> knowing some of the stuff they do behind the scenes to ensure safety

Any info on this? I would love to learn more about what they're doing.

https://waymo.com/safety/ is probably your best resource. You can follow the links from there to learn about all the things they're doing with simulation, etc.

Absolutely. Glad California DMV is taking the more level headed approach here and evaluating them case by case based on merit.

In the case of that particular incident, it seems that the pedestrian was “thrown” into the lane with Cruise when it was hit by another (human) driver.

Ah yeah, when something doesn’t conform to your narrative, it surely is an astroturf. All coming from accounts that are many years old and have thousands of karma points that leave meaningful posts on all types of topics. But the second they post about Waymo in a positive light, that instantly means astroturfing and not their true opinion based on their lived experiences.

Realistically, there are a lot of Googlers and Xooglers on Hacker News.

One of the messages in this thread is literally stating they have special knowledge of the behinds the scene workings of Waymo.

I work in the AV space and I honestly don't think there's much of a competitive mentality between the companies in the space, but at the same time there is a tendency of people to psudeo-astroturf these brands: where they know someone working at the company and take great personal pride because "I know so and so who works on self-driving cars!"

Astroturfing isn't "I have worked here and am therefore biased when expressing support for the product", it's (to quote Wikipedia) "the practice of hiding the sponsors of a message or organization (e.g., political, advertising, religious, or public relations) to make it appear as though it originates from, and is supported by, grassroots participants."

If they explicitly disclose their internal knowledge or affiliation, it's not astroturfing. If they used to work for the company but are now unaffiliated, it's not astroturfing. If the company isn't knowledgeable or supporting of the effort, it's not astroturfing. etc.

Most of them did not mention any relationship: hence the "psuedo-astroturfing"

They have partial relationships that they don't feel the need to disclose.

> They have partial relationships that they don't feel the need to disclose.

Well, that's your assumption. I would preface that with "probably", since we don't know this for a fact.

> Most of them did not mention any relationship: hence the "psuedo-astroturfing"

Side note: I think you mean "quasi-", not "pseudo-". [1]

But, more to your intended point: the point I'm getting at is that that you're really stretching the definition of astroturfing here, even with the weasel wording added on it.

Astroturfing is a central campaign coordinated by a vendor in a manner as to appear decentralized. It's really an accusation against the central entity rather than as the pawns they're using in the process. So quasi-astroturfing might be a situation where (say) the vendor doesn't actually organize the campaign, but pays an advertiser who internally organizes the campaign without the knowledge of the vendor. The vendor might overlook signs that this is happening, without directly supporting it. Or they might not realize it at all. That's something that would be close to astroturfing, but not exactly it.

However, people advocating for a product due to their genuine personal opinions in a way that's clearly uncoordinated and unsupported by the vendor is not remotely "astroturfing" in any sense. It's a biased expression of opinions, with a potential conflict of interest (if they're still affiliated with that vendor).

All of which I think matters not only because words matter, but also particularly because this site's rules treat astroturfing differently from "someone is biased on the internet", and the admins very much don't want accusations of astroturfing without evidence.

[1] Pseudo- means fake, like how pseudoscience is fake science. Pseudo-astroturfing would be fake astroturfing, or the faking of fake grassroots support. Meaning you'd be accusing people of being real grassroots supporters who are pretending to be fake ones, which would make for a funny accusation.

Sure thing.

People taking it upon themselves to talk up a brand is the opposite of astroturfing, not pseudo-astroturfing.

When you're indirectly related to the brand and that's affecting your eagerness, it's something akin to astroturfing in why astroturfing is looked down on: ie. you're biased

But it doesn't have the malice/intent of astroturfing.

Everyone is biased. There are no perfectly unbiased opinions. And "having a bias" is not the same thing as astroturfing.

Thank you. Contrary to other replies, you do not have to be employed directly by [Institution] expressly for the purposes of covertly pushing an agenda to be guilty of astroturfing.

What's important is the existence or appearance of a conflict of interest, particularly if you're not disclosing your connection. Having a friend who works for Google who might be upset at criticism counts. Being a former employee with stock counts. Having any kind of monetary or professional interest in FSD companies succeeding counts.

If someone is presenting their statements as that of a person who is not involved, and they actually are, their behavior is duplicitous, and justifiably characterized as astroturfing.

It's justifiably characterized as biased. We lose a useful term for distinguishing normal bias from marketing or PR instigated efforts. That's the point of the term astroturfing. It's useful in that regard. Or I suppose, if it's definition has truly shifted, was useful.

Can a stand-alone complex be an astroturf? I think so, if the social conditions that create the behavior stem from a single entity that stands to benefit.

The point of astroturfing (in comparison to the term it derives from, grassroots) is that it's not people coming together to support a policy, it's people with a vested interest in an entity taking cues from that entity to act in their interests in such a way that it appears as the former. The definition never shifted, the tactics just became less obvious. I would be less suspicious if more comments were complaining or giving measured thoughts, but they weren't.

If I may really stretch the metaphor, I'd say individuals with vested interests are more like sod. They really hold those opinions, and are motivated to express them, but do so "unnaturally". Vs commercial efforts that are entirely fake and are therefore astroturf :D

Bias exists. But by removing the financial incentive from the definition of astroturfing, you make it useless when it comes to distinguishing from bias. Otherwise you turn it into "how many degrees of separation can I find to Waymos" and it gets infinitely tenuous.

I would agree that Waymo employees or shareholders shouldn't be commenting here without disclosing their affiliation. But that's a relatively small class: Googlers (and, increasingly tenuous, Xooglers) don't get paid by Waymo and don't get exposure to it through their GSUs. You can make the claim that Waymo's success adds to Google's brand prestige and has knock-on effects, and maybe that's true to some limited extent. But comparing that to being directly paid by the company to spread propaganda is quite the reach.

>But by removing the financial incentive from the definition of astroturfing

I didn't. I pointed out the less obvious ways a financial incentive might exist.

Excuse me for being skeptical of someone who uses a cutesy nickname for former Google employees.

> I didn't. I pointed out the less obvious ways a financial incentive might exist.

You did no such thing, which is exactly the issue.

>What's important is the existence or appearance of a conflict of interest, particularly if you're not disclosing your connection. Having a friend who works for Google who might be upset at criticism counts. Being a former employee with stock counts. Having any kind of monetary or professional interest in FSD companies succeeding counts.

> One of the messages in this thread is literally stating they have special knowledge of the behinds the scene workings of Waymo.

I may have hinted as such (being an ex-Googler), and I know this thread has gotten kinda buried but I wanted to address the implication anyway.

When I left Google I sold all of my stock. My investment portfolio is independently managed and I do not advise any specific positions except for the exclusion of my current and former employers. I do not stand to gain financially from any company in the AV space.

The guidelines ask you not to accuse others of being shills.

I feel kinda the same way about Waymo like I did about Apple as a relatively early adopter. The experience is genuinely magical, and there’s a massive gulf between Cruise and Waymo’s capabilities.

A sufficiently good product may produce customer reviews that are indistinguishable from astroturfing.

Brian might be right but he is also an anti-car / pro public transport activist. He also hates electric cars, in case you were wondering.

(I kind of agree with him, but you have to understand that there is NO implementation of self driving cars that he would be happy with).

Just because some people's opinions differ from yours doesn't mean astroturfing is at play.

Did they give that reason?

"unable to move and blocking traffic" is not a new failure mode for cars. I get that it's particularly infuriating to be blocked by a car that is, mechanically, in good shape, but I don't find the story particularly damning. I've blocked traffic that long waiting for a tow truck.

The case he described is different in that, apparently, the vehicles in question lock all of their wheels and can't be accessed, even by law enforcement. A special tow has to be called in. It's a design decision made to protect the company at the expense of the public.

What's your opinion on Prop22?

Can any engineers in the field comment on whether the [very cool!] Waymo vs Cruise performance anecdotes being discussed here would have been the result of Google's millimeter-level scanning of SF? Or is it just better algos only? Or both?

Cruise vehicles have sufficient sensor payloads to produce a 'milimeter-level scanning of SF' with many fewer hours than they've been on the road there. I doubt that provides the biggest difference. It's all down the software at this point. Especially because they're not trying to do anything crazy like vision only.

We keep hearing this is crazy, but is there evidence to back that up? At this point with hundreds of thousands of Tesla vehicles logging millions of miles a month, I feel like we’d be hearing non-stop reports of destruction on the highways if it really is as crazy as the average HN reader has been led to believe.

>with hundreds of thousands of Tesla vehicles logging millions of miles a month, I feel like we’d be hearing non-stop reports of destruction

Why do you believe that Tesla vehicles are driving millions of miles a month on autopilot?

I suspect that autopilot accounts for a vanishingly small percentage of Tesla miles-driven. Mostly because it sucks, but also because of how hard it is to not miss a nag on a long road trip, with the result that you are locked out of autopilot until your next stop.

(Yes, even when paying attention to the road, with your hands on the wheel. Hell, autopilot locks you out if you exceed 85 mph for more than a few seconds by pressing the accelerator, such as when passing. This is true even in places where the speed limit is 85 mph, and the slower end of the flow of traffic is 95 mph.)

I love the performance of my Model S Plaid. Autopilot, however, is a joke.

"This is true even in places where the speed limit is 85 mph, and the slower end of the flow of traffic is 95 mph."

Where do you drive where the speed limit is 85 mph?

Poland? 140km/h is the motorway speed limit. Admittedly there aren't that many Teslas here. But they are extremely popular in Germany and you see people drive them at silly speeds.

no

>>no

Care to expand?

"Extremely popular in Germany" is a bit of a stretch, from googling around, there are only around 40k Teslas out of 48 million cars overall in Germany (which is about 0.08%).

Well, yes - I think what I really meant was that they are extremely popular[among electric cars]. Whenever I drive through Germany and stop at an autohof all tesla chargers are always occupied, and you see quite a lot of them on the roads just driving around. But you're right, 40k is nothing in such a big country.

It's a bit of a stretch to just randomly drop some numbers... In the US in the last three Q's the marketshare of Tesla is 0,003868139%

Some German cars are driven at sillier speeds. I've been on the autobahn driving my Model S at well over 200 kph and been passed quite quickly by big German cars.

Highway 130 outside Austin, Texas has an 85 mph zone.

There are a number of highways in Texas with an 85 mph limit.

> This is true even in places where the speed limit is 85 mph, and the slower end of the flow of traffic is 95 mph.

Once traffic rules are enforced by autonomous cars that don't let you go over the speed limit, we can start raising the limit to more realistic levels. I guess currently it is priced in that everyone goes over by x km/h.

>Once traffic rules are enforced by autonomous cars that don't let you go over the speed limit, we can start raising the limit to more realistic levels. I guess currently it is priced in that everyone goes over by x km/h.

Bloody noghtmare fuel that, but it's been the inexhorable direction manufacturers and Governments have been pushing for for a while.

Stop making things that enable people more than it constrains them. The act of velvet glove societal manipulation is pretty damn obnoxious once you realize it's a thing.

Tesla reported that there were 400,000 cars in the US and Canada who had access to FSD Beta in January 2023, so I don’t think that’s at all hard to believe.

Edit: Looks like back in January they were adding 10M cumulative miles driven per month (roughly) - https://insideevs.com/news/633328/tesla-fsd-beta-now-active-...

Assuming those data are accurate, it looks like you’re right.

Surprising, but it is what it is.

I probably use AP >80% of the time on the highway. From observing comments in the local Tesla groups, I'm sure there are plenty of people who behave similarly.

Millions of miles per month doesn't surprise me at all.

Wow, I've had much more positive experiences with Autopilot on multi-hour highway trips. No shutoffs, no phantom braking in the last year or two. Autopilot on the highway is so much better than any other car I've driven with adaptive cruise. Anyway, the parent is comparing city and highway self-driving and they are completely different.

the only reason we don't hear non-stop about destruction is that you're required to keep both hands on the wheel/yolk at ALL times while FSD is activated. comparing the current cameras on a Tesla to the human eye is just silly really, sure they cover 360 which is superior, but in almost any other way, the human eye is superior having better dynamic range, better contrast, better resolution and best of all, we have eye lids. if you would just sit down and think about it for 10 mins you'll realize that gimping your self driving car by using cameras only is just a very very silly proposition when we have all this cool tech like lidar essentially giving the car super powers.

If FSD drivers were having to constantly intervene because the car couldn’t accurately map obstacles, we’d be hearing a lot more about it. I drive with FSD all the time - I could give you a list of things it needs to improve on, but not a single one has anything to do with its accuracy of understanding its surroundings.

I do see stories about FSD seemingly trying to drive into obstacles fairly often. It’s true that it does see most obstacles, but most is not good enough for this.

>not a single one has anything to do with its accuracy of understanding its surroundings.

This has been my gripe for a long time. I feel like many in tech have conflated two problems. With current software the problem of perception (ie "understanding its surroundings") is largely solved*, but this shouldn't be conflated with the much more difficult problem of self-driving.

*for sure, there have been issues with perception. A glaring example is the Uber fatality in AZ.

This exactly. Reading the comments and understanding the huge gap between perception and reality of FSD is eye opening. There are a lot of armchair experts here who wouldn’t be caught dead driving a Tesla but are so confident in their understanding of its strengths, weaknesses, and the underlying reasons.

Accuracy of surroundings is absolutely something it could improve on. Adding a different modality (like lidar) would be like adding another sense. Seeing an 18 wheeler without under guards would be easier with an additional sense. It makes the intelligence part easier because the algorithm can be more sure about it's interpretation of the environment.

No longer true. Now they use the camera to verify you're looking at the road.

Also, we can drive just fine with two cameras and a neural net, no LIDAR.

In engineering, you don't add something unless it's necessary, not because "well we may as well have it."

you listened to Musk a bit too much.

And no, a neural net and two cameras are not "just fine". The day cameras will be as good as your eyes and your neural net will be on the level of human intelligence (AGI) then maybe it would be possible. But until then you will need to rely on extra hardware to get there.

Go check on youtube how FSD behaves in city with 1/10th the complexity of SF/Waymo. And remember the difficulty is with the long tail of unexpected events.

We don't drive just fine, we routinely kill each other, often because of poor visibility or not noticing motion. Backing into a busy street? Bam. Open your door without checking? Biker down. Passing a bus at a crosswalk? Pedestrian dead. Driving at night with heavy fog? Off the cliff.

Even your basic non-fancy non-AI car these days have a variety of sonar/radar assists to help out their cameras. Tesla is just being cheap (and getting people killed because of it).

Speeding and drunk driving are the two big ones that kill people.

I’d think better algorithms. A detailed scan of a city’s streets and layout is obsolete before it’s completed.

A driver (bot or not) needs very good reactions to bad situations observed in the field. And ideally a good sense for when a situation is one step from trouble. As a driver, if I can’t see through a bus then I’m going to assume a crazy pedestrian is waiting to jump in front of me.

(A lot of drivers are quite bad frankly and don’t do most of the stuff you’d want someone to do if they’re directing a one-ton blunt weapon around town at 35mph.)

"A lot of drivers are quite bad frankly and don’t do most of the stuff you’d want"

I'd say 90% of driving is good/safe driving is mental, with at least half of that bening based on the decision making capability. Yet our tests are mostly focused on physical ability with a small amount of memorizing a subset of the rules.

I wish we put our drivers in driving sims the way pilots have to accrue flight sim hours.

Pilots (at least to get the PPL) train at least around 40-60 hours on a live aircraft before flying solo. That's more rigorous, yet very close in principle to what we have in place to obtain a driver's permit, I'd argue.

I certainly didn't have 40-60 hours in a car before I got my license. Much less learning permit.

In Germany it’s 28 theory lessons, at least 12 practical lessons (including night driving and Autobahn), 1 multiple choice exam, 1 practical exam to get your driver’s license. Additionally, you need to successfully attend a full-day first aid course.

The multiple choice and practical exams are not done by the school, but by an accredited / governmental institution. Each have a ~30% failure rate.

Counting 1 lesson unit as 45 minutes it’s definitely not quite as rigorous as a PPL, but certainly in the same spirit, I’d argue.

I think it only took me a half dozen hours or so to get my license? But also that only tested me in best case scenarios with good visibility, other good drivers, etc. A better sim should test and train my response to unexpected events in less than ideal conditions.

Mine was 30 hours of lessons, 7 of behind the wheel practice. Most of that practice was very plain and simple. Parking lots, side streets, etc.

Being completely honest, I did not feel totally comfortable behind the wheel until maybe a week or of actually driving myself, post-license. (Not that I didn't have a lot more to learn, obviously)

During that week, I could have easily gotten into an accident and did have one or two closer calls.

I would not all be opposed to maybe an additional ~10 hours of more rigorous behind-the-wheel driving, or a very good-quality sim for that same amount of time.

But realistically, I'd imagine making this a hard requirement would hit some massive walls. Good simulators are expensive, everyone wants to drive, and states have very limited budgets for this sort of thing.

It's the sensors and Lidar.

> Waymo is amazing, knowing some of the stuff they do behind the scenes to ensure safety

Such as?

> My biggest fear has always been that Cruise or Tesla would shit the bed so bad we don't get any self-driving, either because of regulatory constraints or ruining the public perception of them

Exactly this. When I saw Tesla releasing their half-backed crap I feared they will damage the public perception so badly that we wouldn't have self-driving cars for years.

I know most of it was PR but the stuff Waymo did about showing a test course they built and actually having "unit tests" related to driving showed me they're not dumping these things on the road to make money, they're trying to get it right.

The competition:

Uber kills someone and turns out they disabled "breaking due to objects in the way" because it makes it jerkier.

Tesla is often killing the drivers because they marketed it as "auto-pilot"

Cruise rolled over someone for 20 seconds.

I would have to agree. I honestly hate regulatory capture in most industries, but perhaps Waymo's "unit tests" should be turned over to the DMV and made as the gold standard driving test for AI drivers.