I highly recommend everyone try it out if you're in SF. It's an incredibly smooth and sure ride. The cars are really nice too (Jaguar I-Pace electric cars), clean and spacious.
The first time you ride in one, it feels truly sci-fi. But within 5 minutes, you're almost bored of it - that's how good it is. If I had to choose between an Uber of questionable cleanliness and driver temperament and a Waymo with a slightly longer wait and slightly more fare, I'd choose the Waymo every time.
(I have no affiliation with Waymo, Google or any related industry - it's just an amazing service!)
I love how we've gone from "Taxis are gross and dirty, that's why I love Uber!" to"Ubers are gross and dirty, that's why I love Waymo!" in the span of 5 years.
What do you think comes next? These cars are literally unsupervised.
Unsupervised in what sense? There's internal cameras that are periodically checked. Weight and seatbelt sensors that give alerts if a passenger is or puts objects in the drivers seat, or if too many individuals get in the car.
I'd be shocked if a similar or greater level of observability doesn't also exist outside the car.
I've had a dude with a Tagalog accent take over the speakers in my car and asked me and the folks I was inside with to leave because I decided to clown car it with friends visiting from out of state. With that said, there for sure is someone monitoring, but it's similar to how checkout at Amazon Go stores went.
I'd be curious to know how this monitoring scales over time.
We have computers that can drive a car in city traffic and you're worried that we can't have a computer look at the inside of a car and tell me if someone's left a mess? A non-ML model background subtraction algorithm from the 2000's could tell you that.
“Massive fart detected. Initializing ejector seat in 3, 2…”
I’m not sure how you could make that work. People that eat Indian food step in and it might get set off, so people will cry racism. You drive next to a particular industrial plant, etc..
I don’t work for Amazon but I have worked with the tech. The media really ran with a misunderstanding there - the error rate for JWO is dramatically lower than people seem to think.
Unsupervised in the sense that there's no driver there to clean the car when it gets dirty. At the same time, presumably part of the odor of an uber comes from the driver, so that helps. I wonder if waymos roll down the window (weather permitting) between riders, to air it out. Might be dangerous at a red light, since people could throw stuff in or even dive in through open windows.
True! Presumably they'd refuel the vehicles in these places as well. The number of depots would depend on how much time you want to waste having vehicles driving out-of-direction and (1) the cost of additional real estate plus (2) the cost of employees at each location.
> How do you remotely detect if car stinks or not?
Why do you think they need to? These cars go to a centralize location every night, I would assume, and cleaning at scale can happen. How often do you think an Uber driver cleans their car in comparison?
Not to mention the problem is pretty eminently solvable: let users flag a car as having an unacceptable condition. If that happens dispatch a new car and send the dirty car directly back to home base for cleaning.
The main problem with Uber/taxi quality is that the responsibility for cleanliness is on individual drivers, with widely varying results as a natural outcome. In fact a lot of problems with Ubers and taxis is downstream from the ownership and responsibility model.
The advantage of something like Waymo is that the responsibility is now on Waymo itself.
Worth noting that this is a problem in other lines of business: it's harder to ensure quality in franchises vs. stores operated directly by the brand. The more independent parties you have in the mix the more incentives become misaligned and the fewer levers you have available to ensure compliance to some standard.
This also isn't impossible to solve with human drivers, because ultimately this isn't a technological problem but an organizational one. Livery car services where drivers are employees (as opposed to independent owner-operators contractors) can centralize cleaning and training, and have more means to ensure compliance to a standard.
The "downside" of such a model is that there are many more laws to ensure you can't shovel your own expenses onto the employee.
> It's not like people constantly trash hotel rooms
I worked as a house attendant in an upscale hotel for years. Let me tell you from experience: a shockingly high percentage of people trash and damage their rooms. Not the majority, but enough to keep staff busy every day.
I can't believe I have to specify this, but I obviously meant "other than by the human watching the room". Equally I don't expect the Uber driver to be the one vandalizing their own taxi.
Uber started as off duty black car drivers doing gig work. Those were the default right? The random guy with a car gig was rolled out as uberx later?
But I'm completely with you on the unsupervised part. People doing all sorts of things back there that an ML might not identify. Now if they hire an army remotely to monitor, I guess that could scale because of wage disparities.
Not just wage disparities. You can probably watch the typical video of an individual sitting calmly scrolling on their phone back at 32x speed. More complex scenarios with multiple people at 16-4x. Even assuming they pay 'car monitors' as much as drivers (which you're right, they probably don't) the cost for the monitors is still probably less than 10% the cost for equivalent drivers.
Not just wage disparities, but also time / location.
Quickly going through CCTV footage from a ride to catch "unusual things" takes much less time than driving. Even if somebody had to watch the entire ride at 1x speed and just one ride at a time, there's no need to do it when the car is idle or driving to pick up the next passenger. You can also hire a lot less watchers than you have drivers, because a surge in demand in one city can be serviced by watchers in many other cities. There's no need for night shifts either, you can make everybody work 9-to-5 in whatever timezone they're at.
Well it's because as Uber expanded and had to desperately fight for its margins, it started to lower its standards for UberX. Before, cars had to be fairly new and in a relatively good condition. Now it's no longer the case.
You have ID of who was in the car and when, and a 'report' button in the app for the next passenger. Easy enough, if something is reported you have the choice to wait for the next one while the trashed car goes to be cleaned or ride anyway at a discount if you solve it yourself - e.g: put the trash in a bag in the boot of the car. After N reports attributed to you, you get banned.
Uber can have reporting as much as this, or more. I can think of some: the driver owning the car incentivizes keeping it clean; bad reviews for dirty cars.
They supposedly use trip info for their own advertising. Says also "in some cases" shared with third parties for ads.
"We collect usage data that includes trip history, buttons or links you click on our mobile app, in-vehicle interfaces, wait times for our vehicles, and other actions you take with our products and services"
"We collect information about your location in a few different ways. When you take a trip, we collect the pickup and drop-off location and details about the vehicle’s route."
"We may also use your information to personalize services, advertisements, content, and features, communicate with you including marketing (which you can opt out of), service, and account messages, or communicate other information we think will be of interest to you "
"We will retain information we associate with your Waymo account, such as name, email and trip history, while your account remains active."
"Waymo will not disclose your personal information with a third party unless one of the following applies:
...
- We're involved in a merger, acquisition, reorganization or sale or transfer of some or all of our assets
- In some cases to help us tailor ads and offers to your interests as detailed in the U.S. state law requirements section
- Comply with any applicable law, regulation, legal process, or governmental request
...
"
Public transportation is not a true panacea even where it’s seen most favorable. Car ownership increased 14% in Europe from 2012 to 2022, across the board:
The future will be a combination of trains, a more extensive robotaxi network for the last mile, which can be cars or vans, and a much smaller percentage of personal vehicles compared to today. Buses as they are today will decline.
You joke, but I've heard more than one argument about how self-driving cars will drastically improve traffic because, for inter-city travel, they will be able to drive very close together on the highway in a long convoy, leaving almost no space between cars...
I don't think it's such a bad idea, actually. There's value in having something that can drive in a "train", but that can also disconnect when needed and drive on its own.
If you're driving long distance, you get the advantages of a train, but door-to-door, with no scheduling conflicts and no egregious stop times. If your destination is 5 minutes away, it's still a car.
Or, you know, you could drive (or bike! or waymo!) to the local train station, which should have <30 min, well-scheduled, predictable headways to major cities, then take a train that goes 150+mph, and then drive (or bike! or waymo!) to your final destination, for _far_ less total energy expenditure.
Every car in a train-caravan is schlepping along its own engine, crumple zones, airbags....and with rubber tires...
I got access to Waymo in LA a few weeks ago and have taken it 4 times. It's capabilities are impressive for sure, but I'm not sure I'd go as far as "smooth and sure ride". The car's skills seems to vary between impressive and "nervous new driver". It drives like someone that got randomly stuck into a much bigger car than they are used to.
When I rode in one, admittedly a long time ago when cruise was still operating in sf too, the waymo car pulled over for a firetruck. However, the firetruck was merely crossing the road we were on and it was 3 streets away. And the waymo pulled into a bus stop to do this. The safety driver had to nudge it back into traffic.
Prove that. Because last time I checked in there wasn't a self-driving system in play that wouldn't faceplant into wet cement, blow through crosswalks, or find itself trapped in a cleverly deployed ring of salt. Additionally, by their own admission Waymo is nowhere near level 5 autonomy, which means they still haven't reached parity with what a mediocre human driver is capable of.
Well they just announced that they are doing tens of thousands of rides per week. That either works or it doesn't. And they seem pretty comfortable doing that. Also there's a distinct lack of horror stories involving Waymo cars. And I assume that they don't have thousands of mechanical turks wielding a joystick somewhere, which means these things are mostly working as advertised (i.e. autonomously) with the very occasional manual intervention.
So, what you are asserting and that cannot be true at the same time. So, my conclusion is that whatever you think you know here is probably wrong.
Idgaf if they're round-tripping the equivalent to mars and back weekly, the fact of the matter is their tech cannot perform at the level of a human driver who's had their license for a week, by their own admission. No amount of argle-bargle changes that as the autonomy levels are very well-defined and Waymo self-reports as L4 under optimal conditions and no clear path to L5 in sight in the next decade. Let's not confuse a legislative fuckup on the part of cdot with actual technical prowess yeah?
You are splitting hairs. Tens of thousands of rides per week. Autonomously. Those are the keywords. Other things they are bragging about involve such things as 24/7, night and day, and foggy conditions. I would suggest actually reading the short press release. They make quite a few interesting claims in it. Anyway, anyone in the SFO area will probably be reporting all the wonderful and uneventful rides they are enjoying with Waymo soon.
As for people that recently got their drivers license. I'm pretty sure that demographic is over-represented in the statistics of who drives the least safely, traffic fatalities, etc. Also insurers and rental car agencies have policies that reflect those cold, hard statistics. It will be interesting to see what they do when level 5 starts happening (probably sooner rather than later). My guess is that they'll charge people extra for the privilege of taking control of the car as they are far more likely to damage the vehicle and otherwise cause trouble.
And obviously one of the points Waymo is trying to make with their press release is that they are already safer. It's a press release of course and not the same as cold hard facts. And you make a fair point about self reporting. But it suggests the obvious notion that computers are getting pretty good at not crashing into stuff (or people). I find that entirely unsurprising, BTW. It does not seem like a particularly hard problem.
I'm not splitting hairs. The difference between L4 and L5 autonomy is enormous. I submit that you cannot get enhanced safety outcomes out of a system that cannot perform the task at hand at least as well as a mediocre human, and realistically not until you're routinely out-performing the bulk of human operators. And if you think not crashing into random shit in a perfectly chaotic environment is an easily solved problem please explain to me why store PoS tech (orders of magnitude smaller problem space) has been a roiling dumpster fire for decades now. Software doesn't have an amazing track record at solving problems. It has proven to be absolutely phenomenal at shifting problem spaces in weird ways and introducing spectacularly stupid unplanned side effects however.
Not only will safety take a long time for technical reasons, but its extremely predictable under what financial conditions corporate execs sweep safety issues under the carpet. If Alphabet has a few tricky quarters good luck to everyone.
Oh sure yeah. It'd take exactly one legislative push to place full liability in the case of accidents onto the vehicle manufacturer (where it clearly belongs) to fold up every autonomy division in the industry like a wet towel. What galls the shit out of me is there's apparently enough dumb money afoot to float R&D spend equivalent to the GDP of a middle of the road 3rd world country just to make cabbies lives even more miserable.
That dumb money mountain will go to fighting legislation as we have seen with big oil, telco, pharma, wall st etc. The money always flows and aggregates around anything that promises to become a monopoly tomorrow. And monopolies can then collect rent at whatever rate they want. This end state is what attracts money more and more. The money is not just to replace the existing solution, its reqd to capture and lock in talent, attention, suppliers, lobbyists, politicians etc. The more the spend the more competitors exit too, the closer to market capture. But they have been spending for a long time to end up in just 2 cities so the pressure to scale, monetize, compromise will keep growing.
took two waymo rides in SF and two nightmare cab rides in SF and memphis in the last couple of days.
one waymo was perfect, the other ran a red light then stopped in the intersection diagonally across two lanes of traffic until the light turned green. didn't endanger anyone but felt awkward!
on the cab ride to SFO my driver kept falling asleep and veering into the next lane, i tried talking with him to keep him awake but it was clear he'd been driving all morning and was exhausted.
the cab ride from memphis international was interesting, the cab was falling apart, driver was nice but he tried to convert me to christianity for the last ten minutes of the ride.
the waymos will end up safer but totally devoid of character.
The possibility of "character" is always my biggest dread when getting into a taxi or an Uber.
I just want to get from point A to point B without having to make small talk or delicately navigate a political discussion with a stranger. A stranger who both controls the vehicle and, in the case of Uber/Lyft, will be giving me a rating at the end of the ride.
From my (hopefully) neutral point of view I think Waymo running a red traffic light is the worst of all and most likely will end up in a disaster.
For huge metropolitan cities for examples London, Istanbul, Delhi, Tokyo, Jakarta or Sao Paulo running a red traffic with a car can most probably cause fatal accidents. I am now more than convinced that level 5 autonomous driving cannot be achieved in my lifetime, and much better efforts should be better directed toward non-invasive highly accurate early detection of high mortality diseases like CVDs and efforts for properly mitigating climate changes.
I wonder if the cleanliness will be maintained over time? Presumably this requires humans to physically clean periodically, and it seems like this would suffer when the company starts to squeeze costs to improve margins.
The biggest problem with cleaning is that compared to Uber where there's a human driver, the autonomy factor will almost certainly lead to more people fucking shit up simply because they can. Look at anything else that goes unattended in the public way: ebikes, delivery bot things, scooters... nasty combo of "I want to fuck with The Man" and "haha nobody is here watching I don't need to be as careful with my messy sandwich"
> At some point, the people who tent to leave a mess in Waymos just... won't be allowed to ride in Waymos any more.
This would reductively mean a more effective form of punishing vandalism, which would in of itself be a slight net good overall.
In the ideal case, there would be a recompense program to pay for the damages made + extra, in return for being able to ride them again.
Meta-point: This discussion has significant overlap with busses/trains/public transportation. How we deal with bad actors in that space would translate well into this space.
At some point, the people who tent to leave a mess in Waymos just... won't be allowed to ride in Waymos any more.
Starting to sound a bit like a credit system or some other dystopian shit show. You do something dumb, drunk while in college and you're cut off...sounds amazing.
No, the point was that billing the damage to the malicious actor in question will prevent damage.
But for some people, a fine is just the cost of doing the thing. And for every bad actor you ban, there's another bad actor willing to take their place.
They'll accept cash if you fork out a $300 deposit, and the people who are likely to both pay cash and trash a room rarely have that much extra money on hand.
A friend of mine works night audit in a rough part of the valley. As soon as he says the magic words 'we require a major credit card or a $300 deposit', they immediately hang up.
This heavily depends on the country, plenty of places in the US and other credit-card-loving countries are credit (and specifically credit, NOT debit) card only for this reason. Plenty of tourists from countries where such cards aren't as popular get bitten by this.
> Some taxis have that too. Where does it get them?
The keyword there being "some". If an account is tied to the person, recourse for negligent/malicious behavior can be applied, but that only works if there's an account to add the penalty to.
On Waymo's part, requiring an account to use the service would be in the benefit of the service as a whole.
1. The cars are not unattended 2. Public transit and the like get rekt because they are cheap and there is no consequences whatsoever. When you get $400 full detail bill like they do for rentals and/or get banned for life (even as a +1 because they have your face) that tends to have a chilling effect
That's my biggest concern too. My fear is in 20 years nobody will be excited to go in these because the interiors will be whittled down to dilapidated NY subway seats.
Once these become more common I am sure they'll adapt the interior to be more similar to the subway (plastic seats, no carpet, hard plastic everywhere) instead of having luxury car seats.
You need an account, probably with a credit card, if you're gonna ride. If you intentionally mess up the car, you'll probably just get banned from the service.
It's probably a little bit better since a car is a closed environment. That won't save it entirely, but it'll at least mitigate the issue I expect as compared to e.g., a public bus.
I imagine the cars all head to some centralized location for cleaning and whatever other maintenance. That probably makes cleaning a very economic, factory-like procedure. Beats something like Uber where drivers have to bring their cars to a car wash or whatever (although presumably the Uber drivers aren't being paid to wash the cars because they aren't employees, so the unpaid labor there gives Uber an advantage).
I wonder if they'd ever offer riders a discount for taking trash out that was left behind by other riders. They could even include a vacuum and give people a couple bucks off for cleaning it out during/after their ride.
I think eventually they’ll just make a car (like Zooks is) without all the messy crud (steering wheel etc) and with fixed, waterproof seats and such. It will be designed for automatic cleaned by driving into a depot and being hosed down.
Right. This is a prototype. You've got a whole empty seat not being used, and a steering wheel sat there baiting someone to play with it. The vehicle was never designed for the purpose for which it is being used.
I totally agree we will start to see very specific driverless designs that are very easy to clean.
The nhtsa used to require all vehicles have steering wheels. Looks like that rule changed in March 2022 for autonomous vehicles so we should so start seeing those. I guess it just takes a while for new autonomous vehicles to be designed and built to take advantage of the change in rules.
> Presumably this requires humans to physically clean periodically.
That's a good robotics project. All the cars are the same. The cars do not have any objects inside that belong to the occupants and should stay. So robotic interior vacuuming could be quite practical, as a station in the car wash. A vision system can inspect for damage and route that car to the maintenance line.
So smooth they apparently opt to blow red lights instead of stopping abruptly.
source: I commute by skateboard in SF daily. Just yesterday an empty Waymo cruised straight through a fresh red, narrowly missing my entry into the crosswalk.
But don't get me started on what I've seen human drivers do on the same streets. Just annoyed that Waymo's aren't better.
Competing anecdata: Yesterday I was running downtown and a Waymo stopped for me at a green light because it wasn't sure if I was gonna jaywalk (jayrun?). Only once I stood still for a few seconds did it continue to make its turn.
This was in a turn-left-only intersection with a separate pedestrian light. Maybe the Waymo got confused and thought I also had green.
Yes, but SF seems to have some truly reckless pedestrians, often presumably under the influence of substances. There are graduations between "cross at a red light when you think it's safe to do so" and "barge into a 4 lane street in the middle of a block because the voices in your head tell you to".
I sometimes wait on pedestrians when turning right on a green and pedestrians do a little dance on the corner that makes it hard to tell if/when they're gonna cross.
Why would it be illegal to enter on yellow? I thought that's the entire point of yellow. Otherwise we'd just have red and green, and hopefully you can imagine how that would work out.
> So smooth they apparently opt to blow red lights instead of stopping abruptly.
Surprised traffic lights aren't updated to communicate with cars directly. Why not have traffic lights broadcast 'stop' message along with turning red?
I suspect traffic lights, roads and cities will have to be updated to work with driverless cars.
Have also nearly been hit as a pedestrian by a Waymo. I also last week saw a Waymo blow through a yellow-red very aggressively, so I think the downvotes on above comment are intensely biased.
I’d say the same, maybe. But yesterday I had an Uber with a guy that had a degree in history and a love for SF lore, and told me about the tomb of Starr King, and then I went on a quest to find it with my friend. Just saying that there’s magic out there that’s not technological.
> Just saying that there’s magic out there that’s not technological.
Dark magic too, I've been on one too many rides where the driver insists on monologuing on topics that range from detestable (politics) to alarming (the driver was armed, and had picked me up from the airport)
No, I added the pick-up locationbut to make it clear the driver knew I was unarmed as I was coming off a flight, since one can't legally fly with a gun.
In most countries, including US and all EU member states, one can legally fly with firearms and ammunition. There are some rules that must be followed. Sure, you can not take them to the cabin.
Yeah, I've ridden with two cool drivers with history degrees, one jazz musician from Ethiopia, and one who I'd already met elsewhere. Zero interest in riding in a Waymo. (Disclaimer: Alphabet shareholder)
Personally I dont understand why they went with Jaguars. I wish they had gone with Hondas and struck a deal with them to integrate their level 3 Autonomous tech as it is the most advanced in the industry.
They're worth trying once, but they're not ready to replace Lyft/Uber. The cost is similar or sometimes even more for Waymo, and the trip often takes longer thanks to the car's skittishness. The ride length estimates are also quite optimistic which can lead to arriving later than one intends. I don't see a real compelling reason to keep using them.
Is there any kind of limit to them besides the geofence? Can you get a Waymo at night? In the rain? I suppose it never snows there. How about roadworks? How do they react to vehicles with emergency signals? Can they follow directions of a cop in the street?
While they pick up in almost every location the drop off is sometimes "close by" like 3min walk to final destination (the app tells you in advance tho so you can decide to order or not). This is quite annoying sometimes and I picked uber instead.
Are there no freeways inside the geofence? Someone in the thread mentions that they will be adding freeways soon, so I understand it can't do those now.
Based on my limited experience with Telsa model 3 FSD and my Toyota lane assist/radar cruise control, congested highways are the absolute sweet spot for self driving. Not much happening on the sides, and stop and go traffic being quite tedious for the human. It's happy to speed up and slow down over and over again.
uh, in southern california there used to be an unspoken rule: if you can't pass the driver's test with freeway included, you go to NorCo and take the test there. No freeway - the hardest thing is the "parallel park and reverse" which is just being told prior "when you pull over, take as long as you need to get as parallel as you can to the curb, then just go in reverse, look over your shoulder, and don't touch the wheel." Well, that and you automatically fail if you hit a cow.
Freeways in CA aren't the worst in the country (lookin at you, TX), but they're still not easier than surface streets, especially during off-peak hours.
Can you expand a bit on why you feel freeway driving is "easier"? the only thing i can think of is you're much less likely to get into a head-on[0] collision.
There's fewer pedestrians that might or might not be about to step in front of the car. Less need to predict human actions. There's mostly no perpendicular intersections. Generally visibility is good, don't have to remember that there was a car getting ready to enter traffic behind the peach tree. No skate boarders.
In Redwood City (easiest dmv test route I know of), they ask you if you want freeway on the test, you say no and they cross off that whole section. The parallel parking section is done on a completely empty block and the only way to screw that up is to park blocking someone's driveway.
As to why freeway is easier: maybe not as a very new driver, but as soon as you're comfortable handling a car you can get on the freeway and the mental processing loop is much easier. Just follow the car ahead at a safe distance for the speed. Very little else, can be done with just half a brain while you think about other things.
No idea. I'm just reporting on what's already actually happened in the past (my ride experiences) instead of speculating on what might happen in the future.
It's a function of the filtered population. The cars are at full capacity day and night so the increased number of users won't affect a single car's cleanliness nearly as much as the type of people that will be riding in them.
That new group is people who aren't at the frontier of trying out new tech.
The way this will manifest is the drunk idiot who'll puke all over the car. The bored asshole who carves his initials in the seat. The edgelord who gets their jollies out of destroying other people's stuff.
Good chance you had some in the original group as well, but early adopters are usually mostly people who deeply understand tech. Once that falls away, you have a less thorough understanding of tech, and a very surprised realization that booking something under your name that has cameras all over will likely result in you being held accountable for what you did. But after you did it.
How that'll play out in the long run is anyone's guess. If Waymo maintains rigorous enforcement and the courts actually play along, it might just work out. It's still going to be capital-intensive because we seem to have created a world where being a major asshole in public without consequences is kind of an entitlement people expect to have, and Waymo will need a very strong "yeah, not here" vibe to prevent that. Which requires a number of high-profile incidents.
Yes, the subtext of the question likely was "are you discriminating against the poor!?" If it was indeed, the answer to that question is "no, the ride pricing will do that".
I don't know what 'deeply understanding tech' has to do with any of that. Plenty of people who 'deeply understand tech' get drunk and puke in places. Honestly, you sound like a Victorian-era petty lord looking down his nose at 'commoners'.
You read the part where I said the understanding constrains the existing user base from antisocial behavior, right?
This isn't about "commoners". It's that any large group of people contains assholes, assholes restrain themselves depending on circumstances, and the new users might well lack that understanding.
It's fundamentally not about "commoners", but about a knowledge-based restraint falling away. (As for the "commoners", you might also want to read the last part of my comment, instead of outraging)
That person did read it, and the fact that you believe anything you wrote in the entire comment would not contribute to the notion that your are a classist snob is hilarious.
A loaded question is a trick question, which presupposes at least one unverified assumption that the person being questioned is likely to disagree with.
That's a loaded question, but sure, let's go there. The problem with public transportation is that the public is allowed to use them, and the rules, legal and social, are not well defined or enforced. Assaulting other passengers is generally tolerated by the system, depending on the type of assault. Physical assault is considered too far and doesn't go unnoticed, but chemical and audio assault on fellow passengers usually goes unreported. The types of people are those who would assault others on some fashion.
Whether this translates to Waymos smelling like meth or fentanyl when you get in them thanks to the previous rider remains to be seen. Or just needles, foil, or used condoms left behind. They record video, so Google could close the person's account so they won't be able to book Waymos with that account again, so we'll end up having to see how hard it is to create new accounts to use Waymo on to ban nuisance riders.
Breaking news, google introduces social credit system that "only applies to waymo and think of the childen". More at 9
/s
I agree with your point about the approach to the argument, but I think google has enough info to make a way to vet passengers by identifying if you are likely to trash a vehicle.
Granted in my work I've never made the attempt to smell fentanyl, but it isn't one with a reputation of having an odor. I assume you mean the smell of recreational users of fentanyl.
I Guess you must not be riding it as much as you think you do. I’m an infrequent rider and I’ve seen 2 altercations and multiple (10+) disturbances since January.
The funniest (or not if you’re not from SF) was a guy boarding a full bus with a 7 foot long dining table with the legs attached, arguing and threatening anyone who protested. Eventually he dropped the thing on someone’s foot and starting an altercation which delayed everyone by 15 mins.
I've ridden the 22 through the Mission at midnight every Tuesday for much of the last seven years. I assure you that's exactly as much as I thought I do.
You might be magic. Crazy, crazy shit goes down on Muni. I used it to commute from outer Richmond to downtown for years and the bystander effect was fully powered up on many of those rides. And it wasn't always the homeless people, little old Asian ladies could be hella scary.
I might be. I also commuted downtown from the outer Richmond for years, on the 31, the 38, the 1, and the 5. Oh, I've seen plenty of crazy stuff, but only one assault.
Okay, the point I think the poster up there was trying to make is that there are degrees of assault. I never saw anyone murdered, and only twice some sort of physical struggle. The number of times crazy yelling broke out is uncountable by me, though.
Potentially. But, with each driver exercising quality control over their own vehicles, the actual result will likely vary from hitting as good or better a standard to being worse. The Waymo standards, thus far, are pretty high, so I would expect on average Uber/Lyft/Etc. would fair worse on average.
I have no doubt that Google is waiting for more adoption before starting to cut costs everywhere and before you know it your puked out ride will direct you to www.waymo.hr/help to find an article which resolves your issue
Uber has partner drivers which have their own companies, their own rating, and can be punished for their behaviour. Once a company completely vertically integrates (like Google would like), meaning they have their own cars, they no longer want to punish themselves for bad behaviour/cars. Since they have to choose between short term cost of higher maintenance fee or long term cost of loss of quality of service their managers will start to optimize for quarterly results: cutting short term costs. What they want is to first entrench the market, push out competitors, introduce complex regulation and fees which prevents new competitors into the market and then start cutting costs everywhere they can and increase prices.
Since you mention Uber, I can definitely see in my city how the quality of cars decreased and they started using almost inclusively cheap immigrants who realistically couldn't pass a drivers exam in my country and have on multiple occasions driven into wrong directions/ran red lights etc.
Pixel phones? Nest? The Bayview hotel rooms are pretty nice. Hell even Gmail feels pretty premium to me, but I guess this word might be considered subjective.
Also, Waymo isn't even Google. You might accuse me of overstating this, but truthfully they operate as a different company.
Pixel phones have less of a brand than Samsung, much less Apple (not talking about actual product quality, just brand positioning). Nest doesn't stand out, although I don't think any of the smart home things are really established enough to know which are good or bad. I've never heard of the Bayview hotel rooms. Gmail is good but it's not a "premium" feel.
Yup. Plus, if Waymo can clean its cars with greater efficiency at lower cost than Uber can, then all other things being equal, Waymo will have cleaner cars.
The drivers are not the same people who activate their account.
There are schemes where undocumented immigrants ask someone to activate their account on their behalf. In practice, the person giving you a ride could be literally anyone.
Uber doesn't really have a way to increase profit through messier cars. But they can do things like increase prices after taking over a market, which they have not been at all shy about doing.
> Uber doesn't really have a way to increase profit through messier cars
Don't they? Allowing messier, older, and less pleasant cars would increase the supply of drivers, allowing Uber to place lower bids on those drivers, lower their prices, increase volume and revenue, and increase profit.
They can definitely have more beat up cars which over time I can observe in my city. As for prices, luckily they have a lot stronger competition with bolt and local taxi apps as creating a local taxi app is really not that hard
I would be surprised if the Waymp cars weren't exceptionally nice in this phase, while they're trying hard to gain market share and trust. Google Search was a clean and delightful experience once upon a time. The aspect I'm more interested in is what the experience will be like if they ever become a dominant transport option.
Yeah they probably pipe that video to an algorithm that does background subtraction so they're able to assert that the vehicle is clear of foreign objects, but those cameras don't detect smell. If someone defecates and wipes it somewhere the camera doesn't see, then what?
The rider reports an issue and Waymo sends a new car while they send the dirty one back to the depot back to be cleaned? Offending previous rider is charged a fee and then banned for life.
Hard for me to imagine this being a widespread problem. Probably about as much problem as there is with something similar happening in the elevator of a private building.
Why deal with real people when we can plug ourselves into the matrix? Until we get so bored that we ask the matrix to create fake people for us to interact with.
Because of safety concerns or because you like the idea that in principle humans should be driving?
In other words, what kind of safety metrics do you think would be necessary to change your mind?
If in N years 90% of your peers use self-driving cars, and the only accidents you hear about are in the news about people you don't know, would that change your feelings about safety?
Primarily because I Lack trust (safety concerns) and prefer the flexibility of human drivers. Besiides, I think eliminating "simple" jobs like being a driver will result in more social problems.
I am blind. General statistics and adoption rate in my peer group (sighted coworkers) are not representative for me at all. None of these datapoint give a good indication how much I am (as a visually non-participating pedestrian) in danger.
Given my life experience, it is safe to assume nobody really thought about blind people. I am regularily tripping over (and hurting myself with) badly placed bicycles. I am assuming a high percentage of these bicycle owners are well-meaning, green-voting people. However, their political and/or personal attitude doesn't help people like me at all, we still get hurt on a regular basis, and nobody talks about it. I would be surprised if this changed with AVs.
Another comment, another thread mentioned that Waymo requires more walking than an equivalent Uber ride - to the pickup location, from the drop-off location. Anyone know why this might be true?
I'll hazard a guess: because Uber drivers are sometimes willing to stop for a minute in an illegal spot to park to do a quick pick up or drop off, and Waymos are never willing to do that (presumably).
During one of my last Waymo rides the car stopped on Powell between Bush and Sutter (facing South stopping on the regular lane a bit before the Powell/Sutter crossing). This caused other drivers to drive on the cable car tracks to go around the Waymo (which are separated from the driving lane with a double yellow line) and it caused a truck to do a right turn directly from the cable car tracks (as there wasn't enough space to merge back into the lane).
Not sure if was legal or not for the Waymo to stop there, but given that Waymo stops take quite a bit longer than stops with Uber/Lyft (as it takes a while for the car to continue driving) this was one of the worst places possible to stop. Especially as there would have been space available right after the crossing next to Walgreens.
Honestly, it’s almost always legal to stop in a lane. Regardless of the fact that other drivers had bad behavior, that was the right thing to do. Some of the future of all this is that we will need to install curbs in places we don’t have them now to prevent bad human behavior.
This is absolutely false in California, please don't spread dangerous misinformation.
See CVC 22400(a):
No person shall drive upon a highway at such a slow speed as to impede or block the normal and reasonable movement of traffic unless the reduced speed is necessary for safe operation, becauseof a grade, or in compliance with law.
No person shall bring a vehicle to a complete stop upon a highway so as to impede or block the normal and reasonable movement of traffic unless the stop is necessary for safe operation or in compliance with law.
The above was referencing stopping on a city street ("Powell between Bush and Sutter"). You're talking about stopping on a highway. These things are not particularly comparable.
In the California Vehicle Code section 360, a "highway" is defined for the purposes of the vehicle code as "a way or place of whatever nature, publicly maintained and open to the use of the public for purposes of vehicular travel. Highway includes street." [0]
Last night a waymo dropped me off near the giants game. Presumably somewhere you aren’t meant to stop as I heard a security guard on loudspeaker asking “the car with the display on top” to move along as I was walking away, but the car wouldn’t move as there was still relatively fast traffic moving past it.
I’ve had to walk a few times near steep hills, I was wondering if partly it was due to the angle of approach and the sensor view being blocked so they avoid the area?
Fwiw there is a checkbox if you want to absolutely minimize walking. It'll often do things like drive around the block so you don't have to cross the street yourself.
One thing I actually think is really cool about the Waymo ux (full disclosure, I work for Google but not Waymo) is how it elevates the pickup/dropoff locations to feel like more of a first class feature (compared to just typing in an address and your Uber driver dropping you off in the general area).
I don't think I've ever had to walk more than 30 feet to a pickup location, but I have had it drop me off at the nearest cross street (usually on streets where it would've had to double park in front of my exact drop-off location)
Human taxi drivers are okay parking illegally to do pick-ups/drop-offs, and we as a society usually tolerate that, as long as they're quick about it.
But Waymo probably isn't comfortable telling its cars to park in many illegal parts of streets, even if it's going to be quick. Partially because determining which illegal parking jobs are socially acceptable vs unacceptable is a hard determination for a robot to make.
I definitely would not want to be a beta tester of this software. I know too well how the sausage is made, I guess.
We are way too aggressive with allowing self-driving cars on our roads. Companies are incapable of building a toaster without bugs, our banking apps are a burning garbage dumpster fire, and yet we somehow think it will be different this time?
Living here for the last 10 years, it's been jarring how just in the last few years, driverless taxis went from "it'll never happen" to "is this the default now?"
The Waymos are genuinely good drivers. I look forward to taking them every time.
One thing that really impressed me about riding Waymo is that when you start to open the door, it'll give really good warnings if you're potentially opening the door into traffic. Even specifying if there's a passing car/bike/scooter
I don’t look forward to taking them and choose other drivers, mostly because the price and wait time dynamics are a little funny, but I am glad I did take a ride or two. They’re much better drivers in the sense of “not interested in pushing any limits.” They navigated around a parked truck effectively, queueing and waiting their turn to go into the opposing lane behind some other cars. The perception display of surrounding people and cars was very comforting. My only moment of fear was a sudden stop because a wrong-way bicyclist had lurched out into traffic — that’d happen with any driver, unless we hit the guy. Yeah, I guess you can cone them, they’re that conservative of drivers.
It’s clear that they’re not the cars for me to worry about out there on a bike / on foot / etc.
This makes me wonder why this shouldn't become the default mode of public transport — i.e., subsidized by the local govt in the same way as busses/trolleys. It seems they could actually replace the busses and provide better service with a sufficient inventory of cars. Could even provide Waymo-style vans on regular++ routes, with the "++" being ability to divert to nearby residences, especially for the mobility-impaired.
It's us in the EU who are having the sacrifice though. As we're left behind economically and technologically as a load of octogenarian bureaucrats decide to ban genetic engineering, AI, etc.
There is someone to take over as soon as the car encounters something it cannot handle. It's not self driving. By definition it's not driverless, there is always a driver assigned who will take care of it if needed. One of the reason they cannot scale beyond small areas.
Do you think the remote operators are driving the vehicles? They can only give the car instructions (pull over, go back to depot etc). In no sense are they driving.
Regardless of the definition, I can't buy and own such a "driverless car" because I don't operate a supervision center with people available to "assist" the car. Whatever "assist" means. Nor can any company that's not large enough to operate the infrastructure.
I'm sure at least some of the hype is in regards to people owning such "driverless" cars.
It's driverless by definition because it does not have a driver, but it's not autonomous by definition, because it requires outside help to function. Cool hair splitting. It's an impressive technical feat for sure. In regards to perceived benefits for society, it's part of the way there, as it reduces the number of humans required to driver from 1 per car to 0.something per car.
These cars are not meant to be owned. They're part of a fleet.
They are driverless and specifically L4 autonomous, which is the best autonomy you can get right now. The vehicles that can operate without any help doesn't exist. You'll be waiting for a long time if that's your expectation.
Doesn’t seem impossible to offer remote assistance as a paid subscription. The real showstoppers are cost and maintenance of the self-driving hardware which is astronomical
> vehicles that can operate without any help doesn't exist. You'll be waiting for a long time if that's your expectation
Eh, I'd say we're a decade out from an L5 vehicle. It'll officially be L4, on road only. But that matches a good fraction of American drivers' capabilities.
Since the assist doesn't require real time response (e.g., the failure mode is the car not moving, not you dying unlike a tesla) it seems like a non taxi version could simply set you up as the assist monitor.
If you fall asleep or don't respond it's fine, you're just stuck.
I don't think Google plans to do this for a variety of reasons, but I don't think there's fundamentally any reason they couldn't.
> If you fall asleep or don't respond it's fine, you're just stuck.
"Just" being stuck on a highway doesn't sound too safe. If you have multiple supervisors watching the fleet it's less of a problem than when it's one fallible supervisor.
It would go against their operating license for a remote driver to completely take over control of the car. The car is always in ultimate control even when it's getting "assistance" from a remote operator.
If you could rent/purchase a vehicle without a steering wheel, would you not want there to be someone available who can help out when the system runs into trouble?
As long as there is no driver in the car, does it matter how it drives? Is that not just an irrelevant implementation detail that has no bearing on your passenger experience?
As long as it requires some humans to be available to "assist", I'm not sure anyone would sell or buy such a vehicle. Or rent in the "car rental" sense. Taxi service is what makes sense.
If I own the car, and I'm sending it to do errands for me while I work or sleep, seems like the cost of someone being available to "assist" it at short notice would be prohibitive. Unless it requires assistance once a month or so.
If you're sending it to do errands, you might consider also sending it out to take riders, as a side hustle. Do that, and you're now a taxicab service.
Point being, I suspect many people don't want to own any vehicle, whatever level of automation it has (which could be none). The reason we do is that taking taxis everywhere is too expensive. If that cost can be brought down, however it's done (computer brain, cab driver, capuchin monkey, whatever), many people will be happy to forgo owning a vehicle.
Where did you get this information? There are definitely people ready to respond to any requests from the car or passenger when they come up, but I've never read anything to indicate that each ride is actively monitored by a person.
Search for "driver monitoring system". As soon as there is something unexpected there is an actual driver taking care of it. Also the reason why it's not level 5, but stuck at level 4 self driving.
That's not "always a driver monitoring". They're not actively monitoring all of them all of the time, they're responding to requests from the car.
And they're not drivers, they're supervisors. They don't take over driving, AFAICT they issue high-level commands like "safe to proceed", "take the right lane", et cetera.
There is always somebody ready to respond. Of course they can't take full control, that would be stupid dangerous, but they can still control the car. They even have people who go to your car to get it unstuck manually when needed.
Nobody is driving the car remotely, that would be an enormous liability. The car calls in and either asks the operator to make a decision for them or tell it where to go. Once they do that, the car handles the how. You're making a lot of proclamations about this stuff without a strong understanding about how it works.
- Elon (and his pro-analysts) heavily weight the future of the co's valuation on their ability to deploy a taxi network and has been promising it just around corner for years
- Alphabet via Waymo seems to have "solved" robotaxis for city-proximity driving and has deployed as a business.
Beyond the obvious "reality distortion field" argument is Tesla actually in a position to win here due to their manufacturing capability / current deployment of Tesla's?
The only selling point of FSD (Supervised) is that it (can) work "everywhere." This is because it only relies on navigation information and what the car can see.
Waymo and similar companies all use HD Mapping. Ignoring the specifics, it can be thought of as a centimeter-level perfect reconstruction of the environment, including additional metadata such as slopes, exact lane positions, road markings, barriers, traffic signs, and much more.
HD Mapping is great when it's accurate and available. But it requires a ton of data and constant updating, or the car will get "lost," and realistically will never be implemented in general, at best in certain cities.
Reliance on HD Mapping gets you to "robotaxis" quicker and easier, but it doesn't and likely cannot scale.
It remains to be seen if Tesla can generalize FSD enough to reach the same level as HD Mapping everywhere. Still, they have shown that the current limiting factor is not what the car sees or knows but what it does with that information. It is unclear how or why HD mapping would help them at that point.
> HD Mapping is great when it's accurate and available. But it requires a ton of data and constant updating, or the car will get "lost," and realistically will never be implemented in general, at best in certain cities.
Waymo have said time and again they don’t rely on maps being 100% accurate to be able to drive. It's one of the key assumptions of the system. They use it as prior knowledge to aid in decision making. If they got "lost" whenever there was a road change, they wouldn't be successfully navigating construction zones in San Francisco as we've seen in many videos.
> They can also do constant updates because the cars themselves are able to detect road changes, self update maps and rollout changes to the entire fleet.
Which leads to mapping failures being unchecked, as the system that generated the data is the one checking the data by driving it. See bullet point 1 in their recent recall for an example.
> Prior to the Waymo ADS receiving the remedy described in this report, a collision could occur if the Waymo ADS encountered a pole or pole-like permanent object and all of the following were true:
> 1) the object was within the the boundaries of the road and the map did not include a hard road edge between the object and the driveable surface;
> 2) the Waymo ADS’s perception system assigned a low damage score to the object;
> 3) the object was located within the Waymo ADS’s intended path (e.g. when executing a pullover near the object); and
> 4) there were no other objects near the pole that the ADS would react to and avoid.
> the Waymo ADS’s perception system assigned a low damage score to the object;
and Tesla would do better how in this case? It also routinely crashes into stationary objects, presumably because the system assumes it wouldn't cause damage.
> and Tesla would do better how in this case? It also routinely crashes into stationary objects, presumably because the system assumes it wouldn't cause damage.
Are the Teslas in the room with you right now?
Please point out in my comment where I mentioned Tesla. I can wait.
The changes can be checked additionally by humans, although not always.
> We’ve automated most of that process to ensure it’s efficient and scalable. Every time our cars detect changes on the road, they automatically upload the data, which gets shared with the rest of the fleet after, in some cases, being additionally checked by our mapping team.
Doesn’t mean it’s foolproof. But the benefits far outweigh the drawbacks.
Waymo doesn't serve any snowy locales yet. But sure, years and years ago mapping was worse than it is today? The mapping used today is working quite well in warm weather locales.
> Reliance on HD Mapping gets you to "robotaxis" quicker and easier, but it doesn't and likely cannot scale.
If you can make the unit economics work for a large quantity of individual cars, mapping is a small fixed cost.
I agree that it's not economical to map every city and road in the US, since you need to generate revenue from every mapped road and city. So you can think of HD maps as amounting to building roads. They will be built in lucrative places. Cruise and Waymo won't make money from putting taxis in nowhere Arkansas, so they don't need to map it.
> the current limiting factor is not what the car sees or knows but what it does with that information. It is unclear how or why HD mapping would help them at that point.
That's simply untrue. All the hard stuff continues to be reliability and sensor gated. Cruise and Waymo have amazing sensors and even they struggle with sensor range, sensor reliability, model performance on tail cases, etc. For example, at night these cars typically do not have IR or Thermal sensing. They are relying on the limited dynamic range of their cameras + active illumination + hoping laser gets enough points / your object is reflective enough. Laser perception also hits limits when lasers shine on small objects (think: skinny railroad arm). Cars also have limits with regard to interpreting written signs, which is a big part of driving.
Occlusions are still public enemy #1. Waymo killed a dog. Cruise crashed into a fire truck coming out of a blind intersection even though their sensors saw the truck within 100ms.
LiDAR and HD mapping together are supremely useful, even if you don't drive with it, for enabling you to simulate accurately. You cannot simulate reliably while guessing at distances and locations. HD maps let you use visual odometry to localize, and distance measurements grounded in physics backstop the realism of your simulation at least in terms of the world's shape.
Tesla lacks the ability to resim counterfactuals with confidence since they don't have HD ground truth. There are believers at the company that maybe you could make "good enough" ground truth from imagery alone but that in and of itself is a huge risk, and it's what skipping steps looks like. Most in the industry agree that barring a major change in strategy they just have no way to regression test their software to the level of reliability required for L4 / no human supervision.
The obvious thing to do is to just have every Waymo robotaxi or car with licensed Waymo tech report in its daily mapping/obstacle data to the mothership, so you can get new changes almost immediately.
I dunno if said data would be as high quality as dedicated HD mapping cars, but it's probably at least decent, given the variety of cameras and lidars every Waymo car has.
Further, it seems to me that if you brake hard to avoid a dog, your car should warn me as I’m approaching. I’m not sure why we are trying to teach each car to drive when we could be teaching all the cars and the road to drive.
> Further, it seems to me that if you brake hard to avoid a dog, your car should warn me as I’m approaching.
What does this mean? Electric cars are already required to emit a sound as they drive.
I guess if it has to brake hard for something, honking might be a good idea, but I wouldn't want cars to constantly be beeping at everything in their vicinity if there's no imminent crash.
> I’m not sure why we are trying to teach each car to drive when we could be teaching all the cars and the road to drive.
I'm not sure what you mean. Presumably Waymo's software is the same across its fleet. They're not training one car's model at a time.
Well, if your car brakes hard to avoid a dog, your car should warn me. I’m not sure how to make this concept simpler so I can only repeat it.
> Electric cars are already required to emit a sound as they drive.
I know.
> I guess if it has to brake hard for something, honking might be a good idea, but I wouldn't want cars to constantly be beeping at everything in their vicinity if there's no imminent crash.
If you think in a discussion about robot cars that drive themselves being conducted on a hacker website, I’m suggesting that cars communicate their sensor data to each other by honking their horns, in really not sure what to tell you other than yes, this would be profoundly dim witted.
> I'm not sure what you mean.
I believe it.
> Presumably Waymo's software is the same across its fleet. They're not training one car's model at a time.
I believe it. I also believe you’re deeply missing the point, perhaps intentionally.
Agreed, but having the raw data is still useful, especially for less-used routes where it's not economically feasible to send out dedicated mapping cars all the time.
I'm just speculating here, but I can envision a few ways of dealing with the cost problem in scaling an HD mapping-based robotaxi fleet:
1. Robotaxi companies might simply stand to make enough money to cover the cost of routine HD mapping. Anywhere the revenue of putting taxi services in a new city outweighs the cost of implementing the necessary updates sufficiently, won't companies do it? We could think of these companies as having similar economics to Uber, but replacing the cost of paying drivers with the cost of routine HD mapping updates.
2. Smaller towns have less frequent construction, so the update costs might be lower as you target less dense areas.
3. I could see a single company that specializes in providing routinely updated maps to a variety of fleet-operating companies. This could potentially be a utility or somehow subsidized by the government. It would also be possible for government to coordinate construction with HD mapping updates. After all, by lowering the rate of accidents and decreasing square footage devoted to cars, governments have a vested interest in seeing robotaxis replace human-owned and driven cars.
> Tesla lacks the ability to resim counterfactuals with confidence since they don't have HD ground truth.
Tesla does have HD ground truth data for verification generated by their own LIDAR-equipped vehicles. However, according to a recent tweet by Elon Musk [1], they don't need LIDAR for that anymore.
> That's simply untrue. All the hard stuff continues to be reliability and sensor gated.
IR and thermal sensing are unnecessary if the bar is human level and neither is the lidar. The point is overused, but humans rely on two eyes in the driver seat. I don't see any evidence to suggest the modern model that Tesla has developed for their vision system is their limiting factor in the slightest to reach L4/L5.
Dogs jump into the road in front of cars all the time and get killed, and kids get endangered at school bus crossings. That's a reality of life that robotaxis do not need to solve.
That vision-only argument is marketing spin from Tesla. The biggest thing it leaves out is that humans process their vision input with a human brain, which Tesla vehicles very much do not have. If and when we create true AGI they will have a good argument, but a world where that exists will be wildly different from our current one and who knows if Tesla's tech will even be relevent anymore.
Why are you so confident that AGI, or a human brain, is necessary to be able to drive a car with only cameras?
I get annoyed with statements like this because technology changes and advances so quickly, and Tesla has made substantial technical leaps in this field of machine learning. They have the state-of-the-art vision -> voxels/depth models and are only improving.
Tesla, who use cameras only, have not demonstrated full self driving, despite trying for a decade. Elon Musk has stated "It is increasingly clear that all roads lead to AGI. Tesla is building an extremely compute-efficient mini AGI for FSD" [1]
Waymo, who use additional sensors like lidar, have a driverless taxi service which needs no safety drivers.
Waymo does have safety drivers, they're just driving the vehicle remotely when it's in certain areas instead of being in the vehicle. So it isn't "full" self driving either.
> Tesla, who use cameras only, have not demonstrated full self driving
There are entire youtube channels with hours of continuous video showing Teslas driving around SF, but also other parts of California, with no human intervention.
No, Waymo is not driving remotely. Remote operators can only answer simple questions. They're at the point of commercialization so it's all about unit economics. There's no point in driving remotely especially since it does not scale cost-wise.
Waymo is geofenced, but within its geofence it requires zero human intervention. Tesla on the other hand is famous for mistaking the moon for a traffic light. Saying "Tesla has so many miles on YouTube" is hilarious because first of all there are channels with lots of Cruise & Waymo footage too, and more importantly it's not the # of miles that matters, but the # of non-trivial scenarios you can handle.
I don't see why Tesla can't handle those scenarios if they also use remote operators. I wouldn't be surprised if they do.
Btw Waymo is nowhere near achieving unit economics. Their cars cost like 5 times what Teslas cost, and the sensors require a lot of upkeep and maintenance.
Who’s to say what unit economics they’ve achieved but I’d hazard to guess that their investors wouldn’t support expanding their fleet and service unless the unit economics are at least close to break even. Cost for sensors and overall BOM keeps going down as more suppliers enter the market.
Are you saying that there are times when a Waymo car's ability to respond to events is at the mercy of a random Internet connection? What happens if the safety driver is steering remotely, from another town, and there's packet loss for a couple of seconds in the middle of a curve?
Again, they don’t drive or steer remotely. What sometimes happens is a multiple choice question is presented to the operator in an ambiguous situation:
<photo of construction zone>
Can I drive through here?
[Yes] [No]
When this is happening, the car is stopped and lets the passenger know that it’s reaching out for remote help to figure out what to do. For me this has happened two times across my 125 Waymo rides (571 miles) so far, and was resolved in under 20 seconds. Though I must say, 20 seconds feels like ages when you’re in the car and blocking traffic!
Eyes which are orders of magnitude capable than the best cameras, and Teslas come with mediocre cameras, not the best. Eyes which are connected to a brain, and ML is a looooong ways from rivaling that.
> That's a reality of life that robotaxis do not need to solve.
Robotaxis do not need to account for things jumping out unexpectedly in front of them?
I am not sure that the vision in Teslas is adequate with -any- amount of processing to drive a car. Spatial resolution is limited, as is seeing distant vehicles during merges, etc.
Secondarily, there is no guarantee that the amount of processing is enough, because the extant human systems use much more.
“Cheating” by using more sensors to simplify out complexities and to cover for the shortcomings of other sensors in the suite seems wise.
Orders of magnitude more capable than the best cameras? I wish. I need corrective lenses for my eyes to even work at all. With that fixed they feed my brain an image that's upside down, black and white except in the centre, which is covered in blood vessels and which has a blind spot. They also take a long time to adjust to sudden changes in lighting conditions, don't do any true depth sensing, suffer frequent frame drops and can't run for more than about 20 hours at a time before they basically stop working.
My brain tries to hides all this from me, and makes me think that I see the world in glorious 3D technicolor all the time, but that's a lie as revealed by the many amusing optical illusions that have been discovered over the years.
Meanwhile, today I used ML that knows more than me, can think and type faster than me, which is a much better artist than me and which can read and react far faster than me to visual stimuli. Oh, it can also easily look in every direction simultaneously without pausing or ever getting distracted or bored.
Somehow it doesn't feel like I have a big advantage over computers when it comes to driving.
Are we talking about Tesla's cameras or the "best" cameras? There are smartphone cameras that do depth sensing and HDR, and cameras are cheaper than eyeballs so composing them to get more angular resolution seems OK.
ToF/structured illumination cameras are honestly not that capable.
The maximum dynamic range of the eye is ~130dB. It's very difficult to push an imaging system to work well at the dark end of what the eye will do with any decent frame rate.
It's not as different as it used to be, but even so: the Mk. I eyeball does pretty damn well compared to quite fancy cameras.
> There are smartphone cameras that do depth sensing and HDR
Depth sensing is again, estimated or using time of flight sensors which is pretty much short-range lidar. HDR is used already in AV perception, but still loses to your eyeballs in dynamic range and processing time.
Eyeballs have high dynamic range but with high mode switching times. Walk from a bright area to a dark area and it'll take seconds for your eyes to adjust. Cameras are so cheap you can just have a regular day camera and a dedicated night vision camera together, switching between feeds can be done in milliseconds.
Robots aren't humans. You need accurate depth perception to maneuver a robot precisely, and you need ground truth depth measurements to train learned depth perceivers as well as to understand their overall performance. Humans learn it by combining their other senses and integrating over very long time using very powerful compute hardware (brain). To date, robots learn it best when you just get the raw supervision signal directly using LiDAR.
> Walk from a bright area to a dark area and it'll take seconds for your eyes to adjust
You do realize cameras have the same issue, and that HDR isn't free / is very computationally intensive?
Your brain is _really really_ good at surmounting challenges including many that you did not mention. We don't know how to get close to this in terms of reliability when using cameras and ML alone. Cameras and ML alone can go very far, but every roboticist understands the problem of compounding errors and catastrophic failure. Every ML person understands how slow our learning loops are.
Consider that ML models used in the field have to get by with a fixed amount of power and ram. If you want to process time context of say 5 seconds, and with temporal context 10Hz and with resolution 1080p, how much data bandwidth are you looking at? Comparing what you see with your eyes with a series of 1080p photos, which is better? Up it to 4k: how long does it take to even run object detection and tracking with a limited temporal context?
Your brain is working with more temporal context, more world context, and has a much more robust active learning loop than the artificial systems we're composing today. It's really impressive what we can achieve, but to those who've worked on the problem it feels laughable to say you can solve it with just cameras and compute.
There are plenty of well respected researchers who think only data and active learning loops are the bottlenecks. In my experience they're focused on treating the self driving task as a canned research problem and not a robotics problem. There are as many if not more respected researchers who've worked on the self-driving problem and see deeper seated issues -- ones that cannot be surmounted without technologies like high fidelity sensors grounded in physics and HD maps.
Even if breadth of data is the problem and Tesla's approach is supposedly yielding more data -- there is also the question of the fidelity of said data (e.g. the distances and velocities from camera-only systems are estimated and have noiser gaussians than ones generated with LiDAR). If you make what you measure, and your measurements are noisy, how can you convince yourself or your loss function for that matter that it's doing a good job of learning?
It's relatively straightforward to build toy systems where subsystems have something on the order of 95% reliability. But robotics requires you to cut the tail much further. https://wheretheroadmapends.com/game-of-9s.html
Agree 100%. And IMO it is worth remembering that a really significant share of collisions are caused by well known risk factors. For those of us who avoid being in those situations to begin with, the robotaxi would need to be a good bit safer than our average.
> I don't see any evidence to suggest the modern model that Tesla has developed for their vision system is their limiting factor in the slightest to reach L4/L5
For one, frame rate and processing rate on human eyes is way higher than cameras. Dynamic range is another. Also, Cruise and Waymo are some of the only companies that have hard internal data / ability to simulate how well their safety drivers do, and in the very same scenario what their software driver will do. Without LiDAR you can't build that simulation, and once you have that data if you continue to use HD Maps and LiDAR there's probably a good reason.
> Dogs jump into the road in front of cars all the time and get killed, and kids get endangered at school bus crossings. That's a reality of life that robotaxis do not need to solve.
Robotaxis need to avoid any accident that a human would be able to avoid.
> IR and thermal sensing are unnecessary if the bar is human level
See, you could say this if you had some data that showed that incidents per X miles (when the vehicle is driving at night) is sufficiently low, + if the software passes some contrived scenarios to gut-check its ability to see in the dark with the necessary reliability. But you don't have that data, do you? Someone has it though :) and I'd argue regulators should have it too.
> For one, frame rate and processing rate on human eyes is way higher than cameras.
I don't think it's exciting to say that you must have theoretical parity with something to use it for this use case. Tesla's solution monitors ~6? cameras at once with accurate depth in each. That's 6x more views than a human can see. I wish people would stop comparing apples to oranges.
> Robotaxis need to avoid any accident that a human would be able to avoid.
I never said anything to the contrary. Animals get hit all the time, not just because a human wasn't paying attention.
> Tesla's solution monitors ~6? cameras at once with accurate depth in each
No, the depth is estimated. It's not accurate, at least not in the way you need for L4.
> I never said anything to the contrary. Animals get hit all the time, not just because a human wasn't paying attention.
I was just clarifying what the bar is. The bar is that avoidable accidents need to be avoided. Nobody will get mad if a plane crashes due to unavoidable circumstances (freak accident where two engines go out due to bird strikes or something). People will stop flying in the plane when it becomes clear that the airline is not doing everything it can to avoid fatalities.
> The only selling point of FSD (Supervised) is that it (can) work "everywhere."
I seem to recall Musk saying in the last couple years that "full self driving will basically require AGI." This appeared to me to be extremely honest and accurate, though I believe that in the moment he was trying to promote the idea that Tesla was an AGI company.
I guess the cars can and will update the mapping in real time ?
> at best in certain cities
If mapping a city is possible, so it's mapping a highway, even easier.
If cars do update the maps themselves, they require might just a couple of human-driven passes of the standard WWaymo cars on a highway to generate the maps.
The obvious question here is "why not both". Use mapping data where you can, LIDAR and other sensors where you can, and visual cameras when you must. There's no reason to limit yourself to just one input type. Elon claims that, sure, but it doesn't seem like a given at all.
1. Robotaxi is a better target than general self-driving because the human baseline is much lower for robotaxis (most people dislike their experience with uber, while most people think that they are a better-than-average driver)
2. Google took the high road on safety. The move-slow-and-dont-break-things DNA of Google (that hurts them in so many domains) is a golden asset in self-driving.
Tbh I love my experience with Uber. I know people who don't own a car because they think it's cheaper to use Uber. But you're right - I am an above-average driver.
Waymo/Google benefitted massively from regulation (and lack of regulation) early on when they hacked up the car’s code for a demo and caused an accident with injuries
I'm typically very skeptical of Tesla's strategy here, but to play devil's advocate for a moment:
Waymo has shown they can make robotaxis work, but the big catch so far is that it takes them a long time to open in a new city. They have several phases before they open fully, from what I've seen it seems to be: safety driver no passenger testing, safety drivers with employee passengers, driverless with employee passengers, limited rollout to paid passengers under NDA, wider rollout but with waitlist, and finally getting rid of the waitlist.
This means that hitting even all the major metro areas in just the US is going to take them a long time, let alone the rest of the world (or at least developed world). That does give Tesla some time to potentially catch up, since they don't seem to be bounded by geography in the same way.
Now, that said, I personally don't think Tesla's strategy is workable except maybe the very long term. Doing this with only vision seems like taking something that was already enormously challenging and making it nearly impossible instead. Their slow progress and inability to get their cars to avoid even basic errors frequently, despite near a decade of development now, I think points to this strategy just being bad.
> They have several phases before they open fully: safety driver no passenger testing, safety drivers with employee passengers, driverless with employee passengers, limited rollout to paid passengers under NDA, wider rollout but with waitlist, and finally getting rid of the waitlist.
It's certainly true that they need to do a bunch of extensive mapping for each city, but I don't think we should expect their roll-out speed in later cities to be as slow as the first couple of cities. Most of the stuff they are learning in the initial roll-out will generalize to other location; it's not all city-specific learning.
Well you can definitely bet it will be faster not slower than the first two, especially given the basic (i.e. shared/city-agnostic) engineering required and the policy component, which will get easier and easier with each city as risk aversion turns to FOMO.
My argument is based on theory. We know that a lot of the learning is facing unusual situations (trucks delivering traffic lights, etc) that can happen in many places. And we have some idea of how long the mapping takes.
As a potential customer, Waymo's careful approach seems much more appealing to me. I don't want to ride in a move fast and break things robotaxi when it's snowing in Chicago.
Same, though playing Devil's Advocate some more, I can certainly see why "everywhere all at once" sounds more appealing to many people than "incremental rollout to major metros". While I'm guessing that eventually Waymo will cover pretty much any paved public road, that's not actually certain, let alone when it would happen.
I don’t think every city is a brand new learning experience, there will definitely be takeaways that will speed up deployments in new cities. Plus, a lot of these deployments can happen in parallel so seeing them come online in 20 places at a time simultaneously doesn’t seem extraordinary.
There was an episode of The All In podcast a month or two ago. Friedberg brought up driverless Waymo being available in San Fran. Chamath hadn't even heard of it. He looked it up live and it blew his mind.
These guys are all about tech and couldn't believe there were companies ahead of Tesla, what do you think the normies know?
This is a good question. Will the robot taxi company beat the company that hasn’t made a single autonomous vehicle in the robot taxi business? It is hard to say
> is Tesla actually in a position to win here due to their manufacturing capability / current deployment of Tesla's?
Tesla has much lower costs. If they can beat Waymo on customer experience (better driving primarily, but also better in-car entertainment, better mobile app, and match Waymo's pricing) they'll win.
Waymo might have a regulatory advantage since a lot of politians don't like Musk.
> They only have a car that can drive with a human partner behind the wheel who's ready at any second to take over and prevent it from crashing.
Even that is a faulty system.
I really don't understand how people can think Tesla and Waymo are anywhere close to each other. Making a car drive itself is a much harder problem to solve than scaling production. If you look at what Tesla has, today, it is a vehicle that can sometimes drive on its own with the need for a human behind the wheel. The ability to remove the human from the front seat is the other 95% of the work. In systems like this where human lives are at stake, getting "mostly there" with FSD is essentially worthless.
The problems of logistics and scaling are well-known and well traveled. The invention of a self-driving car is not. Once Waymo has solved the issue of entering new cities and quickly gathering data to feed into their model, they've won.
Scaling production isn't the issue for Waymo, it's scaling and maintaining operational capabilities within cities at a cost that still lets them be profitable.
I think it's possible, but it's definitely less trivial than scaling production of more self driving cars.
Those go hand in hand. In theory the successful mass production of their vehicles should drop the cost so that they can turn a profit. The increased number of vehicles and ridership cover the staff to maintain them. As I mentioned elsewhere I would not be surprised if the plush leather seats give way to easily cleaned hard plastic chairs.
No, I'm saying the problem is scaling that's not just building and maintaining the cars themselves, but rather building and maintaining capability around the cars.
You are assuming Tesla is anywhere close to the capabilities of Waymo. Ignoring Elon and his history of hyping things far beyond reality, Tesla does not appear to have the equipment, data, or organizational culture to achieve what Waymo has done.
Waymo doesn't have to convince its vehicle owners to let strangers ride in their car unsupervised.
Yes there can be cleaning fees and vetting, but all it takes is one or two people puking in your Tesla, and you'll have no interest in providing a Robotaxi for Musk's Mission.
Tesla is rather good at mass producing their own vehicles so they are not reliant on existing Tesla owners providing their vehicles for the fleet. If the rumor mill is to be believed, they will also be making a dedicated robotaxi vehicle with no steering wheel so there isn't a wasted empty driver seat.
> Yes there can be cleaning fees and vetting, but all it takes is one or two people puking in your Tesla, and you'll have no interest in providing a Robotaxi for Musk's Mission.
Plenty of Tesla owners (and owners of all sorts of other vehicles including high-end sports cars and luxury vehicles) already rent out their vehicles to strangers on Turo. Listing them on Tesla's platform instead won't be a large change. A lot of owners might actually prefer it if it allows more fine-grained options (eg. no rides past 9pm, no rides more than 50 miles from home) instead of renting it out for 24 hour periods.
> Plenty of Tesla owners (and owners of all sorts of other vehicles including high-end sports cars and luxury vehicles) already rent out their vehicles to strangers on Turo. Listing them on Tesla's platform instead won't be a large change. A lot of owners might actually prefer it if it allows more fine-grained options (eg. no rides past 9pm, no rides more than 50 miles from home) instead of renting it out for 24 hour periods.
The entry bar for renting a vehicle on Turo is quite a bit more substantial than hailing an Uber. I looked at making my cars available.
You can restrict the age of the driver (anything vaguely high end I see required you to be 30+), require substantial deposits ($750), proof of your own insurance, mileage and location limitations, and more. Versus "Downloaded the Robotaxi app and put in a credit card number".
People aren't even entering a credit card number, they're just using Apple Pay!
The Tesla-owned fleet will likely have the same requirements as riding an Uber. If the rumors are to be believed, those will be dedicated robotaxi vehicles with no steering wheel.
Of course for peer-to-peer you will be able to set your own requirements based on what you're comfortable with, or even just not join the network at all so your car stays exclusively yours.
I really don't think robotaxi's are viable with just consumer grade cameras. Lidar's are what make them truly safe. Aka: tesla's training data is garbage.
Alphabet is far ahead of Tesla in the category of "deploying a taxi network". No one can dispute that. They also use a different technology. What I don't know today is how fast can Waymo scale to more cities. I assume if Tesla cracks the "taxi network nut" they can scale faster and will catch up to Alphabet.
Tesla seem stuck-ish to me. They do have some incremental improvements each year, but even after several years of development, their cars want to randomly run into parked cars and other stationary obstacles on a frequent basis. We're not talking about edge cases, your cars shouldn't be regularly trying to hit a concrete wall after this much engineering effort.
Waymos do occasionally screw up, but if they did it as much as Tesla's FSD, it'd be chaos in the streets in SF, so it seems like it must be fairly infrequent.
I'm not sure how true this is anymore. FSD has improved significantly this year that they're on their new NN architecture.
It's worth remembering that Waymo required their users to sign NDA's during beta, while the Tesla FSD beta was open to everyone with no NDA. So there was a lot more Tesla content being posted and going viral.
I've heard "Autopilot/FSD has improved a lot recently" or "the next release is going to be a huge improvement" many times from Tesla superfans. And certainly there has been improvement, but it's still at a stage where it's making very basic mistakes in operating the car.
It's not even at the point where the challenge is handling weird edge cases with construction or strange intersections, it's still struggling with not running into parked cars and walls. How many years do you think it should take a self-driving system to be able to handle those basic tasks in the general case? Because Tesla has been working on self-driving for almost a decade at this point, and they still seem to be barely past the starting line.
> it's still at a stage where it's making very basic mistakes in operating the car.
> it's still struggling with not running into parked cars and walls
Based on the phrasing of your sentences, I must ask which FSD version you are running, and if you can share footage.
> Because Tesla has been working on self-driving for almost a decade at this point, and they still seem to be barely past the starting line.
Waymo/Google has been working on this since 2009, which was itself based on a Stanford project (whose team was hired from Stanford to Google) that started in 2004. So that's either 15 or 20 years depending on how you count, and it doesn't include the much harder tasks of mass producing vehicles and making electric vehicles commercially viable.
The NDA's ended when they were confident enough in their system that the good press would outweigh the bad.
So essentially we're comparing footage from Waymo when it was at the end stages of its development to footage from Tesla at the early stages of its development.
I don't know if they're wildly different at this point. Sitting in the shotgun seat, comparing the latest FSD vs the latest Waymo, on the same pickup and dropoff in San Francisco, I couldn't tell much difference. On the one hand, Waymo definitely chooses slower, quieter roads, and weird pickup/dropoff points - which means it's a slower ride. On the other hand, most people don't have access to actually try FSD so they rely on videos which are typically older FSD versions and spliced to only show "highlights" instead of being a raw 20 minute ride footage video.
I don't think we'll actually know until Tesla has an actual robotaxi product. When Cruise had one, most people who had tried both Cruise and Waymo said Waymo was better. That was my opinion as well.
> Sitting in the shotgun seat, comparing the latest FSD vs the latest Waymo, on the same pickup and dropoff in San Francisco, I couldn't tell much difference.
Well, except for the fact that one is doing it completely driverless. And it has to do that every single time without having the luxury of a driver to prevent accidents.
Big difference in reliability, which makes them wildly different.
There were no interventions, so both of them were doing it completely driverless.
We can't make an apples to apples comparison until Tesla also has a robotaxi product, but even then there will be questions around the role of remote operators.
> There were no interventions, so both of them were doing it completely driverless.
Well, no. A Tesla doesn't operate without a driver's supervision, so it can't be driverless. It did that particular drive without intervention, that's it. The stats [1] clearly show it's nowhere near capable of doing it without a driver in the seat. Community tracker puts them at 30 miles per disengagement.
FWIW a quick google search turns up Waymo reporting they have 0.41 incidents with injuries per million miles driven [0], whereas Tesla vehicles using autopilot had 0.152 incidents with or without injuries, per million miles driven [1].
So Waymo has 2.7 times more incidents with injuries then Teslas using autopilot have incidents, with or without injuries.
Maybe if I checked more sites they'd give different numbers, but from those initial numbers it seems your perception of reality of Waymo "screwing up" less is not accurate.
This is a ridiculous, apples-to-oranges comparison. You’re comparing fully driverless miles to driver assist miles with humans actively preventing accidents without controlling for any variables.
This is an extraordinarily disingenuous comparison. A big reason why Tesla superfans have such a poor reputation is because of bad faith arguments like this that frequently pop up in these discussions.
Tesla cars with FSD have a driver behind the wheel who can instantly take over if the car is about to crash into a stationary object. Any time a Tesla would've crashed into something an object but its human driver saved it, that doesn't count in stats like these. Many Tesla owners have reported that they have to regularly disengage FSD because it's trying to do something dangerous or looks like it's headed for a crash.
In contrast, Waymo cars do not have a human who can take the wheel if they try to run into a wall. The closest equivalent is that if Waymo cars get confused and don't know how to proceed, they can stop, then phone home and ask a human navigator to give them 'advice' or a general path; these people don't directly control the car, they're more comparable to a human navigator in the front passenger seat. It's still human assistance obviously, but it's not gonna save the car from running into an object that it didn't think was there.
> Many Tesla owners have reported that they have to regularly disengage FSD because it's trying to do something dangerous or looks like it's headed for a crash.
With Tesla the responsibility is on the person in the driver's seat, so there is a (rightfully!) a bias for overreaction on the part of the driver. We will never know many of these disengagements were necessary.
The only way to get a true comparison of data is to compare robotaxis with robotaxis.
It's true that not all of them would be crashes, but many would be, because, well, the car was about to crash. The car isn't just joking around when it swerves towards some parked cars.
> The only way to get a true comparison of data is to compare robotaxis with robotaxis.
100% agreed. And so far, Tesla hasn't taken the step of actually letting the cars be driverless.
Older versions of Tesla FSD tended to make steering adjustments that were short in duration, but at a higher turn angle. Human drivers in a similar situation would turn the wheel slightly but keep it turned for longer before returning to centerline.
People saw the steering wheel turn and perceived it to be the system going haywire, or thought that "the car was about to crash" as you put it, and intervened.
The newer NN based FSD acts more like what a human would do.
Yes, every year Tesla fans talk about how much it's improved, and every year it's still failing on basic driving tasks.
And there's definitely cases where the Tesla in question just tries to run into parked cars or similar for no apparent reason, but Tesla fans always have some excuse about why that's irrelevant, especially if it's not on whatever the absolute latest version is.
Then they accuse the people horrified at Teslas making basic errors and trying to crash of being "anti-Musk".
I actually watched the entire video in the article.
There were some private driveway situations where the uploader intervened to back out to go to a new destination (but Waymo drops you off half a block away and makes you walk instead of entering your driveway, so it's not possible to compare). And there were some situations where a human driver honked - this has happened to me in Waymo as well. There was one situation where the Tesla didn't seem sure if it could proceed, but Waymo in that scenario would ask a remote operator (this has happened to me in both Waymo and Cruise) and presumably Tesla robotaxi can also have remote operators.
The only case where he actually disengaged was at a stop sign with a slip lane, and the car turned right at the stop sign instead of turning right using the slip lane. He went there again at the end of the video and the car used the slip lane. I don't see this as an unfixable problem, because clearly the car can use slip lanes to turn, it just needs to be taught to always prefer slip lanes when turning.
So, your own video disproves what you're saying. It isn't failing at basic driving tasks.
I think the mistake you're making is assuming that they will never be good enough. A lot of people said the same thing about Google/Waymo until they actually rode in one.
Tesla is willing to sell to people who will pay $$$ for a self driving car. Waymo isn't. That's probably more important for now. Taxi drivers don't earn that much, and have some advantages AI can't easily replace (able to help with luggage, use petrol stations etc). Replacing them requires undercutting them which in turn means you can't generate a ton of revenue from that. Yet Waymo's business model, such that it is, has put them many billions into the red already. I wonder if anyone has done some ROI calculations and if so how long it'd take. The LIDARs alone would require a huge number of trips just to pay them off, then you have the cars, the decade+ of enormously high software development salaries... if Waymo were another YouTube where it could hide amongst Google's other profitable businesses that'd be one thing. As a separate business with its own accountings, how long will it take until it's turned a profit?
> Tesla is willing to sell to people who will pay $$$ for a self driving car
The very low take-up rate of FSD during the trial period indicates that most people are buying Tesla's because they are arguably the best EV with the best charging network.
Because the trial period showed that FSD will routinely try to kill you, on the routes I take around San Jose (interstate and normal roads). But having solar power and electric cars is awesome, and the car is a lot faster and more fun than my Prius was.
Remember that the passenger cars are not the only thing that can scale; if you can automate the mapping and data preparation part of the process sufficiently, you may even be able to reduce it to mostly a matter of driving a few sensor cars around for a few weeks; maybe even cars that are adapted versions of your normal taxi vehicles, but with a human driver behind the wheel while you are mapping.
I would imagine that while Waymo's mapping efforts have been very human effort-intensive so far, they will be looking at developing this automatic map-making capability as a high priority for rolling out new cities. Scaling the rate of expansion is then mostly a matter of throwing hardware and compute at the problem.
1) Waymo has not "solved" robotaxis as a business. They are not profitable and the vehicles are not truly autonomous (the humans monitoring the vehicles are merely remote. We don't know how many humans are needed per vehicle.)
2) Tesla has zero even remotely monitored, let alone autonomous, miles driven. So no, there is no reason to believe Tesla is close to true robotaxis.
Really, you can't repeat this point enough. Tesla has zero experience in autonomous operation. Their vehicle has not ever driven itself any distance, under any circumstances. There is no reason to believe their software is on the cusp of a sudden improvement. They simply release new major version numbers that have different sets of flaws.
You are correct, Waymo has not solved the economics of robotaxis yet. However Waymo does have a huge head start on the solution. Waymo has been able to manage their scale growth to manage the cost of finding these solutions. It seems like a competitor that hasn't had that will have to pay a lot more to catch up.
as a Alphabet and Tesla shareholder, this is what is important.
The rate of innovation at Tesla > Waymo
The cost of building Tesla FSD = 1/100 * cost of building Waymo FSD
The cost of delivering Tesla FSD = 1/10 * cost of delivering Waymo FSD
Tesla has economies of scale. Waymo has all the details figured out. Waymo can never get to the scale of Tesla (it can never buy 5 Million FSD cars, while Tesla is delivering them every 2 years)
Mathematically, Tesla has an upper hand over Waymo and it'll play out as that.
Larry, Sergei are extremely poor capital allocators. Musk is brilliant (despite him being a narcisstic a*hole).
Larry/Sergei left Waymo at a limbo state because they don't think in terms of economics, just coolness.
Waymo is successful enough to not kill it, but also not a cash-flow positive to scale it up
Edit : Tch, Tch expected HN anti-Musk hate showing up in downvotes.
move fast and break things may be ok for social networks, but never ok for endangering the safety of others
to launch a 6th model of a marque, so well established, theoretically knowing a lot about cars, a model that dies when it gets wet, it's just embarrassing
"The cost of building Tesla FSD = 1/100 * cost of building Waymo FSD
The cost of delivering Tesla FSD = 1/10 * cost of delivering Waymo FSD"
This seems like one of the key assumptions, but is not proven out at all because Tesla does not even have a level 4 vehicle. So the cost of delivering one comparable to waymo is infinite right now!
Your other key assumption is "Waymo can never get to the scale of Tesla (it can never buy 5 Million FSD cars, while Tesla is delivering them every 2 years)".
Both the assertion in the first part and the second part seem like super-strange assumptions, and not obviously true at all, yet are also critical to your analysis.
Waymo could get to the scale of tesla. It may or may not be too expensive to do.
It could in fact, buy 5 million FSD cars. It may or may not be too expensive to do.
"Mathematically, Tesla has an upper hand over Waymo and it'll play out as that."
Or you know, if needed, Waymo could change?
It's funny to watch someone say "this one company will be able to adapt in every possible way to it's advantage, and nobody else can or will"
That almost never happens.
Your retort is then that you are getting downvoted because of anti-musk hate.
Have you considered that maybe you just don't have that good of an argument instead, and that your comment comes off as more of a tesla fanboy (regardless if you are) than a useful contribution?
I could write the literal opposite comment of what you did, in favor of Waymo.
That would not be a useful contribution either for the same reasons.
Tesla is still stuck at Level 3 while Waymo has been operating at level 4 for years.
If Tesla does manag to jump straight from level 3 to level 5, they have a chance to compete, but that seems unlikely. They also might move to level 4 and be able to expand level 4 coverage faster, but that still remains to be seen.
Waymo has years of experience with the other hard part of self driving taxis: actually picking up and dropping off people without a human driver.
Anti-musk partisanship frustrates me, but I suspect it is your fan-boy talking points that drive the downvotes of your comment
A fan-boy would never call Musk a narcisstic a*hole.
Technology progress is non-linear. Yes, Tesla is at L3 and Waymo at L4.
But, you completely missed the point of rate-of-innovation, which was why I made that as first point.
My GOOG holdings are 4x than TSLA. But in terms of who will deliver FSD at better margins, it's hands down TSLA. It's simple Math and Elon's obsession about cutting material costs & process to it's barest minimum. Waymo has no such discipline or culture.
What good is being able to produce cheaper vehicles at scale if they're not capable of providing the same service? Tesla has had years to catch up but still hasn't, and there is no proof that they can.
The main concern many observers (and I) have that is that Elon's insistence on not using LIDAR may mean that it's not possible to reach L4 with the current Tesla hardware stack, in which case TSLA can't even compete.
Any driver that makes this kind of move is typically in the top 5%ile of driving skills. Yes, it's an asshole and slightly illegal move, but the level of intelligence that needs to be applied for this move is all you need to know about Tesla's advantages.
I have been extremely critical of Musk's reliance on only cameras, but I'm impressed with the progress they have made because of the new architecture in 12.x series. Considering it'll be trained with 100x more compute, I'm willing to bet that Tesla will overtake Waymo's capabilities.
Yes, it's a speculation and you can disagree, but unless proven you can't tell it's wrong
> Yes, it's an asshole and slightly illegal move, but the level of intelligence that needs to be applied for this move is all you need to know about Tesla's advantages.
Every idiot can drive like an impatient asshole. Choosing not to do is is the true sign of intelligence. The car is literally driving itself at this point. What does getting there three minutes faster so you can doom scroll Instagram from the lobby instead of doing that while your car is driving itself do for you? Are you more important than everyone else and thus deserve not to wait your turn?
If it helps, Waymo's will run red lights, so it's not like Tesla's got a monopoly on being a bad driver.
The level of skills / intelligence required to make a smooth merge in these situations without impeding anyone requires superior skills and intelligence. Something people who don't make this move won't understand
I'd say these drivers are in the lower 5%ile because what usually happens is that they now block two lanes or almost crash trying to go into the corner together.
It was luck that the left lane was moving and that there was space. I don't want to see this kind of move from a driverless car.
It wasn't luck. There is always space between cars during stop-and-go traffic as cars take time to accelerate.
A skilled driver always makes the smooth transition into those gaps. I know most HNers are goody-rule-followers and can never appreciate the skill required to make that maneuver consistently. But, this is a clear example of separation of intelligence / skill and I'm happy with the bet that these are signals of intelligence
As long as it doesn't impede traffic, and make these kind of smooth merge, it isn't.
This is the difference between normies and first-principle-thinkers. They are being brainwashed into thinking all rules / laws are there to maximize total good for society
I don't know where you live, but here in Europe, the driver would lose the driving licence for 1 month for this little stunt. Specifically, changing lanes near the intersection over the solid line, and cutting those waiting in line.
What you call first-principle-thinkers, the rest of the world calls dicks. Everybody sees that manoeuvre, it doesn't require a genius. Most of the people don't do it because they are afraid of repercussions (if they get caught) or are civilised enough to realise that their time is not more important then the time of others. Yes, that manoeuvre doesn't "save you 5 minutes", it steals 1-2 seconds of everybody else's time.
That's why Musk's approach will win. Waymo will no doubt provide a safe, rule-following driving experience, but Tesla will have frontier breakthroughs and provide more human-like, adapting and sufficiently aggressive driving experience
The problem with Tesla's approach is that it learns from humans. Specifically, it learns the mediocrity of humans. That's why it made that idiotic manoeuvre, it learned it from the typical people that drive other Teslas.
So Tesla will necessarily converge to the performance of your average driver, because that's where the data leads it.
Waymo's is developing a new type of understanding and modelling of the world. It perceives and tracks items above the capability of humans to track and understand. Therefore its limits are outside of the bounds of the most capable human drivers.
Your own argument highlights exactly why Tesla's approach is a dead end (exactly in the same manners that LLMs will lead to the dead internet), while Waymo's approach will likely generate super-intelligence.
> you completely missed the point of rate-of-innovation
How is the "rate of innovation" higher but yet they've innovated less and have a less functional product? It's made up metric and was ignored because it adds no value to the conversation unless you have some actual data.
> Waymo has spent more money, resources and started earlier than Tesla and yet it is only marginally better than Tesla at this point.
Tesla isn't even comparable to Waymo at this point because Tesla has zero level 4 capability. Teslas don't even have the redundant sensors needed for level 4+. All Tesla has is empty promises and an u unclear path for ever getting past level 3.
Many people predict that AI is going to explode, and afterward nothing will be the same. If that happens, Telsa is in a better position than anyone else to simply update their software and deliver self driving cars.
I was in urgent need of a hacker to help recover my lost files and hack my partner phone. I met a hacker who scammed me of my money. I was scared of hiring another hacker not until i saw good recommendations about ( spyrecovery36) of his good works. i gave him a try cos i was desperate to get my files back and hack my partner phone, Good to see there are still great hackers like ( spyrecovery36 @ gm ail com ). i got my job done and can recommend him to anyone else. He also offer services like delete criminal record, recover wallet, track GPS, facebook recovery and many more...
Add freeways and airport rides, both of which they are very close to doing, Waymo will become much more of a complete service and a true Uber/Lyft replacement.
In a year's time, we could genuinely see them operating at scale in 6-8 major cities (SF, Phoenix, LA, Austin and new cities), especially with their new dedicated robotaxi from Zeekr. A possible hold up would be China import tariffs imposed by the US government.
1: Airport rides are already available in Phoenix between 10PM and 6AM.
2: I have never seen a Waymo on the freeway without someone driving it (not monitored self-driving, but physically driving it). Take this with a grain of sand but it is my understanding that full self-driving on highways is still far away given the limited range of the (relatively) small sensors and the speed of travel required. That is: the sensors cannot see far enough ahead to react comfortably when going above N miles-per-hour. That might be a dated understanding of the issue though...
Source: I've been riding Weymo since 2019 when it first went public beta (NDA-restricted use). Rode one last night coming home from the airport!
Airport rides in Phoenix are always available, but apart from the hours you note, Waymo drops riders off at the 24th or 44th Street Sky Train locations.
Freeway driving is more straightforward in many ways compared to urban driving, but the problem is that the impact of screwing up is a LOT more severe.
If you screw up at typical speeds in SF, probably nobody will die. On a freeway, odds are a lot higher.
If a car is confused in an urban area, stopping in the middle of the street isn't 100% safe, but it's not super dangerous, whereas stopping in the middle of freeway suddenly is obviously extremely dangerous.
> That is: the sensors cannot see far enough ahead to react comfortably when going above N miles-per-hour. That might be a dated understanding of the issue though...
It's more of a challenge when your stopping distance is low (more problematic for trucking than it is for TaaS vehicles since trucks are so damn heavy).
I suspect some challenges are risks like tailgater collisions and high-stakes lane changes at speed, and even more problematic is just profitability (what if your cars are tied up in traffic jams, what if they don't have enough charge left to make a long round trip, etc).
I meant airport rides to the terminal at all hours of the day.
Waymo is already giving driverless freeway rides to employees in Phoenix, so I don’t agree that it’s far away. It’s definitely a different challenge though.
I know from their announcements that they give freeway rides to employees in Phoenix area. They're also frequently spotted in Bay Area and Austin freeways, but with safety drivers.
If these vehicles tackle the roads of Boston, there will be no stopping this company. That place has one of the most confusing road networks I’ve ever seen in addition to some of the most confused drivers.
NYC is another obvious challenge but Boston seems like a challenging middle ground.
The issue is not a confusing road network, it's that Waymos require a modicum of grace in other drivers to operate. They aren't super aggressive, they'll defer to pedestrians and other drivers, they drive safely.
That I think is where they'll find problems when expanding to locales with less than conscientious drivers.
This is both amazing and horrifying. I'm actually confident this automation will save lives. Well of course any system can fail, Uber drivers are often distracted by 30 things, they're fiddling with the app, on personal calls, while navigating tricky traffic situations.
However I predict within a decade or so we're going to get to a point where gig work is no longer feasible. It'll take a bit of trickery, but I'm sure you could have restaurants opt in to putting their own food in the backseat of these. And then as a consumer you would just get your own food from the car .
So think about every delivery driver, and every Uber driver, and many other gig workers. All of these people are going to be out of work very soon. Plus tons of creatives will be replaced by AI. AI will reduce the need for junior software engineers .
I don't think the modern economy is ready for this. If I had one wish, it would be to at least decouple employment from health care. As is, let's say you have a serious illness that requires you to resign or otherwise not have employment for an extended period of time. You're now stuck with a serious illness and no health care. Depending on the state unless you're a child or parent you're not qualifying for Medicare period.
Has anyone figured out, who exactly gets sued when one of these Waymo's hits someone.
Gig work is already not feasible. The only reason anyone undertakes it is financial illiteracy. Uber is largely funded by the irrational sacrifice of numerous individuals of the residual value of their own cars.
When the option is a choice between going possibly hungry for the day, you and your family, and not going hungry then that is not called financial illiteracy.
Now, we're talking. Often, there will be handwringing about low-wage jobs from people who wouldn't dream of working low-wage jobs and who benefit from the existence of low-wage jobs. Often, that will be invoked as a rebuke to automation. But, automation in general isn't a threat to workers or well-being. In fact it's productivity gain which benefits well-being by lowering costs and raising living standards. If a portion of those productivity gains aren't used to help the people they displace, that's a matter of policy not of technology.
Oddly enough people are self-regulating, so few people are having children now because modern economics makes it a bad idea for most people .
In my grandfather's time you could work a normal job and buy yourself a three-bedroom house in what's now one of the most expensive cities in America.
If I wanted to buy the same house at the same city, me and a partner would both have to make 150k to 200k each. And even then we'd barely be middle class, we wouldn't really be doing exceptionally well.
Homeownership shouldn't be regulated to the top 5% of income earners.
However I don't realistically imagine a tech utopia. I think we're headed towards unholy levels of income inequality. And income inequality isn't a problem by itself. Hypothetically if Billy Bob makes 60k a year, but he owns a house worth 400k and he's able to leave it to his kids, that's just fine .
The problem emerges when he's renting his home from a mega corporation that raises rent by 15% per year, until he ends up either homeless ( unhoused, shelter challenged, pick your semantic softening) or sharing a room with 2 other people.
It's a complicated problem with no clear solution.
If you're the age I think you are (mid-30s), then during your grandfather's time most women didn't work and many racial minorities were barred from many jobs and forms of housing. If you cut the labor supply in half or more, then wages naturally go up because demand doesn't change that much per capita.
That said I completely agree with you about everything else. I think the US needs to firm up its social safety net (hopefully through UBI, though I'm not hopeful) ASAP. I don't think AI will take over the world or even remove all jobs, but it will remove enough in our lifetime that we'll have a crisis of low-skilled employees out of work who just won't be able to transition into a job that pays well enough to pay four healthcare, housing, and transportation. Housing itself is a really hard problem too for various other complicated reasons.
The biggest issue is zoning and other restrictions on construction.
Back in the 1950s, it was much easier to build. Now it takes much longer and it's more expensive to just get permits. It's kinda wacky to think we're still using housing stock from 100 years ago. New construction is rare.
A middle class person, say a Firefighter needs to be able to buy a home where he works. Many city workers in California have to ride long Metrolink rides just to get to work. This is also a safety concern.
Do you really want a subway conductor to have to get up at 5am to get to their 9am to 6pm shift?
This has been an issue for decades, but it's coming to a head now.
100%. Our local bus system can't hire operators because they're priced out of the area for housing, so all operators have to come from a 4 hour Metrolink ride away.
I feel you. Thing is, I think there is a solution or at least part of one, on paper. We know what to do. Trouble is, we can't do it because of politics, politics that transcends the usual divisions. It's hard to have a tech utopia and deep income inequality without policy to make it that way. Policy creates markets, and markets create and distribute wealth. How that's done, whether evenly or unevenly, is determined by politics, policy, and the market structures they create.
Patents, copyrights, subsidies, ZIRP, foreign policy, etc., even without reporting to value judgments of "good" and "bad" and sticking safely with neutral terms, all these things involve trade-offs. They have consequences. They do tend to pick winners and losers.
The point is that the shape of society is a matter of choice not chance, technology, or divine providence.
Couple of anecdotal data points from a new rider. Here's the first time rider spiel they give you [1]. Overall experience is pretty polished. It's handled some tricky situations with confidence like this left hand turn from a two way stop on the inside of a blind curve [2]. Then other times it drives like the most cautious newbie driver. Maybe it didn't like me being so closed to the road, but this time it awkwardly failed to pull close enough on pickup [3]. Definitely got some stink eye from drivers trying to squeeze by.
While it’s still a drop in the bucket compared to human driven taxis, it’s remarkable that Waymo will like reach 50 million passenger only miles this year. And will surpass 100 million passenger only miles sometime in 2025.
With that much data the safety case should become very clear.
If you mean actual taxis in San Francisco, it's not a drop in the bucket. There are 1,800 taxis and compared to 300 Waymos, and the latter have a much higher duty cycle. It's true that the number of Uber/Lyft is a lot higher, something like 40k drivers (who work a widely varying number of hours per week).
Yes, the 100M miles scale is very important, because that's about how many miles humans drive until they cause a death.
Except that they have remote humans monitoring every vehicle so the whole thing is an illusion and we don't know the truth of how safe truly autonomous vehicles are (since they don't exist.)
1:1 remote human monitoring would not scale from a unit economics perspective, and even if they did that, the remote operators can't drive the car, only offer small feedbacks. So the car is really driving itself.
The safety story is an interesting one. Companies like Cruise and Waymo are not forthcoming with their incident data. They share infrequently and through spreadsheets that do not capture every incident. It's pretty ass, and I'd be wary of trusting their self-reported data. I imagine their insurance companies have slightly better data than the gov't, but even then maybe not.
Spreadsheets are the best and most open way to share data because it lets other people analyze it. But they put out plenty of their own research if you want to read that.
I'm really not sure what you mean by "infrequently". They release new raw data every 3 or 6 months or something, and every single accident is trumpeted in the news. What other industry has more publicly accessible safety data?
Every software release is a totally new driver. If they release software at a cadence of say, 6 or 8 weeks, would you feel comfortable riding? You don't know how safe the car was during that time -- any part of the driver can regress. You're basically trusting that they know how to do simulation, that they bore the cost of running that simulation, and that their simulator is realistic enough to yield trustworthy statistics.
Sadly, our regulatory agencies are currently set up for very delayed decisionmaking. Since every company can release software at any time, you could imagine a regulatory platform that tracks software releases and their corresponding safety statistics in real time.
I’m quite confident in it and personally ride it monthly. Happy to bet it will turn out to be no more than twice as dangerous as the median driver today, ie, as dangerous as driving in the 1980s. In the long term it will be vastly safer, obviously.
I'm glad you feel that way (and I mostly feel the same / ride Waymos frequently), but it doesn't change the reality that you're putting all your faith in a self-regulating company that does not cooperate with regulators more than it absolutely needs to, and where a change in leadership / safety culture / process can cause a catastrophe overnight. And this isn't hyperbole; 1 software push is all that separates a car that crashes in scenario X from one where it was found to not crash in that same scenario. Do you trust the code to be bug free? Do you trust the bugs to be discovered before they maybe kill you?
I have a vested interest in the success of AVs (worked in the industry for the better part of a decade) but the companies building them are far from perfect in terms of how they conduct themselves wrt safety. There needs to be some development in terms of the supporting tech.
The way I think about this is: planes came before the FAA and key technologies that we use to manage flights, such as radar and radio being used in conjunction. AVs are in their "just learned to fly" era.
In the next couple years Waymo will have accumulated more than 100M miles driven, which is the number a human drives statistically before causing 1 death on average. Are you still going to be worried when they drive 1B miles and have 2 deaths? What about 100B miles with 30 deaths? At what point will you accept the statistical evidence?
A new software release has driven zero miles. You do not know what kind of performance it has -- regressions both happen and are expected with every launch.
The number of miles a release ends up driving is mostly in simulation + some human-supervised on-road testing, and may be low / influenced by the availability of resources ($$$). There are little to no regulations or 3rd party checks on the process used to clear releases.
They do not simulate every release on all X miles driven. They can't, it's super expensive, and that's to speak nothing of challenges like simulation realism or compatibility of the simulator with multiple generations of hardware that isn't backwards compatible.
They could have no deaths for years, then suddenly a whole lot after a bad software push. Recently they issued a recall because of a bug where the car could hit a telephone pole. Imagine if the software had a bug when their cars were ubiquitous on highways, slamming cars into concrete medians? Imagine if a fault could bork critical technologies like the lidar at a high rate across the fleet?
What happens to Waymo if they kill even 1 child due to a bug or lapse in model quality? I imagine many would wake up to the reality of their safety story and demand a way to audit software releases. Or would you continue to let your kid ride Waymo? This is also to say nothing of sensor failures and other ways the car can mess up.
The probability of your dying in a Waymo is not # incidents / # miles, because past performance does not guarantee future performance when it comes to new releases. It's unknown quantity L (likelihood of encountering a dangerous bug per unit time) * T (time spent in car). Without more data about L (which changes based on many factors, like the company's process for certifying releases), L*T could go from 0 to something unacceptably high really fast.
You trust the car you drive today because the parts are quality controlled and regulated. Software drivers are a very new part added to this chain of things to trust. Their quality control process is not yet publicly regulated. The reality is there needs to be an advancement in technology to ensure software quality / safety, ideally from the role of an unbiased auditor that can catch faults and ground the fleet asap, then restore public faith by ensuring gaps in the qualification processes are fixed. No such mechanism exists today.
> They could have no deaths for years, then suddenly a whole lot after a bad software push.
The software release process itself becomes the thing that is tested statistically. If they release 100 versions, and each is safer than the last, it’s silly to think that one can’t be confident in the safety of the 101st without some gov’t software approval/validation process.
> What happens to Waymo if they kill even 1 child due to a bug or lapse in model quality? …would you continue to let your kid ride Waymo?
If they had driven billions of death-less miles and then a software update killed a child, I would obviously let my kid ride Waymo. It wouldn’t be close.
> If they release 100 versions, and each is safer than the last, it’s silly to think that one can’t be confident in the safety of the 101st without some gov’t software approval/validation process.
That's just the thing. Each release is not always safer than the last. Even ascertaining that with any reasonable level of certainty requires a lot of technology. In a company's hands alone, this "trust me bro" amounts to a ticking time bomb since companies have a tendency to trade safety for profit.
> If they had driven billions of death-less miles and then a software update killed a child, I would obviously let my kid ride Waymo. It wouldn’t be close.
You don't think the outcry at Boeing applies to Waymo?
I feel like this concern is mostly a thing having to do with safe rollouts. Just as with any other software, I'm sure there's a gradual rollout to ensure bugs don't just kill a ton of people at once. Google and other large cos have largely solved the reliability problem with web services. while cars are ofc different i think the same learnings apply
> while cars are ofc different i think the same learnings apply
They are very different. Unit tests and stress tests can be conducted on pure software. You can monitor them and if there are failures you can retry.
Robot software needs all the stuff you use for regular software, but then you have this huge space of counterfactuals that can only be verified using large scale simulation. How do you know the simulator is realistic? That it has a sufficient level of scenario coverage? That the simulation was even run? That the car didn't catastrophically regress in a way that will lead to a crash after 1M miles of driving (which in a scaled service will happen quite frequently)?
Oh and a fault can result in a death for Waymo, but not for Google Search. So that's kind of a major difference.
We don't know how often the humans interject. That's my whole point. It's an illusion that the car is operating alone. When I drive a Tesla on FSD i only need to interject periodically, but it's enough that the car cannot be called autonomous IMHO. How many remote human supervisors are needed for Waymo vehicles? How often do they interject? Without that data it is absurd to call Waymo autonomous.
Waymo operators do not ever drive the vehicle. My understanding is that Waymo operators can specify things like "take this path" (e.g., by drawing on a map or something) or "yes that's safe" but this doesn't correspond to the actual driving inputs.
I'd appreciate if they were open and honest about the reality of how they operate. But since they keep it very secret, I have no choice but to assume the worst. If it was impressive rathee than detrimental to their valuation, they'd be open about it.
"the autonomous driver can reach out to a human fleet response agent for additional information to contextualize its environment. The Waymo Driver does not rely solely on the inputs it receives from the fleet response agent and it is in control of the vehicle at all times. As the Waymo Driver waits for input from fleet response, and even after receiving it, the Waymo Driver continues using available information to inform its decisions. This is important because, given the dynamic conditions on the road, the environment around the car can change, which either remedies the situation or influences how the Waymo Driver should proceed. In fact, the vast majority of such situations are resolved, without assistance, by the Waymo Driver."
But let's assume they're lying and you're right: what kind of throughput and latency do you expect to be required between the car and the remote driver for this to work safely? How does the car know if the remote driver is actually looking at the data and acting on that? What happens if the connection stutters? Does the car take control again? How would it decide when to alternate between remote and local control? What if the two sides disagree on what to do?
One blog post lacking in any detail is not being open about how it works, and I think you know that. I was very clear in the kinds of detail I think they should legally be required to provide to operate on public roads.
I don't know that. You asked for numbers on how many supervisors per car are needed (I bet it's around 1:100, probably less) and how often they intervene (a few times a day at most? Otherwise you'd read articles about cars being stuck all the time). If you knew the actual numbers, you still wouldn't know HOW it works. The post does shed light on the actual workflow and gives you important details that you wouldn't glean from the numbers alone, e.g. that the car is always the one in control.
I remember almost 10 years ago, maybe 8 i first saw one of these up in the city. It was parked and the technician was outside of it with an ipad, and i happened to strike up a conversation with the gentleman. I was just curious and at first he was remarkably cagey, i was quite confused and just continued to be friendly and express hope for the projects success and for automation in general.
I got the feeling that man had been accosted many times by angry locals and I may have been the first to give a word of encouragement, he was very polite after the initial tension wore away and he felt my shared enthusiasm. He must have been one of the early engineers, I had never seen or heard the name waymo but I was aware google had been competing in the level 1/2 dessert tests.
The man was very friendly and i was surprised how his behavior must be a reflection of society's view towards technical automation. Seeing the videos of people kicking food delivery robots and now my own tendency to flip off elon musks tesla cameras all these years later I am starting to get why he was nervous.
Cheers to the future I suppose, but hopefully the future has less cars and more walkable cities.
In the future, when you're old and quite possibly disabled, you might rethink the whole "walkable cities" thing.
I mean, it's easy to say "just walk (or ride a bike)" when you're 22 years old and in prime health, but the population in most First World countries is rapidly aging.
"Walkable cities" doesn't mean that everyone is forced to walk everywhere. It means that there are destinations (grocery store, restaurant, pub, library, cinema, hospital etc) you can reach within walking distance without having to play frogger on a highway. They are absolutely a boon to someone old and disabled. They can use assistive devices, public transport, and even cars to get around.
I think you have the wrong impression about what "walkable cities" mean.
A walkable city doesn't preclude driving. In fact it often improves it. If almost everybody is walking or biking or using transit that means there are few other cars and driving becomes much nicer too.
Better driving can only get you from 1 curb to another. You still have to walk to your actual destination whether it’s to the other end of the mall, or to your home’s door. In no way is this any worse in a walkable city which will not only allow you to walk everywhere when you’re able, therefore delaying any loss in walking capabilities, but will also include better and more accessible public and private curb to curb transportation.
A walkable city will also make the curb to actual destination far more walkable.
A walkable city will also mean a lot more options are accessible to the disabled through their wheelchairs, etc.
The biggest reason people loose muscle mass as they age is due to lack of use, so a built environment that encourages walking is going to help keep people in shape to walk as they age.
NY is an AARP top-10 most livable city alongside other pedestrian-centric cities like Boston and San Francisco. Orlando and Tulsa conspicuously absent. Aging in car cities sucks.
Old and handicapped people have lived before cars, and in less car-centric cities forever. Walkable cities mainly mean that things are close to you so you _can_ walk to them.
Actual car-free areas, like the access-controlled dense old towns in Europe, are possible (if built-out where they don't exist), but not necessarily to be walkable.
> WAY more old people can still drive than can walk a mile round trip carrying bags of groceries
Groceries are not a mile away when you live in a walkable neighborhood. My grandparents could easily walk to the corner store to buy their daily groceries until they passed away. Chances are that walking a little every day for their whole lives helped them stay in better shape.
> I mean, if y'all like "walkable cities" so much, why not just move to one
The few walkable neighborhoods where I live are, unsurprisingly, highly desirable and thus unaffordable. In my forties I don't feel like immigrating yet again to another country.
> But no, everyone else has to conform to what you think is best
The immense majority of the large metropolitan area where I live only allows single family homes to be built. I would like walkable neighborhoods to be made legal again so that offer can meet the existing demand.
Don't you worry, car-dependent suburbs are not at any risk of disappearing in North America.
Many more disabled people are disabled in a way that prevents them from driving a car than in ways that prevent them from walking. There is also a huge class of people who are physically capable of driving but legally barred from it.
In fact when I scroll back through your recent comments I find it hard to find any that don't break the site guidelines with some sort of aggressive or nasty edge.
We have to ban accounts that post like this. This is not what HN is for, and destroys what it is for. It has also, unfortunately, been a problem for years:
I'm not going to ban your account right now because you've been here a long time, but if you want to keep posting here then we need you to properly fix this going forward. That means being respectful and curious, and it also means not being irritable, mean, or any of the other things that lead people into the flamewar style. So if you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
Yeah I wonder what people do, and have done, around the world in walkable communities predating the car. Perhaps people in these communities are on average more mobile into old age because they frequently walked?
They were (mostly) taken care of by extended multi-generational households. They obviously died and/or suffered to the extent that medical and QoL technology was insufficient. The whole status quo of pensioners migrating to Florida-style retirement community necropolis with a F-150 and a rubber stamped driver license is a modern US-centric phenomenon.
You are naming disparate & absurd things as the reason that European like walkable communities are bad for old people, like the difficulties of 1930's cities (or perhaps pre-industrial, you aren't very clear), the difference in landmass between the US and Europe (?) and the fact that there is a small range of disabilities that allow someone to drive and shop to Costco but not to go to a cornershop.
The reality is that being old sucks ass, but being old and sedentary (when you can avoid it) sucks worse. I've taken care of multiple senior family members & friends, the ones that were active, i.e walking everywhere, not avoiding stairs completely, have lived longer and happier. Sure, even active people gradually lose the ability to do day-to-day stuff and cars or microcars[1] can help them and other people with mobility concerns, but that's not an argument against walkable places.
Making streets non-hostile to pedestrians, having necessities closer to residences and prioritizing public transport makes people stop preferring cars as medium of transport if they don't need it (youngsters). This actually helps with traffic and in turn helps the people that need cars like someone with a rolling walker and arthritis or a delivery van.
Implying that people advocating for walkability are heartless youngins that don't care about old people just because you are losing arguments left and right, is actually really unempathetic.
> “When I first saw it before filming, it was all alone, no peds, no other moving traffic. So, I don’t think it was yielding for any of the peds, just stopped for the steam.”
This is so so huge and one where a tech ten years ago had an incredible interest and since then the companies started folding and only waymo survived. That's vision..10x
Do these need a significant amount of local infrastructure (parking, charging, etc) specific to Waymo that would make them moving into a new city (Austin is next) a big investment? Like will there need to be a big Waymo fleet center/office built in each city?
I'm not sure if these park and someone plugs them in or what, who maintains the actual Jaguars, etc.
Depends on your definition of “significant”. They require at least one depot in each city where they operate. The depot handles charging, cleaning, inspection, repair, and parking during low-demand times. Each depot is different since they are generally leased buildings or lots in light industrial areas. You can search on YouTube to see them.
Thanks! Found the perfect video. He talks about the depots here, there are some that are full service and some that are not. These are actually a lot bigger than I was expecting and have a bunch of staff.
I'm always struct by how much more slowly driverless cars has been deployed than what I expected when the first won the darpa grand challenge. I guess I understand the need for caution. By the time this goes nationwide, I expect it will be pretty damned solid.
To be clear: Waymo can take you to basically any address in those city limits, but it doesn't do so by taking the freeways like 101 and 280 that pass through it.
I've only taken them twice but I was surprised by how smooth it was. They're pretty defensive drivers, which is what you want, but were still able to navigate some tricky stop signs (e.g. a biker biking along stopped cars, pedestrians trying to cross in both directions) without getting totally paralyzed.
It's a really uncanny feeling when it pulls out into traffic for the first time and the car is driving itself. I have a Tesla, so I've played around with autopilot, but being in the back seat and not being responsible for the car at all is a crazy feeling. Legitimately one of the most "oh my god I'm living in the future moments" I've experienced in recent memory.
Pick up and drop offs get snapped to specific areas so I had to walk ~1.5 blocks on either end which wasn't a big deal for me, but could see how that might be annoying/difficult for others.
They are extremely nice compact SUVs (Jaguars), extremely nice interiors with displays (like Teslas) that show the entire surrounding area with visualizations of what the car can see, you can choose music and air conditioning, and the driving is extremely good — equivalent to a typical safe (albeit slow) Uber driver.
About half the Ubers I get in have overpowering air "fresheners" (perfumes). I'll always roll down a window unless the driver has disabled it (which is one more thing not to like about Ubers).
The big hope to me is the idea of cheap, cheap rides due to the lack of labor expenses, tipping and eventually gas. I don't see how that happens if there isn't competition, but I don't see how it doesn't happen if there are competing companies that can offer a similar service.
They've invested untold billions into getting this to work and now they need to recoup their investment. They'll get "cheap"
eventually, but don't hold your breath.
All it takes is one competitor in a hot space to get companies offering venture-subsidized prices, like we've seen when rideshare and delivery companies. And though the upfront investment may be huge, the price per trip can end up close to zero, unlike rideshares and delivery which have a much higher floor.
Not that there's any competitor that seems even close to Waymo.
Seems like a massive stretch to insert Tesla into the conversation.
Waymo has been building up to this for years in San Francisco. Cruise has been looking for a new CEO for many months. Rivian has been in well known financial trouble.
I don’t think anyone’s worried about a Tesla robotaxi. It’s another vaporware in their long list of vaporwares. Going by their FSD track record, that thing won’t exist for a number of years.
Thank you. I didn't know about that. That's a little vague, is there any more information? Like is this a special model of Tesla or a technology or software package for an existing model? Just wondering because I bet there are people here who have insight.
Does Uber regret shutting down its autonomous efforts? The problem has emerged as a tractable one. Or were they just too far behind and the capital expenditures too high to even remotely consider continuing? Does Uber exist in ten years?
How long before you have to stop and spend 5 minutes in a store on your route? Google is an advertising business, and can now literally drive traffic to your store.
The cost of the extra drive time would seem counter to that kind of thing vs just serving more drives. It takes next to nothing to e.g. show a business with more emphasis on Google Maps so cheap ads for emphasizing your business are palatable for customers to buy and Google to serve. I'd expect something more like that (recommended destinations, recommended side stops) where the showing the option and getting a decline has basically no impact to the operation but you still have a chance to get the bite anyways.
They've been in Phoenix for the past two years (generally available for the past year). They're so commonplace now, I don't think anyone really notices them.
Alternative Apple ID / Google account with region set in US. Always useful to have! For Android it's much easier, I have a bunch (for every country I've lived in) set up on my phone simultaneously and I can just switch in the Play Store.
i drove for uber around 7 years ago. it was the first time that i experienced first-hand a mass delusion. more on that later though. at the time i had just read the first samples of GTP output from openai and i immediately understood that we were probably going to experience an explosion of AI progress. and it weighed on me because i was afraid that the economic disruption caused by the resulting automation would ruin my life. i remember sitting in my car thinking that its completely pointless because they are just going to automate ubers… so today is a funny day for me because this really is the death of human uber drivers in SF, even if generative AI isnt at play here. that day has come where people who need a job to tide them over for a few years wont be able to drive ubers. one less option.
driving an uber costs about 20 cents per mile accounting for everything including gas, maintenance, replacement, insurance. including tips you are often paid around a dollar per mile. its just dead simple math. the idea that ubers lose money is a mass delusion fueled by outrage culture, echo chambers and the media. i made tons of money doing it. im screaming into the void and it wont change anyones mind
I dont get why anyone cheers shit like this. It just takes jobs away from more people. You will just get more homeless and drug addicted people. And more rich techies. And it will make no difference to the rider - they get where they need to be either way. It's just another bit wealth transfer from poor to rich. Yet this is cheered.
Before Uber taxis were basically unaffordable for me. You had to call them and wait in queue till somebody send you one. Check if they aren't cheating etc. I think now much more people are using them due to convenience. Robotaxis can possibly expand market similar how Uber have done it.
They solved the wrong problem. I still have to sit in a car for 30 minutes after work. If instead they made a robotic chef cook, that would actually save me 30 minutes every day.
I can't imagine this turning out well if there is any social media movement around these cars. It's reminding me of that one robot travelling across the country which ended up getting destroyed by some people.
They have been slowly expanding service in SF for several years. Tens of thousands of people have taken a ride. Actually, they give more than 10k rides a week, so probably over a hundred thousand people. Scaling up to the whole city (pop: 800k) is not that big of a jump.
Some people will get real mad, but eventually they'll become so commonplace that they'll be accepted in general. Nobody's gonna be smashing up the self-driving cars they see every single day.
i've always felt that a purpose built robotaxi should be a completely round blob without any exterior protrusions, to maximally avoid coning and subway-surfing-style incidents (waymosurfing?)
Comments are really underestimating Tesla FSD in here. As an owner and user of FSD, it's scary how quickly their new Neural Net version is improving. There's a reason Elon and other internal Tesla engineers are so bullish on it, they've seen the data on 12.5 and beyond. Once their 100,000 GPU cluster is complete this year, how many iterations until FSD is comparable to Waymo? One year? Maybe 3 years?
The reason Waymo should be worried is because once FSD hits that "crossing", a flick of a switch at Tesla and 3-8 million robo taxis could hit the market all at once. For comparison, Waymo has around 1,000 cars.
It's cool what Waymo's doing, but as an actual FSD user it's plainly obvious "vision only" will work, which does not bode well for all the other car companies + waymo.
With Elon, these types of massive changes are always just a year or two away and somehow never come or are not delivered as promised. They're bullish because it makes the stock price go up.
It could be better than Waymo today and it wouldn't matter. As long as Elon runs that compant with his middle finger extended to regulators, they'll never be allowed to do what Waymo has done by being so careful.
Look at Cruise. They had actual robotaxis on the streets of SF and were firmly in second place. Then they had a bad accident and covered it up. Now they're set back by years and billions of dollars. Now imagine Tesla's behavior ...
There are cheating devices that simulate hands on the wheel. If FSD becomes good enough they could go from suicidal to popular. Regulators wouldn't be able to stop illegal aftermarket modifications for obvious reasons. I still think Waymo will win, but this is how I think a Tesla victory would look.
> a flick of a switch at Tesla and 3-8 million robo taxis could hit the market all at once
Assuming this is a feature that car owners (and presumably, states/provinces/countries) have to opt into, what does the insurance and liabilities look like for this?
I can't believe that Tesla wouldn't shove all of it onto the end car owner, even if (or rather, when) they take a sizeable cut.
I think there's still significant regulatory hurdles to overcome before it's just a flick of the switch.
I haven't ridden in a Tesla FSD. Have you ridden in a Waymo? It seems tough to compare where they're "at" without having experienced both.
I also wonder how much Waymo will have improved in the next 1-3 years. This whole field seems like the sort of thing where improvements come more rapidly as more data is acquired and data is acquired more rapidly the more vehicles come online. I could easily see a situation where Waymo masters their process and expands to hundreds of thousands of vehicles in a short period of time.
Also, while I think I do believe in the ability to do FSD well with just cameras, I wonder if there will be fringe safety benefits to using accompanying lidar that will end up winning over public perception. People seem very interested in the ability for cars to do driving more safely than humans. I feel like this conflicts, mildly, with the desire to do FSD with pure vision on the premise that humans do it with pure vision. It's a true premise, but I'm not sure it's especially marketable.
Except they are not "doing it" as long as it requires driver supervision. They are doing something that's few levels below Waymo. Removing the driver is the whole problem domain, which Waymo has done years ago.
There's a reason Tesla hasn't been able to remove the driver (or even let them take their eyes off). It's because the car has "just" a few cameras.
Is much known about the science behind Waymo? I am impressed that the cars can operate autonomously in real life when I see them around me. At the same time, they have an enormous number of sensors like multiple spinning LIDARs (?). I also read that they have to map everything ahead of time to be able to operate in an area. That seems a bit like cheating to me. It may work and may even be valuable to customers, but it doesn’t seem like as big a breakthrough as achieving autonomy with the same sensors as humans.
> it doesn’t seem like as big a breakthrough as achieving autonomy with the same sensors as humans.
As far as I know, nobody is currently attempting to do that. Even Tesla has way more cameras on their cars than you have eyeballs, covering far more angles than your head can. There is that one org that does L2 automation with just a phone IIRC, but I don't think they're trying for full autonomy.
Realistically, it's hard for the car to be as smart as the human mental model while driving, and so having vastly superior sensors helps compensate for that.
But...they are a company, their goal isn't to make breakthroughs. Their goal is (eventually?) to make a profit. They'll simply use the technology and breakthroughs that allow them to get there faster. That's not cheating.. that's business
Nobody knows the details, e.g. how many humans are monitoring per car. It came out at one point that Cruise was using >1 human per car. What is Waymo doing? We dunno, we just keep seeing these absurd puff pieces.
The first time you ride in one, it feels truly sci-fi. But within 5 minutes, you're almost bored of it - that's how good it is. If I had to choose between an Uber of questionable cleanliness and driver temperament and a Waymo with a slightly longer wait and slightly more fare, I'd choose the Waymo every time.
(I have no affiliation with Waymo, Google or any related industry - it's just an amazing service!)