Hacker News Reader

Comment by themafia | Hacker News Reader

If your premise is "robotaxis are so much better than human drivers" then this is almost a disaster. This is only the 10th city they've deployed to, all in the south, and nowhere there's significantly inclement weather. It does not bode well for their expansion plans.

View on HN

Better is an arbitrary statement. By number of jobs robots lose, by number of sexual assaults by taxi drivers they win. Pick the wights for very factors and you can select anything as the best in category.

Safer, cheaper, etc are less arbitrary.

> This is only the 10th city they've deployed to, all in the south, and nowhere there's significantly inclement weather

You may be relieved to hear Waymo is rolling out to Portland, Oregon. It's not in the south, and with over 150 rainy days per year, it ranks among the rainiest US cities.

I'll be relieved when I hear that they did it without killing anyone. Considering they didn't bother to work out how to handle floods before they put people's lives at risk everywhere else, it's not all that reassuring that they're now going to YOLO it in Portland

Nobody has ever been killed by a Waymo. You're being dramatic.

Over a given driving distance, compared to humans, Waymos produce a 90% reduction in serious injury, 90% reduction in pedestrian strikes, 83% reduction in airbag deployments, 85% reduction in cyclist strikes [1].

We currently sit in the ballpark of 300,000 pedestrian deaths per year worldwide [2]. You should be relieved every time they deploy to a new city.

[1] - https://waymo.com/safety/impact/

[2] - https://ourworldindata.org/data-insights/more-than-a-million...

Rain is one thing, but despite the rain Oregon is almost dead-last among all the states in terms of flood risk. It gets constant drizzles, not sporadic deluges.

This guy knows what he’s talking about.

Born and raised in GA, it wasn’t until I moved to CA, the bay specifically, after college that I realized things like flood warnings multiple times a month and, flooded out roads during the summer weren’t just part of life lolll

My ex moved to ATL from Seattle, and it was just WILD watching her go… “you guys have RAIN, here… like it comes down HARD”

When Waymo came here and also when Tesla started doing self driving (I drive a Tesla with FSD ) majority of the time, I was constantly seeing things that were GA specific that these systems were just clearly not trained to handle.

The data was there but it wouldn’t surprise me if the folks building these ADAS systems had just no clue what to do to handle cases like “ice storm caused all the roads to be iced over and now there’s no lane markings” and “flash flood comes out of no where” and “it’s so dark there no street lights for a couple of miles”

> My ex moved to ATL from Seattle, and it was just WILD watching her go… “you guys have RAIN, here… like it comes down HARD”

So it makes sense to first rollout to a place with frequent, lighter rain - no? As an outsider, Waymo's approach seems to be solving challenges step-by-step, and the criticism in this thread is asking why it hasn't already solved the hardest cases.

> The data was there but it wouldn’t surprise me if the folks building these ADAS systems had just no clue what to do to handle cases like “ice storm caused all the roads to be iced over and now there’s no lane markings” and “flash flood comes out of no where” and “it’s so dark there no street lights for a couple of miles”

I wouldn't be surprised if Waymos are confidently driving into flooded roads because they "know" where the markings are without sensing the markings. Lidar-based GPS + SLAM are now very good at calculating location, as long as features like buildings or trees are still present.

You don't understand! Google is trying to do something difficult, and because they haven't solved all possible theoretical problems with it, they should just give up and go home and never try anything difficult ever.

This is a common retort used to cut cost and push things out the door in tech during mvp. However, given most of the world doesn’t look like the west coast, maybe having human drivers would’ve been a good idea until a couple of seasons had passed.

Also they could’ve tested this in other places too?

I'm a Waymo supporter, but I hate to break it to you, Portland gets less inches of rain per year than most major US cities.

Well, only one Waymo got stuck in that flood, while at least two human-driven cars did, so by pure counting metrics they are better lol. But in my experience driving around them Waymos are much much better than most Atlanta drivers, not that that's a high bar

the real question if you’re attempting to imply what i think you’re implying should be:

how many human driven cars decided not to drive through vs how many waymo’s decided the same?

I'm not sure why you would say there's no significant inclement weather in Atlanta. The flooding this week was not super common, but also not unheard of. It rains here a LOT in the summer

Agreed, this happens here every year, it's why we built O4W park the way it is, and built many other drainage structures similarly. We have a real runoff problem. Waymo picked a great city to train the cars on weird weather and weirder roads. :D

The part of that people aren't considering is that it's very common to get brief, intense thunderstorms that dump a lot of rain quickly. They won't flood the whole city obviously but there's _always_ pockets that have very short-lived, localized flooding on the roads. So it's not a "oh what are the odds of that happening" kind of thing.

It's a delay. The question is how long? Doesn't seem unfixable.

I would assume that after the very first instance you would start moving to fix it. To be in a position where you have to roll back your plans doesn't seem like a simple "delay."

The question is: why haven't you fixed this already?

> The question is: why haven't you fixed this already?

Since you're of the opinion that this is taking too long, what do you think is a reasonable time for a fix, and why? I'm assuming Waymo didn't have a team of flood-detection experts twiddling their thumbs waiting to be prompted into action.

Human drivers are very very bad. Being better than humans is a low bar with plenty of room to be bad as well.

No they aren’t. Billions drive around every day with minimal collisions. Far more people get raped than hit by cars.

It still seems to be a high bar to achieve

How high the bar is isn't worth arguing about. The question is do self driving cars clear it?

Those who work on self driving cars say they have cleared it - but they have an obvious bias. Nobody independent has done a full study of this, so we have no particular reason to believe them, but we also shouldn't completely discount them (when the truth is in their favor everyone with a bias will tell it, and some people are even able to overcome their bias when the truth is against them)

Reply on news.ycombinator.com