Hacker News Reader

Comment by xnx | Hacker News Reader

Meanwhile, Waymo can sensibly move through a street with crowds of people walking around: https://x.com/TechTekedra/status/1828816255404687773

View on HN

The thing I wish we better understood about Waymo is how much the remote human operators are actually intervening on a daily basis - maybe I don’t know where to look, but I’ve never gotten a clear answer here. AFAICT the success of Waymo means Tesla needs to have a similar level of human oversight for its FSD vehicles. But since Waymo’s PR is all about the autonomy, way too many people have the impression that Waymo’s advantage is solely about better algorithms along with humility around limited geographical range. These are probably important factors! But I strongly suspect we don’t appreciate the human oversight.

Their remote operators have pretty much nothing to do with the video shown, though. It's been described in detail the way that remote operators interact with the car and it's not in a way that would be particularly helpful here. You may be under the common misunderstanding that remote operators can directly control the vehicle.

AFAICT it has not been described "in detail," the only writeup I know of is this vague, flowery blog post: https://waymo.com/blog/2024/05/fleet-response/

> it's not in a way that would be particularly helpful here.

Why would you say that? This claim seems completely unsupported. What if the human directed the Waymo to drive more slowly and carefully than it would otherwise? That sort of instruction seems entirely consistent with the blog post and would be critical to the impressive behavior we see in the video.

> You may be under the common misunderstanding that remote operators can directly control the vehicle.

I think you are under the misunderstanding that since humans don't control the vehicle, they must have only a second-order impact on safety. But I am not concerned about self-driving's technical ability to dodge obstacles, I am concerned about its judgment around passing school buses or not slowing down when there are lots of pedestrians. It's this sort of human-level reasoning that I suspect is critical for Waymo's safety record. Nothing in Waymo's blog post / infomercial suggests otherwise.

We can't appreciate it because it's unknown to the public. For Waymo's detractors, the assumption is it's AI stands for Actually Indians; for its fans, it's Nvidia GPUs and Google TPUs with zero human input. The truth is somewhere in-between, but as a customer of their product, how much do I need to care? if the system requires an army of contractors to actually work, rasing costs, that's between the Waymo investor and their balance sheet. Though I guess as someone that owns Google stock I technically am an investor.

Well, as I looked at my stock portfolio today, basically anyone who invests in funds even tangentially related to the tech sector, including ones with a bond component, you're in Google, Apple, Nvidia, etc, considerably.

Meanwhile Waymo just yesterday went the wrong way on the street.

https://abc7.com/post/wrong-waymo-driverless-car-goes-oncomi...

Waymo hits stationary pole.

https://www.youtube.com/watch?v=HAZP-RNSr0s

If Waymo is as good as you claim, then why don't they release data on the number of manual interventions>

It's just the Waymo's missteps are unlikely to show up on HN unlike Tesla's.

FSD can navigate crowds too.

They do release data on interventions in CA during autonomous testing with safety drivers [1][2][3] (distinct from driverless testing). As a ADS system under test, they are legally required to unlike Tesla who characterizes their system as a ADAS and thus absolves them of any responsibility for robust data collection or reporting.

Their numbers indicate ~17,000 miles between any safety driver disengagement. Tesla does not release comparable data, but the highly pro-Tesla biased community-reported numbers[4] average ~42 miles between disengagements.

[1] https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

[2] https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

[3] https://thelastdriverlicenseholder.com/2024/02/03/2023-disen...

[4] https://teslafsdtracker.com/

There are far more Teslas that have purchased FSD than Waymos, and the Waymos are trained on those specific streets.

Maybe Tesla should treat FSD like Waymo, only available in certain cities to a few hundred drivers? As a pedestrian that would make me feel a little better about it.

> Waymo goes the wrong way on the street

I'm familiar with that intersection in Tempe. There is a bar for beating the average human driver on it, and going to wrong way is sadly not above it.

I'm also not aware of any fatalities resulting from a Waymo accident. (Waymo also doesn't seem to be deluded about its system's safety. That's why they have human drivers on standby, remote and in person.)

From the video, it's hard to see what happened there. It's in the middle of an intersection. It kind of looks like the Waymo vehicle was attempting a left turn but something prevented it from finishing it.

If Waymo is as good as you claim why don't they release data on the number of manual interventions?

They didn't even release the video when a Waymo hit a bike while turning left at an intersection a few months ago.

They do. They report "disconnects" to California's DMV.[1] Collisions are reported in detail.

Tesla doesn't report to CA DMV because they are only a "driver assistance system".

[1] https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

I would not be surprised if the Waymo was left "stranded" in the intersection due to red light runners.

In SF, at least, the Waymo vehicles are nigh perfect. I happily occupy the street with them as a pedestrian and bicyclist.

Yeah. You can tell riding in them and watching the little screen, that the Waymo has a superhuman ability to detect pedestrians and cyclists around it.

I would much rather my children bike on a street where every car was a Waymo, than a street full of human-driven SUVs.

I also like that the presence of Waymos has an overall traffic-calming effect even when human drivers are present.

This sounds like a nice side-effect - I wonder why it happens.

Same. If we ever get (bike+AV)-only roads in SF, I'll be happy to let my children to to school on them. Very trustworthy cars.

Reply on news.ycombinator.com