. I can see it working in NYC but "just fine" is a bit of an over confidence... at least not for the first few years before they learn to deal with these issues that they don't face yet in LA and SF.
We do have narrow streets in LA with double parked cars, cars parked in the street only allowing one car through the middle at a time, and plenty of construction closures and obstacles.
Why do so many NYC people think there’s comically no cars in LA or neighborhood streets?
Also, I can assure you LA drivers are a tad bit more aggressive than NYC drivers (less honking and flicking off though, LA people are more a drive you off the road or into the shoulder sort of passive aggressive).
I was born and raised in NYC and have lived in LA for quite some time, still going home often for family. I’m really struggling with reading these “NYC is unique” comments regarding Waymo traffic.
3 replies →
Don't forget the unique NYC challenge of people waiting to cross the street not on the sidewalk but just into the street itself.
7 replies →
What snow and ice? We don't get much of that anymore. That was actually the last thing I am worried about here. I really want to see how Waymo does with NYC drivers and obstacles(double parking on block where sometimes you have to pull in your mirrors just to get by(if you even take the chance instead of just laying on your horn). In some neighborhoods it can be so annoying.
I think what OP means is Waymo's most challenging rollouts will be to places that do get lots of snow and ice.
I think a well designed winter specific FSD system is probably more safe in snow and ice than a human. For instance downshifting to ensure wheels continue to spin on slippery surfaces, subtle corrective steering to keep the vehicle within its lane, etc. should be easier for a FSD car since it won't panic and over-correct like most people do in those situations.
And if the car reduces speed when appropriate and some assholes start tailgating it, it won't suffer the anxiety of holding up 10 cars that want to drive beyond the safe, reasonable speed for the snowy/icy conditions.
Pretty much all electric cars have single speed transmissions, so there's no downshifting. And modern vehicles have electronic stability control, anti-lock brakes, automatic emergency braking, and several other safety systems. It's pretty hard to overreact with those enabled. The main issue is that people exceed safe speeds for the conditions, making them unable to brake or turn in time to avoid a collision.
Right now, most self-driving software will refuse to activate in conditions of poor visibility. I've had that happen with Tesla's FSD, though in that case it was snowing so much that the road should have been closed. Also when the snow is deep enough that your front bumper becomes a plow, it will refuse to activate.
2 replies →
The problem with places that have real winters is that lanes migrate. They are not absolutely positioned. Nor are the sides of the road edges which may project well out into the street and parked cars even further. No road markings visible. Humans make their own lanes. This situation can happen for many weeks-long periods in a typical winter in, say, Minneapolis or Buffalo suburbs.
If a self-driving car does the right thing staying "in lane" while all the human drivers do the wrong thing flocking to new emergent paths (which swing back and forth across the "lanes"), then the self-driving car is wrong and dangerous. I'm not talking about when it's actively snowing either. I mean the snow on the ground just remaining there, covering things.
It's not about dealing with slippage or skill driving, it's about complete lack of context markers. I don't think any current or near future self-driving solution can adapt to this.
1 reply →
> I genuinely think the most complicated challenge for Waymo in NYC will be…winter snow and ice.
Nah, I'm betting it'll be the locals. They'll get pissed off at it remaining stopped when it shouldn't and do shit like start ramming into it. I've had it happen on the island when I stopped at a yellow. NYC is a lot more chaotic than any other US city I've driven.
I would argue those two areas are very different though. The Bay Area is not as dense or as many aggressive drivers as NYC.
My point wasn’t to say they are the same, more that SF and LA (I would guess) have covered and defeated almost every single challenge and obstacle for an urban environment (sans..weather).
LA also has far denser areas than SF, places like DTLA and Koreatown are more dense than most boroughs in NYC (sans Manhattan).
3 replies →
Oof I don’t know about that. Driving in NYC is much different than San Francisco. Frequent lane departures, cutting into heavy traffic despite technically lacking right of way, and other moderate rule breaking is required to get anywhere. Boston will be even more challenging due to the hundreds of convoluted intersections.
We've been getting less and less of those, though. And even then, it's just for a few days. Last year was a bit worse, but two years ago it was very, very mild, I think. Yay global warming?
The thing is waymo at least in LA specifically geolocks you from those hilly areas. Imo it also is not assertive enough and drivers seem to be learning one can bully a waymo on the road.
LA and SF are not close to close to the complexity of nyc traffic and pedestrian culture
TIL.
LA doesn’t have complex traffic? What sort of traffic do we have in LA then?
LA is walkable, it’s lazy (and mostly incorrect) to say LA isn’t walkable.
LA County is massive, and depending on where you want to pick a comparison from, you may prove yourself either right or wrong.
7 replies →
is it just me or its common that in nyc people bike in most cases like there is no traffic lights at all? this to me is prob the most challenging
NYC is also one of the most regularly built out cities, in stark contrast to Boston, for instance. OTOH roads here may be 3 or even 4 levels high at the same point (e.g. where Manhattan bridge meets Brooklyn), and GPS is sometimes way off in canyons between skyscrapers.
> OTOH roads here may be 3 or even 4 levels high at the same point
And here I thought Chicago was complex with lower lower Wacker (just 3 levels).
> GPS is sometimes way off in canyons between skyscrapers
This is probably very challenging for human drivers using navigation, but probably no nearly as much of a problem for a Waymo car with onboard 3D maps of the entire operating area.
HD mapping gives much better than GPS I would imagine.
I think the main difference for NYC is that quite a few streets and intersections routinely have 10x to 100x the pedestrian traffic of the busiest such intersections in pretty much any other American city.
That's not to say that I don't think it'll be able to handle it, just that it'll be a new challenge. I wonder if their current program of apparently trying to positively track every single moving object in range will survive that, or whether they'll need to figure out some algorithm to prioritize objects that are more likely to be of concern to it. And there probably are more than a few places where pedestrians are numerous and densely crowded enough that you can't positively track all of them, even with a bunch of LIDAR sensors.
NYC (at least the parts I've spent a bit of time in) is pretty grid like with fairly simple roads. The drivers are the hard part :)
I am excited to see them tackle Boston at some point because of how strange some of those roads are. The first time I had ever been I came to an intersection that was all one ways and there were like 7 entry/exit points. My GPS said turn left, but there were three paths I'd consider left. Thank god I was walking.
And I don't really pose much doubt because it seems like Waymo's rollout plan has been solid, but I'm just interested to see how well they tackle different cities.
Next time go below 14th street. The grid is great but very misleading. NYC has had 400+ years of development, and it's deeply not homogenous in its design.
Yeah I’m skeptical that robots will ever be perfect drivers but the bar isn’t perfect it’s better than human and that’s certainly possible.
Yup, the data so far seems to indicate that Waymo is substantially safe than average drivers. Obviously it's not inclusive yet since the tech is still new (and while the study I'm thinking of was done by a third party, it's still Waymo that handed over the data and paid them to analyze it).
Roads in Texas specifically just seem to do whatever they want, whenever they want. It's really apparent that Texas local roads used to be wagon trails.
The grid system in NYC seems like a good alternative for a rollout. Though the current NYC human drivers will hate these things. I also expect LOTS of vandalism.
There is a question, NYC driving gets by with everyone driving aggressive and breaking road rules. That is something that does not exist as much in other markets.
My complaint with Tesla city FSD is that it’s not quick or aggressive enough. It will come to long and complete stops and other things that will not work well in NYC.
I think you'd be surprised how aggressive driving has become normalized throughout the world?
The world, for sure. But not in America. I've driven in most major US cities and I drive daily in NYC. There really is no comparison in any other major US city to how we drive in NYC.
I drive when I travel, though not in NYC these days but being a pedestrian gives you a good enough view. LA from a city driving perspective at least for me is not comparable to most of NYC.
The big difference is that NYC is less law abiding and more devious. Unless you've lived here, you have no idea the lengths new yorkers will go to scam and vandalize.
Source: grew up in NY, moved 25 years in SF. Love Waymo, big investment in Google.
What I don't get about the "checkmate robots" mentality is that, like, get it working in sunny California with no snow, and then get it working in the snow, seems like the way to do it, not, solve all possible problems before anyone knows you're even trying and can make fun of you.
This type of edge case covering is pretty essential to the jobs of most on this board. It doesn't surprise me to see it.
>despite having already launched service in several other cities.
Why does having launched in other cities matter if the new city brings up things that none of the other launched cities do?
For example the first thing I can think of new for New York is snow and ice.
It's my understanding that self-driving cars don't really account their acceleration and braking for roads that could sometimes be very slippery due to snow and ice.
> "Why does having launched in other cities matter if the new city brings up things that none of the other launched cities do?"
New requirements come up all the time in technology. The existence of a new requirement isn't in and of itself justification for skepticism - is there a particular reason to believe that Waymo is not capable of solving for the new requirement?
The answer may be yes, but simply "ahah! It would need to do [new thing]!" is insufficient. "[new thing] is likely intractable because [reason]" would be more justification for skepticism.
> "It's my understanding that self-driving cars don't really account their acceleration and braking for roads that could sometimes be very slippery due to snow and ice."
Sure, but like above - is there a reason this is an intractable problem?
I'll throw this out there: your human-driven car already accounts for acceleration and braking on slippery roads, without the need for the human. Traction control systems and electronic stability control systems exist! They're in fact incredibly common on modern cars.
The interesting snow & ice problem for me is that humans will drive in winter conditions that are unsafe -- for example white-out blizzards. Robocars won't be able to drive in a white-out blizzard, so they'll likely refuse to do so. Humans should also refuse to drive, but people drive anyways.
NYC doesn't generally get white-out blizzards, so refusing to drive in them is quite feasible.
5 replies →
>I'll throw this out there: your human-driven car already accounts for acceleration and braking on slippery roads, without the need for the human. Traction control systems and electronic stability control systems exist! They're in fact incredibly common on modern cars.
These systems don't help with the problems I am talking about.
You have to drive completely differently in heavy snow, significantly slower, brake sooner, turn less sharp, accelerate much slower, leave significantly larger gaps, leave space to move out of the way and be ready to move if someone behind you is coming at you too fast and can't stop in time, etc. I've spend my entire life in the midwest.
The traction control system in my 2023 camry didn't help one bit when I applied the brakes on black ice and the car didn't react at all, it just kept sliding at the same speed across the ice.
3 replies →
Launching in other cities with new problems gives experience dealing with new problems, and the meta-learnings transfer to better processes for adapting to new issues. But yeah, ice and snow are definitely major new environmental factors for New York (and DC, and many other places we are starting to see more serious testing).
Autonomous vehicles can and do take into account surface conditions, there’s not really any reason not to. There are pretty good generative models of the physics of vehicles with different surface conditions, and I imagine part of the data collection they are doing is to help build statistical of vehicle performance based on sensed conditions.
A fair point about weather, but a lot of the assertions are like "how will they handle the double parking and suicidal pedestrians jaywalking across the street??" I'd say most of the concerns just don't sound very unique at all.
For weather, Waymo has clearly started out in warmer climates while slowly building out towards places with colder and colder weather, I'm guessing they're just incrementally getting better at it.