This seems like a good time to point out that the argument makes too many assumptions to be useful, like that moving fast and breaking things will in fact lead to faster progress overall. In the case of robotaxis, the group moving carefully and deliberately is the clear leader, and many competitors who took the faster/less careful approaches have shuttered along the way. When uber's self-driving division killed someone, for example, it didn't lead to an earlier arrival of self-driving.
This is relevant to all sorts of business stuff where we're always asked to move faster than we can reliably move. It's astoundingly easy to forget that sometimes bad rollouts can shutter a project even worse than slow rollouts.
Compare to a rideshare driver that will often drop you off right in front of your destination, even if that is an illegal maneuver.
Don’t even get me started on the “didn’t take psych 102: Attention and Memory”-level cluelessness required to believe a human can safely pay attention well enough in a vehicle that reliably tricks you into believing it’s autonomous to take over in the split seconds before a disaster…
I find it hard to believe that the Tesla and Auto Manufacturer positions aren’t knowingly deceptive. I mean, what are they going to say? “It’s too hard so we’re just waiting for Waymo or Cruise to license their tech once it works”?
I’m gonna stop here before I start mocking geohot… I seriously can’t believe the journalists who wrote those early stories were willing to risk their lives like that…
The auto manufacturer approach is also showing progress. In CA and NV you can buy and operate a Mercedes with Drive Pilot, which is Level 3 certified. In the right (very restrictive conditions which essentially come down to "sitting in highway traffic on your commute") you legally do not have to pay attention to the road and can read/watch/work/etc.
I'd personally never trust an autopilot unless it's either backed by human-level AI which has also had years of driving experience, or it's in some very highly constrained environment (maybe airport bus going from gate to plane). Out on a highway or public road system is the most unpredictable environment possible.
This is about the peak of what you can get with automated lane keeping and braking. I don't see any route from this point to anything like level 4.
How do you plan to do that? Will you wrestle the code away from Waymo? Or do you plan to put in the long years of thousands of man hours to develop it and all the costs of the hardware while you do it?
Sadly its self-ownership is only "according to legend" rather than anything battle-tested.
Won't it be great once we have fully self-driving cars? Heck, I could buy a car and then rent it out to other people like a taxi when I'm not using it, and it would pay for itself. Maybe I could even make a profit!
...
If I could make more money than the car costs to purchase and maintain, without any additional work on my part, why would the company sell me the car at that price in the first place rather than just running the taxi service themselves and keeping all of that extra profit?
https://www.forbes.com/sites/cyrusfarivar/2023/12/04/judge-a...
Not sure if you count this as "legit" or not, but I haven't seen similar incidents from Waymo. (Perhaps I've just missed them - if so, links welcome!)
The question of Camera vs LIDAR+Camera is a narrow technical question about how to construct a 3D scene. That's it. It says nothing about making sense of this 3D world for which you you have a 3D point cloud and it says nothing about how to actually navigate that world. Say you're driving down the road and there's a bit of construction, there's a guy holding SLOW/STOP sign directing traffic. LIDAR will tell you it's a hexagonal sign, but it can't tell you what it says, you need a camera to read the sign and tell you what it says. It doesn't tell you how to drive, how fast you should go, how much space to give the guy with the sign etc. Everything AV-related which is not constructing a 3D scene is actually the same across all AV stacks, which includes the hardest part - the actual driving itself.
Your example of needing to read a stop sign isn't a great example. At least in North America, a hexagonal sign is always a stop sign. A better example of your point would be a speed limit sign.
But Waymo never said you don't need cameras. Hell, they have 29 cameras in each vehicle compared to Tesla's 8.
Your point about their approaches being more alike than different is somewhat true, but you wrongly attribute the LiDAR vs camera debate to Waymo marketing. It's Elon and Tesla fans who started it and incessantly repeat it even to this day. Most rational folks say use whatever you can to get it working (which Waymo did) and optimize later.
There are atmospheric conditions and obstructions that lidar can see through that cameras can't.
Cameras also seem prone to being blocked by a small splash of mud/dirt. Is anyone on this thread knowledgeable enough in the domain to know if that's an issue? I thought of it while moving my head sideways to see around a temporary sight obstruction on my windshield. Luckily the windshield is big, and I can move my head. Cameras are small. I guess you just put several so you have an effectively large camera array? It does mean more redundancy is necessary than I would have initially thought.
if you want to drive across the ultra straight highway flyover states it's game changing. if you don't do that, it's not that useful.
I have watched enough recent Tesla self-driving ride along videos on YouTube to suspect you might be mistaken on this point. Tesla intends to launch a cybertaxi fleet and their software looks like it will be good enough to get them there without lidar or additional sensors.
I just watched the latest video from AIDRIVR on YouTube. AIDRIVR is a TSLA pumper-and-dumper who has dedicated their channel to uncritical praise of FSD. In the first third of the video FSD v12 runs two stop signs, once directly into oncoming traffic in a 1-way traffic control and once at a stop where the cross traffic does not stop. This stuff is not even a little bit ready for fully supervised operation. https://youtu.be/fpoXr_z_6a4?t=565
It is one thing to cherry-pick flawless drives on a sunny day and upload it to YouTube while having someone behind the wheel ready to take over the glorified driving assistant system. It is another to run a commercial driverless service open to the public 24/7 in one of the biggest urban areas, knowing that riders will record everything, assuming accident liability, and keeping a nice safety record without someone behind the wheel.
I have a comma.ai in our minivan and it works great. Much better than Honda's built in lane following tech
https://www.teslafsdtracker.com puts miles to disengagement at 30 and miles to critical disengagement at 300 for all v12.x.y versions. Note: this is crowdsourced data and the users themselves get to decide what's critical and what's not.
As far as numbers required to make it fully self driving, it's at least 3 orders of magnitude worse than the big players. Waymo and Cruise routinely had 30,000+ miles per disengagement during their California testing. That's one disengagement for roughly 3 years of driving.
Well, being as how people get to pick and choose when they use it, and that the driver has to remain vigilant at all times, I'm not surprised.
But this is easy to test: stick random people in the car and go to random locations with FSD, see how it works. Why haven't they demonstrated this yet?
I am very skeptical of the "weeks without intervention". It's cool technology, but I never had a single trip where I didn't need to intervene at least once.
It would regularly blow through school zones, failing to read the posted sign.
On a couple of occasions it veered off the road on to the shoulder.
My thinking is the car will never be level 4. It doesn't have sufficient sensors or NN compute power.
Tesla already silently abandoned the "just over the AIR one day" approach with a dedicated car announcement.
However the camera+ultra-sonic radars but no lidar is not only Tesla vision, but other companies too.
We don't know what it costs Waymo to operate their car. The fact that they charge money doesn't make them a real business, just as people paying for FSD doesn't make it a real business.
Both are promises until a breakthrough occurs. Waymo is starting small-scale but for a full setup, even if guided by humans here and there. Tesla starts with millions of cars and multiple countries but with far modest functionality.
Waymo is scaling up; Tesla FSD finally starts to look like the promise, with a high chance of a ride with 0 disengagements still on the scale of many countries and launching it also on a different continent right now.
It's interesting to observe how companies with radically different approaches are about to arrive at the same goal almost simultaneously.
https://arstechnica.com/cars/2024/05/on-self-driving-waymo-i...
Humans suck at driving: https://jakeseliger.com/2019/12/16/maybe-cars-are-just-reall...
Waymos avoid many of the Uber challenges: foul-smelling "air fresheners," dubious music / talk radio choices, etc.
If they are doing 50k rides a day, then they would appear to have a remarkable safety record.
It will be interesting to see if these investigations lead to a repeat of the Cruise debacle or if this will become the price of doing business.
[0] https://www.reuters.com/business/autos-transportation/us-saf...
> But even the most bullish believers in autonomous transportation acknowledge the tech still has a ways to go before it’s reliable enough for widespread deployment on U.S. roads.
Has it though? They've come an impressively long way to have 50,000 rides a week, but that needs to increase a thousand fold to justify the $6B of venture capital and $30B valuation. That's a lot of cars and a lot more work than it takes Uber to bring on another underpaid owner driver (Uber has 23 million rides per day)
The other common mode is secondary to an original crash. Vehicles either are pushed into different roadways, over abutments, or down hills, which causes the vehicle to roll or otherwise crash into pedestrian areas without warning. This is most common in winter conditions.
I've heard they do 50,000 rides per week in SF, LA, Phoenix combined.
Assuming they make $20/ride, that's still $1M/week, or $52M/year. I'm sure they spend in Billions/year.
They would have to scale out to every major city in America and add another 10000 cars before they can turn a profit.
> Please note also that using a self-driving Tesla for car sharing and ride hailing for friends and family is fine, but doing so for revenue purposes will only be permissible on the Tesla Network, details of which will be released next year.
Of course that was in 2016 and as far as I’m aware we are still awaiting those details.
Also, I doubt there is a point in you "renting" your Tesla. Tesla the company has enough money to flood the road with their vehicles, and your vehicle is irrelevant. Have you heard any individual renting their personal Camry to a taxi company or a Uber driver?
If yes, perhaps cities with fewer cars can skip the taxi step and go straight to smart buses.
No need to wait for autonomy. They have (had?) such a service in Cairo, but unlicensed jitney vans were already common there. They never launched it elsewhere.
BYD recently announced that they will not be using BYD's technology. Not good enough for production cars.
Waymo still has rather bulky rotating LIDAR scanners. That technology needs to shrink more before wide deployment. A few years ago, there were lots of LIDAR startups, but few LIDAR buyers, so that industry collapsed.
No, there isn't. We do have a team of folks to support situations where we aren't confident and choose to "phone a friend". This recent blog post covers some of it in more detail:
https://waymo.com/blog/2024/05/fleet-response/
Most importantly: at no point does someone remotely "drive" the vehicle. They can direct it to say "hey make a u-turn and go to this new point", but they aren't remotely driving.
This doesn't line up with other statements made by Waymo though.
That blog post as an example:
> 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.
Yet in an incident in January when a Waymo ran a red light and caused a moped to crash the narrative is
> In January, an incident took place where a Waymo robotaxi incorrectly went through a red light due to an incorrect command from a remote operator, as reported by Waymo.
So I'm curious, is it a case of "The Waymo Driver doesn't always follow road rules itself." or "Remote Ops can make a car run a red light against the Waymo Driver's programming."
[0]: https://www.forbes.com/sites/bradtempleton/2024/03/26/waymo-...
I can't speak to their RVA operations.
Most people in society don't really have a desire to run amok.
1. They're mostly even more local. The Waymo Driver is in your car, the "driver" for say a DLR train is inside the train too. Unlike Waymo they aren't running multiple live feeds to remote oversight, even the emergency human intervention is literally on board with you. There's somebody wearing a uniform telling those tourists that no, Abbey Road is an outer suburb with a sewage pumping station, they're on the wrong train for the famous Beatles photograph. The person in the uniform is trained to drive the train if there's some reason the automation can't do it, nobody can do that from a control room miles away. In the even higher (and rare) GoA systems where nobody aboard can drive the train even if they need to, remote oversight still may need to dispatch a specialist to rescue a failed train.
2. They're mostly "grade separated" that is, they're either underground or suspended in the air, or maybe in fenced off ground-level areas, so you can't use a "hacked" train to hurt anybody except its passengers or maybe, in some cases, passengers on a nearby train.
Turns out cars are a bit too bulky and pricey to repurpose as shanks.
Waymo sometimes does weird, unexpected things - but safely. Once it seemed to change its mind about the optimal route a few times over the course of 10 seconds, switching safely between two lanes back and forth a few times before committing. It used its turn signal fine, and the lanes were clear, so it wasn't a problem, but this isn't something humans do.
Sometimes it behaves oddly, but I have developed confidence that it will do those odd things safely.
Oh, I disagree, this is something I observe and in fact do myself quite a lot. We all run it through our minds which route might be the quickest spending on certain factors. The difference is Waymo (or any tech) will base this on actual data (i.e., getting there quicker) vs humans who will be more emotionally driven (i.e., frustration at the driver in front, wanting to take the more scenic route, being undecided about stopping at that cafe halfway).
I'm all for self driving in highly populated areas. In a perfect world I'd like to see it integrated into all vehicles, and when entering specific areas you are told your car will enter self-driving mode. Arguably this makes the most business sense for Waymo, licence the underlying tech to manufacturers that already have capacity to produce vehicles vs compete.
As far as humans suck at driving, it's not that they suck on average, but that the ones who do suck at it don't always have a sticker saying that they suck.
No one with kids wants to ride in taxis with kids all the time. Ditto for anyone with hobbies that require transporting large things, like kayaks, bikes, etc. Or people with large pets. Or grocery shopping for more than 1-2 people. Or any of the dozens of other conveniences that Americans have come to expect from owning a car over the past century.
I can take my time to get car seats in and kids buckled, without feeling the pressure to hurry from the human driver.
I don't have to feel like my kids misbehaving are going to annoy a human driver, or get me a bad review in Uber/Lyft.
I don't have to worry about tipping, or the driver taking a longer route to charge me more.
I don't have to worry about small-talk, or awkwardly sitting in silence when I normally would be talking with those I'm driving with.
Obviously this doesn't cover all use cases for a car (pretty sure you can't load a kayak onto a Waymo because you'd block sensors), but it seems WAY better to me as someone who doesn't like to deal with the people aspect of Taxis.
You mention American at the end of your comment, but the rest of the world isn’t the same. Waymo doesn’t really have to limit itself to the states once they get the concept worked out.
Americans have become emotionally attached to cars because of what they enable them to do. That might take a while to die. But in Europe cars are more of a pita to own and run because we have less space. I don’t have any great love for mine. As soon as waymo gets here and is reasonably priced I’ll get rid of my car.
People in the Netherlands get fine without a car: kids just bike to school with their friends instead of sitting in the backseat in traffic for 45m every morning. This is because money and space is not spent exclusively in car infrastructure, but cycling and walking and public transport.
> Just because they're driven by computers now isn't going to magically change all the reasons that people didn't use them before (hint: it wasn't because they were driven by humans).
Sort of. The primary reason I don't hire vehicles more often is cost, which is related to the human driver. The wealthiest families I know are much more likely to use a car service to ferry family members around.
If there was a car service that could whisk us to school, work, grocery shopping, etc with no more than 15 minutes advanced notice for less than the cumulative cost of a similarly-sized private vehicle I'd sell one of our cars in a heartbeat. I have no idea whether that future is years or decades away, but when it occurs many families I know would go from 2 or 3 cars down to 1.
I'll admit that going from 1 car to 0 cars would be a tougher sell. For that I'd have to be confident in five nines of availability and vehicles that can haul equipment like bikes and kayaks. But that doesn't seem like an insurmountable problem, just a logistical one that'll take a bit longer.
The driving experience itself is on par with the "best" drivers I've ever ridden with (things like stopping at actual stop signs, for instance, and not racing from one traffic light to the next, and being courteous to bikes and pedestrians), not to mention just the peace and tranquility of being in a car solo when you're not having to drive (I know, I know, mass transit is better for countless reasons and this is actually doubling down on human isolation which is probably not great long term). Anyway, I have zero interest in getting into an Uber at this point. I'd wait longer and pay more for a Waymo if given the choice. And I'm fully aware people will, if this works more broadly, lose jobs bc of it. I'm not insensitive to that, but I don't think the genie is going back in the bottle barring catastrophic incidents by Waymo et al that cause regulators to kill self-driving cars altogether. Note that I did witness an incident where on a road with no lane markings the Waymo straddled a left turn "lane" and a straight-travel lane. It's an intersection I transit often and normal drivers have great trouble with and frankly makes me uneasy every time I turn left there as well. The Waymo was definitely perplexed by it.
For those who talk about how Phoenix's roads are straight and wide... This is not true in Los Angeles (nor in SF though SF is more of a compact grid than LA). For those of you unfamiliar, a lot of the streets in LA where Waymo operates today are very narrow, with cars parked on both sides and so there's inadequate room for two cars to go down them without waiting for another car to pass. These same streets have zero lane markings on them. I've experienced this several times in Waymo to date where the car just "gets it," though it's almost too cautious when it needs to get over to let another car pass when there's not enough space for both. And if you read all of that and say "what about the weather?" It's obviously an issue and I fully agree it will delay the rollout "everywhere."
All that said, I cannot wait until I can jump in one of these things, from Waymo or any other company, and safely go up to the mountains or some other road-trip destination. The economics of longer trips, particularly to rural areas, are likely tricky bc of the inability to count on a return fare, but, man, I do think self-driving cars are a radically important technology that will vastly change how we transit and, really, how we live. That is, if they don't fuck up too much en route to getting there.
I agree with all of your points about Waymo vs. Uber-like ridesharing—the average Uber ride is so much less safe that it’s hard to argue for.
But I also agree with your aside about the growing isolation of society—the longer term implications of every event, meal, and errand being separated by autonomous journeys are staggering.
So the question is, how do the societal isolation factors play into your decision making? (Honest question, not a gotcha, I’m curious how others think about these tradeoffs.)
I live in a big city (larger population than Phoenix) in the Uk and I've never even seen a self-driving car. Anywhere. I don't even think such a thing exists on public roads in my country. That Gibson quote about the future not being evenly distributed, etc.
Just a data-point.
You won't know if people have Level 3 "Self Driving" cars because unlike Level 4, the Level 3 cars always have a human sat in the driving seat, it's just that maybe the human isn't paying attention and maybe the car is driving anyway. It may be difficult to gauge (beyond guessing) how many people you see are bad drivers and how many aren't actually driving at all under L3...
L1 (the machine does some of the work but a human driver is always doing much of the driving) is certainly something you see and don't even think about. Intelligent Cruise control (ie it won't smack into the car ahead but instead slow down) on a motorway, maybe automatic lane keeping on somebody's fancier or newer car, it's not "Self driving" as you'd understand it, but it's something.
The way these "Levels" work is L3 to L4 is the point where we transition from "The human is legally driving but the machine is offering more and more assistance" to "The machine is legally driving and the human is asked less and less often to do anything at all". As a result a person who is literally blind and thus couldn't possible drive the car or obtain a license to do so - can (and they do) use a Waymo, just like they'd use an Uber, but they cannot do the same with Tesla "Full Self Driving".
That’s an interesting way of saying you live in London ;)
(Phoenix urban area is more populous than every urban area in the UK except for London)
Countries develop at different rates on different things.
A significant portion of traffic deaths also occur in special conditions-- at night, with intoxicated persons, in bad weather.
Existing self driving cars won't even drive in those more difficult conditions.
In terms of the passenger miles driven if you compare to non-intoxicate humans the expected number of deaths for self driving cars is still below 1 if they were as safe as non-intoxicated human drivers.
Safer cars are an excellent goal but they're not automatically a given result for self driving.
> Waymos avoid many of the Uber challenges: foul-smelling "air fresheners," dubious music / talk radio choices, etc.
And introduces new ones like being dropped off blocks from your destination because the car refuses to drive on perfectly fine roads, service being unavailable in poor weather, and extending Google's tracking of everything you do online to offline.
:D
Aside, you can just ask uber drivers to turn off the radio.
To put some numbers on it in the US cars are driven about 3.2 x 10^12 miles per year, and around 4 x 10^4 people are killed in car accidents (drivers, passengers, pedestrians, and cyclists).
That's one death per 8 x 10^7 miles.
There are around 2 x 10^6 people non-fatally injured in car accidents per year in the US. That's an injury every 1.6 x 10^6 miles.
There are around 4 x 10^6 non-injury car accidents per year in the US, which is one every 8 x 10^5 miles.
If we assume all miles driving are equally risky and that we drive 40 miles per day 365 days a year, then we would expect to be in a non-injury car accident around once every 55 years, be injured in a car accident around once every 110 years, and be killed in a car accident around once every 5500 years.
Of course almost no one drives all their miles at times and in conditions when the risk per mile is average so when estimating your personal risk you need to take that into account.
Phoenix has the perfect climate for self-driving cars.
It will require a major technological leap in order for them to succeed in the "real world" (fog, rain, snow, etc).
We handle both dense fog and heavy rain on the latest vehicles. The best blog post is probably https://waymo.com/blog/2021/11/a-fog-blog/ but you can find a lot of videos in the rain.
Snow and very cold weather is a challenge for sensor cleaning. We've done some testing in both NYC and Buffalo (https://waymo.com/blog/2023/11/road-trip-how-our-cross-count...) to collect data.
I'm not sure whether this reflects their own preferences, what they think customers want, or if they are just completely oblivious.
Knowing BigCo reputation, I think it’s equally possible that Waymo and/or BigCo accounts will be banned for actual perp, complainant or random rider in-between… what a world…
> The car went onto a freeway, where it travelled past an on-ramp. According to people with knowledge of events that day, the Prius accidentally boxed in another vehicle, a Camry. A human driver could easily have handled the situation by slowing down and letting the Camry merge into traffic, but Google’s software wasn’t prepared for this scenario. The cars continued speeding down the freeway side by side. The Camry’s driver jerked his car onto the right shoulder. Then, apparently trying to avoid a guardrail, he veered to the left; the Camry pinwheeled across the freeway and into the median. Levandowski, who was acting as the safety driver, swerved hard to avoid colliding with the Camry, causing Taylor to injure his spine so severely that he eventually required multiple surgeries.
> Levandowski and Taylor didn’t know how badly damaged the Camry was. They didn’t go back to check on the other driver or to see if anyone else had been hurt. Neither they nor other Google executives made inquiries with the authorities. The police were not informed that a self-driving algorithm had contributed to the accident.
> According to former Google executives, in Project Chauffeur’s early years there were more than a dozen accidents, at least three of which were serious. One of Google’s first test cars, nicknamed kitt, was rear-ended by a pickup truck after it braked suddenly, because it couldn’t distinguish between a yellow and a red traffic light. Two of the Google employees who were in the car later sought medical treatment.
It was a long time ago, but Larry Page was well aware of it, and imagine if that incident received fair coverage and investigation.
I recognize accident lawyer work when I see one :) They charged Waymo’s insurance to the max.
Self driving buses will be such a boon for public transportation. Now you can have 24 hour buses, that operate on holidays as well, or even dynamic, short term routes based on demand (eg: after a concert or sports event), without being dependent on the availability of pre-allocated human drivers.
Electrical autonomous vehicles don't have a need for a driver and electricity is relatively cheap. So you don't get much economies of scale by making them bigger. Most city journeys would be under a kwh. Even at current grid pricing that's cheap.
Eventually, cheap autonomous vehicles could be mass produced at low cost and would have very low operational cost. So the ride cost would be comparable to, or lower than, current public transport options.
Trains (and train-like options such as metros) are vastly more efficient than cars in number of people moved per unit of time per area used. That might not be a big deal in suburbia, but in dense inner cities it's one of the most important drivers of public transport.
If you are not paying for the conductor, can’t you make trains much more appealing? They could run every five minutes, and last mile can be solved with autonomous car that is waiting for you when you arrive.
AVs give us a path toward a world where very few people need to own their own car. We can put all those parking spaces to better use. We can improve equity by giving more people access to safe, reliable, affordable, and convenient point-to-point transportation. Being able to consistently get a ride to where you need to go is something we consistently under-appreciate. It means being able to get a better paying job on the other side of town. Or not having to worry about missing a dialysis appointment, or a meeting with your parole officer or therapist. When the marginal cost of a robotaxi/robobus ride is close to zero is when the AI economic boom will really begin.
Interestingly, no one ever argued for the profitability of cars, so all we can do now is to calculate the overall economic costs and societal benefits and that's where public transport clearly and easily wins.
But the impact of taxis on road traffic in a dense city is comparable to the impact of private cars - perhaps even more so as they're often travelling empty between rides. If every journey which was previously done with a car is done with a taxi, there's no reduction in vehicle traffic - meaning the same problems of congestion and pedestrian safety.
Driverless cars can probably drive closer on highways to increase throughput, but that doesn't really help in cities or residential areas. Ultimately if lots of people shift to driverless taxis to get around, there will be far more vehicles on our streets.
I will give tempe/scottsdale credit though - they have their roads around the major tourist hubs in GREAT shape - the lines crisp and the lights bright and new - I think it makes it much easier for a waymo to get around.
Waymo is a real business serving 50,000 rides each week delivering paying customers to their destination. If you haven't tried it yet, the product is amazing. Private, doesn't cancel, safe, and smooth. I will never take Uber again if I have the choice.
How much money is Waymo bleeding every quarter? Maybe the investors don’t care, but it’s relevant if you want to call it a real business.
The biggest (only?) complaint I had is that it would not pickup/dropoff at the curb at our hotel. So if it was raining, we'd have had to walk out in the rain to meet the car in a parking spot.
The real business is an entire transit system, with purpose-built vehicles of various sizes, centralized routing, etc.
"Lying to judges" (do you mean withholding material information from regulators?) is not something I'm aware of Waymo doing. (Again, links welcome -- and remember Waymo is not Google.) Nor is it a binary thing. It's one thing to cover up e.g. anti-competitive behavior in the free market, but quite another thing to cover up how you might've actually killed a person on the street.
IBM had an AI (Ross) pass the bar years ago...I believe it was actually 'hired' as an attorney at a firm in London for routine paperwork..
0) https://www.linkedin.com/pulse/ibms-artificially-intelligent...
1) https://www.dailymail.co.uk/sciencetech/article-3589795/Your...
I can easily imagine the same (unprincipled) dynamic applying to an AI lawyer.
On top of all that, Waymo does intensive 3d mapping for their service locations. These maps have to be maintained. Then the cars need sensors that take advantage of those 3D maps. If that combination of intensive mapping and LIDAR sensing turns out to be necessary to get beyond FSD's current and near future performance, then Tesla isn't even at the starting line.
> Say you're driving down the road and there's a bit of construction, there's a guy holding SLOW/STOP sign directing traffic.
Here's a picture of a guy holding a hexagonal SLOW sign. They are very common. https://nj1015.com/how-slow-should-you-go-through-constructi...
LIDAR cannot tell you what's on that sign. It cannot read any sign, nor any road markings, nor streetlights, blinkers, stop lights, etc. If you believe that cameras are not capable of being used as input due to dynamic range or any other reason, then that's fine, you just believe that self-driving is impossible (lidar or not). But to believe that one can safely drive using nothing but a completely blank and unlabeled 3D scene (ie, LIDAR-only)? That's pretty crazy.
the back of a stop sign is often a slow sign, or a do not enter sign. The same shape of sign, but different meaning from different directions. The slow variant is often held by construction workers, hence the GPs example.
Tesla FSD sucks extra bad on sunny days in fact, due to its basic optical systems.
Yes, is that not enough?
It's called Level 3 because it is level 3. Mercedes went through an approval process and carries insurance (or a bond iirc, there's a couple options) to comply with California law dictating the use of L3 features. You are legally allowed to stop paying attention under certain conditions and the restrictions to roads or situations is in no way disqualifying, nor is geobounding to only states you are legally allowed to operate in. Also, its available across all of Germany.
The way its actually a marketing gimmick is how few Mercedes has actually made available and the exorbitant cost. They've been allowed to sell in California since June of last year and only have 65 available and 1 sold as of April:
https://fortune.com/2024/04/18/mercedes-self-driving-autonom...
Your mind is going to be blown when you hear about Tesla and the name they give their assisted cruise control.
So the question is: is LIDAR also necessary or are cameras sufficient? IE, can cameras+motion give you an accurate-enough 3D scene the way LIDAR can. And that's a narrow technical question, and it isn't even the most question when you consider self-driving as a whole.
"LIDAR is necessary" is not exclusively a Waymo talking point - it is shared by all companies using LIDAR, suppliers of LIDAR etc. But it is just a talking point, there's no reason to think it's actually true.
1. Only companies who have LiDAR in their stack have shown fully autonomous driving i.e. driverless.
2. Better 3D scene construction, which LiDAR, unquestionably provides permeates as an advantage throughout the stack.
I do think “LiDAR or not” is a narrow argument. But the advantages are massive and undeniable, so it becomes necessary especially in light of rapidly falling costs.
The most crucial insight for Self Driving Cars is that this is not the trolley problem. "I give up, stop where I am" is a valid answer.
We actually have built automation where "just give up" isn't a valid answer. CAT IIIc autoland (on a jet liner) has "Fail Active" scenarios where the machine concludes just before touch down that it no longer has confidence in its position due to one or more sudden sensor failures but the human pilots can't possibly intercede quickly enough to be safe and under IIIc conditiosn they can't see anything anyway, so, although the aeroplane will tell the human pilots that a failure occurred, it will nevertheless attempt to continue the now unsafe landing in this edge case. Most likely despite the reduced sensor validity, this is successful and everybody lives, and if not it's not as though the humans could have reacted in time anyway. But self driving isn't like that, the plane is flying, if it were to just stop flying everybody dies. In contrast a car can just stop and it's merely annoying.
The assumption is that the highly networked car has some kind of security vulnerability that allows malicious users to take control and perform acts of terror.
Since this is in software, the attacker can theoretically scale the attack to involve all cars on the road with the same vulnerabilities, which could be millions of vehicles.
This problem is not unique to self driving cars - just any highly networked car with software control of key systems, like a modern Tesla. However, a full self driving car may not manual overrides that allow the passenger to stop the vehicle.
Working remotely helps a lot since there's no need to drive every day, but it's only a fraction of overall people who have this privilege.
Still, I'm looking forward to seeing Waymo at my town. It would make a good DD and backup in case my car needs repairs, etc.
That makes a kind of sense, a human shouldn't need to run red lights but they do it more often than you'd like, whereas they mustn't hit pedestrians (although sadly they sometimes do). Just the other day I was watching video of a failed London Underground signal which is stuck at red, the driver knows this, and the signaller knows this, and nevertheless the signaller (who is in a position to know as they've got a board full of position data for trains in their sector) has verified that crossing this signal despite the danger aspect is safe.
This happens so often (ie sometimes) that TFL has a recorded announcement to play to passengers when, as the driver, you're about to do this. The train, you see, doesn't know that what you're about to do is fine, so it's going to stop you. So as the announcement explains, the train will move forwards slowly, brake suddenly to a halt, and then after a moment proceed slowly again. The announcement suggests that passengers should sit down if able to do so. The driver having secured permissions from the signaller will drive forward ("at caution" ie slowly enough to stop short of any obstacle discovered), then the safety systems will detect the danger signal, braking the train to a halt, then the driver proceeds to drive slowly again because they already know why it stopped.
The reason is that you may make things worse by running a light, the people driving an emergency vehicle have in some cases trained specifically to run lights (although not all reasons they're driving under lights and sirens would justify doing so) but they aren't trained to make it safe for you to do the same so you should not.
A comment I wrote 3 years ago has more: https://news.ycombinator.com/item?id=26950254
The old "look at the person to your left, now look to the person on your right" meme comes to mind. One of you will probably have an accident with an injury in your lifetime.
I ran the same calculation for dying in a car accident and got a lifetime probability of 0.7%, but I'm not sure I did it right.
https://www.autocar.co.uk/car-news/technology/mercedes-use-t...
maybe it's a cultural difference
Waymos drive fairly fast and aggressively in rush hour traffic too, which is why I enjoy sharing the road with them. I was initially worried they'd drive like a grandma but that hasn't been the case. Also as a cyclist I enjoy riding near them because they know I'm there and they give you enough space in the bike lane.
The hard thing is that every other human car acts randomly because they don't say, have winter tires, and unlike waymo, don't have the very quick control loop.
It is also funny to watch them get stuck behind busses, having followed too close to safely go around them when the bus stops to pickup/dropoff.
Also Ive seen multiple instances of them trying to turn left on red and pull far enough into an intersection to cause issues.
Finally when I am on my skateboard they dont seem to recognize me as they drive very very close and fast, though I havent felt risky enough to really test this.
That's why they've done winter testing in Tahoe (since 2017) and Buffalo NY (last winter).
When I'm out doing something, the car also serves as a reasonably secure private locker where I can store things without carrying them around.
Hell there's people who just use uber exclusively now.
Not every young person spends all day hiding in their room, doom scrolling on social media. Some of them have to get to sports practice with bulky equipment in places that public transit and robo-taxis don't go.
What’s more I don’t think we will have that privilege for long, personal transportation is a luxury in most of the world, it is becoming a luxury here also. We will adapt though like everyone else has.
Renting gear is fine for casual users, but serious practioners in almost every sport are very particular about their gear.
Thanks for sharing. I skimmed through the video and watched a fair amount of it. I got a different impression.
I thought it was impressive how FSD 12 navigated narrow winding roads with parked cars and oncoming traffic and flaggers holding signs that alternate between stop and slow. My impression was while it's not perfect, it's a few iterations away from having very few situations that require disengagement. And keeping in mind that every incident of disengagement is a learning and improving moment for FSD, the following iterations of FSD will continue to get more impressive.
This isn’t just “a few iterations away”.
Most of what you said has been repeated for years from people who just watch curated YouTube videos. I’ve driven on FSD v12 and I’ve intervened multiple times almost every single drive. It’s nowhere near ready and will likely never be with that sensor suite.
I suspect that Tesla is paying attention to the disengagement events and working hard to minimize them in the future, but I truly have no idea.
Also I am sure this question has been asked before but what is good enough for FSD? Perfect in every situation? Better than the average human driver? At par or better than an expert professional driver? I don't have an answer personally but I am curious what others think.
A liability-viable self-driving car needs to be reliable enough that you would expect to see zero significant errors in a typical journey. That's around the point where you will only have a few articles about one of your thousands of cars going wrong each month. Commercially viable needs better than that.
Averages don't work for risk evaluation.
Eg, turn off the phone or turn off notifications, even vibrations. Or at least make a strict rule to ignore it while moving and only look while fully stopped, eg at a stoplight although that's just decent driver grade not excellent driver, because of the temptation the notification will present each time.
A great driver also reviews the route for a couple of minutes before leaving in order to reduce reliance on GPS-- you still use GPS but because it's not the first time you've seen the material, already know the shape of the route, and just need reminders to encourage you you're on the right track, the GPS will genuinely steal far less attention. The two minutes will be well spent, and may save lots of time, because it vastly reduces the likelihood of wrong turns.
What’s good enough for FSD is being able to do it without a driver present like Waymo does. Their crowdsourced reliability data in https://www.teslafsdtracker.com/ suggests they need at least 3 orders of magnitude improvement to remove the driver.
That said I admit if these are the kinds of complaints we are discussing, as opposed to the kinds Uber attracted (like running a woman over in Nevada), Waymo must be doing pretty well. These are nitpicks to gradually address, not fundamental issues. Kudos to waymo, it was always obvious they were nearly the only player seriously trying
Early on, they had those concept cars that looked like they belonged at Disneyland or in a Chevron commercial. Then, they started modding off-the-shelf cars at talking up the Waymo Driver. I think at some point they decided their core competence would be self-driving specifically, leaving the "car of the future" bit to traditional car companies.
I expect that robot taxis will be both consumers and producers of that actual data. They will likely report the traffic conditions they experience back to the company that runs the robot taxi service, and that will become input to the rest of the fleet.
If the time it takes for observations from a given robot taxi to be incorporated into the data received by other robot taxis is short enough it might be possible to get interesting feedback loops. It may even be possible to get oscillations.
Still, a person can dream.
> This is about the peak of what you can get with automated lane keeping and braking.
Are you saying that within 5y, say, we won't see a level three system that's able to handle full highway speeds?
well yeah, that is the definition of level 3. That's not going to change.
They are limited by:
> Our technology relies on a digital high-definition map that provides a 3D image of the road and the surroundings with information on road geometry, route characteristics, traffic signs and special traffic events
IMHO Tesla's "F--- it all just use NN" approach does get around the pre-generated 3D map requirement. Even if it is not much more than Level 3.5 at the moment. Pretty funny to see it accelerate to 65mph in a parking lot b/c it thought it was on the nearby highway though.
It's not even level 3, and couldn't be certified as such. It's not reliable enough for the person in the driver's seat to focus their full attention elsewhere.
>> level 3 is still basically lane keeping and auto cruise control
> well yeah, that is the definition of level 3
No, it's the definition of level 2: https://www.sae.org/blog/sae-j3016-update ("example features: lane centering and adaptive cruise control at the same time")
Am I missing something? Lane keeping, cruise control and offset keeping is level 3. That's what this feature offers according to a few posts up.
It makes no difference to the conversation. I'll stop poking the bear here.
According to Kagi the population of Greater Manchester is 2.8 million vs 1.61 for Phoenix.
Phoenix's metro population alone is 4.8m.
I think my point stands, that even large urban areas in the UK have no SDVs.
I can’t see trusting that a robot sees me, otherwise I guess I’m just giving deference to them now?
At least with humans I can see if they are distracted, I can yell at them to get attention, perhaps others. With a robot taxi, I can’t tell anything about it, and I’m a software dev so I absolutely don’t trust anything running software made by a corporation in search of profits, or any other software really, to operate a car in open roads. Not yet at least, maybe in a few decades.
your reading comprehension sucks
> "The version of FSD I tried in March [of 2024] was clearly not ready for driverless operation. For example, I had to intervene to prevent the Model X from running over a plastic lane divider, a mistake Waymo would not have made in 2020. So while FSD 12.3 seems superior to Waymo’s technology circa 2018, it’s not as good as Waymo’s technology at the end of 2020."
> During a 45-minute test drive in a Tesla Model X, I had to intervene twice to correct mistakes by the FSD software. In contrast, I rode in driverless Waymo vehicles for more than two hours and didn’t notice a single mistake.
That seems pretty significant.
Again, it's just different approaches to solve the problem.
99% of interventions I had were from a driver's experience perspective, not the actual safety.
It's still far away from 99.999 that you would expect.
It seems silly to analyze a 4 year old version of something that is changing extremely rapidly.
Both Waymo and FSD have come a very long way since 2020.
FSD 12.3 is the current version of Tesla's self driving software. The article is comparing that current version of FSD to Waymo's 2020 state and saying that Tesla's self driving code today is worse than Waymo's 4 year ago. That is, according to the quote, Tesla is more than 4 years behind.
What makes trains efficient has more to do with the cost of energy and drivers than anything else. Both of those go away if you have autonomous electrical vehicles.
A random crappy light rail line will do the equivalent of 5 lanes of traffic (each direction). A serious subway more like 20.
https://visual.ly/community/Infographics/transportation/solu...
Even if you run cars with no distance between the bumpers you'll still need room for changing lanes, crossing and the line.
We'll see how this plays out.
You also have to take into account all the other factors that make roads preferable, for example, that rail capacity number assumes perfect utilization. In practice railways often have lots of downtime due to overnight shutdowns, broken signals/trains and labor strikes. None of these affect the roads.
Private motor vehicles: 600 - 1600 per hour
Mixed traffic with frequent buses: 1000 - 2800 per hour
Two-way protected bikeway: 7500 per hour
Dedicated transit lane: 4000 - 8000 per hour
Sidewalk: 9000 per hour
On-street transitway, bus or rail: 10000 - 25000 per hour
https://nacto.org/publication/transit-street-design-guide/in...
And energy isn't free. If we had any intention of becoming net zero, electricity prices had better increase. And driving around 2t empty weight isn't the way to get there.
Incidentally, a trip that's less than 1 kWh (so, less than 6 km) is a trip that could easily be made on foot or by bike.
2. They will need stop some time. Where? Will this block the street?
3. They won’t all go to the same place so there will be delays at junction and side streets
4. No margin for error wouldn’t fly in practice, so the cars cannot be that close together
5. How will pedestrians cross a train of cars?
I just don’t see how this all adds up. Automation doesn’t remove the space constraints of cars in cities.
A train line has 10x the capacity of a single lane of road. Even if trains are only coming every few minutes, its impossible to compete with a train carrying 1k people with cars. Perhaps reasonably loaded busses would be comparable or better, but that's not the argument you're making.
Trains are financially efficient because of cost of energy and drivers (and, arguably, roads + cars move much of the expense to the public, where as everything related to the train is on the operator's balance sheet), but they are also very space efficient, compared to roads + cars.
They seem to undercut Uber and Lyft by a hair. Given the longer wait times, and lack of a need to tip, that seems fair. (In San Francisco, they undercut by a wider margin. But human ride shares are more expensive there for a variety of reasons.)
edit: What to do when people get brain chips?
There would be a legal firm to handle direct tasks, and vetting contractors, then an auditing legal firm. A CEO-type, and various checks/fail-safes.
Periodically hire competent groups to come up with new ideas and have a policy for updating the policy.
One of the "ideas" people would eventually realize the need for failsafes for the failsafes: taking control of the entity's assets would become extremely profitable. So you need some kind of secret society to keep an eye on things and execute the assassinate-and-reset plan when necessary.
Naturally its influence would need to turn malign, after it was too powerful for anyone to stop it…
A friend from Türkiye told me in his village they take turns taking everyone's cows to pasture every morning and back home at the end of the day. Some homes are abandoned, people died, the gates are open, the cows still living there just join the heard and go back home at the end of the day.
Horses or dogs should be able to learn to walk from A to B if they get food there. If you don't pay them they should eventually refuse (specially if it is you again) and require some payment in advance. Younglings can be tied to the cart. I think they will know their route eventually.
You could even have a Musk like figure promising it will work next year for the first few decades.
Nor is it possible in the West for two companies to own 100% of each other's shares (i.e., circular ownership). In most of the West, a subsidiary acquiring its parent's stock is treated as a redemption of those shares (i.e., as if those shares were returned to the parent), assuming the transaction is even allowed.
Winter here changes daily between
- no road lines visible
- snow packed into ice randomly making the road a a camouflage pattern
- snow is fresh/deep so no road is visible and you navigate based on the slight hump in the snow where you know there's a curb
- same as above, but instead of a curb a slight indent where there's a ditch
- slush piles outside the tire lanes, which if hit will suck you in or cause you to spin out
- ice/snow on hills, so time your arrival for rolling stops at intersections because stopping is not an option
- active snowfall (limited camera vision, and I'm guessing reduced/useless signal from lidar)
- hail
- sporadic black ice (its easy to slow down when its icy everywhere, but knowing when and where black ice is likely when it's sporadic is a skill)
- the "lanes" formed by peoples tires in the snow often don't align with the official road, and sometimes a lane goes missing in this situation
And all that's after you deal with sensor cleaning.
Snow drifts in the wind, side roads don't get regularly plowed, and conditions change rapidly so unless they can handle the majority of the list above then they won't be able to handle winter period.
> All it takes is for something non-routine to happen such as a car ahead reacting to an animal, or swerving as the driver reaches for something or spills coffee on themself, or a wheel come off a car (I've seen it happen to a car in front of me), or a car crosses the center median in opposite direction (which left my ex-boss hospitalized for 6 months)
I merged hard right before it was clear. I assumed the car on my right would do the same, and the driver further behind would break in time. They either made the same assumptions as me or took my queues and acted accordingly and everything turned out ok.
It is pretty amazing how people can coordinate on-the-fly.
However, the human has the major advantage of having a brain and being able to understand the consequences and potential outcomes of something in real-time as it is unfolding. I doubt most autopilots would understand the situations I mentioned - certainly not unless they were specifically pre-programmed/trained into the system. Would an autopilot even see what is going on inside a car if a driver is bending down below dashboard, or fighting with passenger, for example?
Your average driver is already at the limits of human capabilities with no room for improvement unless they go train to drive like F1 racers. So comparing humans to the current state of tech seems silly given that the tech will assuredly one day be faster than a human.
If we had the same lawyers available that Tesla does, humans too would be not responsible for much of anything.
Driverless cars run by a company protecting itself from reputational and legal risk seems less dystopian than the status quo.
If Waymo hits a cyclist which leads to death, and Waymo is found to be at fault, that's definitely going to make headlines and potentially lead to a pause of the entire operation.
These regulators should be supported and kept clear of regulatory capture. Other countries can do this, so should the US.
And not sure why you think running stop signs or any anti safety measures would increase profits.
Because this big companies like Google are actually evil. As an example the mobile YouTube app does not let you use it if you turn off the screen. So Google decided that wasting energy and killing batteries is an acceptable thing to do, this is pure evil - I would accept they adding more advertising or whatever but killing the life span of a device and wasting energy is truly evil shit.
Usually the comparison comes up in cities where throughput is the limiting factor, and cars end up moving at near-walking pace anyway.
“In 2022, 13,524 people died in alcohol-impaired driving traffic deaths.”
Source nhtsa.gov
You’re gonna briefly glimpse a passed out drunk driver before it plows through you at 40 mph
> They didn’t go back to check on the other driver or to see if anyone else had been hurt
They should have made sure the driver was okay.
The safety driver was Anthony Levandowski, who left Google for Uber, taking with him a bunch of stolen IP, at Uber ran a cowboy self-driving car division that got pedestrians killed, Levandowski got sued by Google, ended up in prison and Uber laid off the entire division. Later he was pardoned by Trump.
So good news - the callous and uncaring safety driver has been fired, sued, and imprisoned.
I am unfamiliar with the details of this incident and my reading based on the facts presented is similar.
Could someone provide more information?
A good piece by an ex-Googler on his 2-year journey toward recognizing the scale of institutional thought at Google: https://mtlynch.io/why-i-quit-google/
Like banning them altogether following a public outcry? That is the mechanism to hold them accountable.
Also in individual cases it will be very easy to sue them for accidents they caused or contributed to. Already is.
Where does this “no mechanism to hold them accountable” comes from?
>> that's cool. until it's not...same with code updates that will make cyclist life worse...you're not really their main concern
I agree and expect that the wide safety tolerances driverless cars currently have will become tighter as they gain more experience, and that this will make them more efficient but potentially less pleasant to be around than they used to be.
But even if pedestrian and cyclists lives are not a main concern for self-driving car companies, some concern is better than none. For some human drivers their concerns seem to be things like not getting arrested, getting to their destination as quickly as possible, checking social media to satisfy their boredom, and not scratching the paint on their vehicle. Some drivers consider vulnerable road users like cyclists to be sub-human [1].
My point is that the bar in the US has been set so incredibly low that even if the code updates make their products worse for cyclists than they used to be or even killing some vulnerable road users that may still be safer and preferable to the incompetence and complete indifference on the part of human drivers.
Having said that, the same calculus may not apply in countries that don't issue drivers a license to kill people, so the bar for driverless cars is likely to be much higher in such places.
[1] https://www.sciencedirect.com/science/article/pii/S136984782...
You cannot train a F1 driver for highly dynamic events at 20k rpm and 4000 NM forces at the axles, you need automated controllers for that. You can train him for simple things, like gear switching (5ms) and hitting the breakpoints right. But an AI will be at least 10.000 better than a human on this.
You need a slow human brain for the stupid mistakes instead.
This is nowhere near true. The skill range from average driver (who knows nothing about car control or reacting correctly) to F1 drivers is vast. There's a huge gray area between these extremes that would be improve safety is people chose (or were forced to) get some training.
Even something as basic as taking a car control clinic once a year would improve the average driver skill and safety by a huge amount.
I'll be happy to take it seriously after that happens. In the meantime, I'm sticking with my belief that unreliable automation is worse than no automation.
You just dreamed up this statement and immediately believed it.
In a world where waymo works as a taxi, it also works to deliver a human-drivable rental car right to your door (and send it on to the next customer when you're done with it).
So now the short term car rental user experience should be dramatically better, even if the robotaxi isn't appropriate for all the tasks.
This or an equivalent will arrive to a robo taxi near you when the service inevitably gets enshittified to hell and back.
Ads, trips shared with other humans, pay extra for heated seats, etc.
Is it Yuck? Yes of course. But it also seems extremely unlikely. And it's a lot less Yuck than thinking about what's in the sand in public playgrounds that all kids visit constantly. And while I have a reasonable chance of preventing them from licking the seats, I have no chance of preventing them from eating some sand.
This is just a bizarre irrational worry...
Do you and your kids take public transport regularly? In a failing city/society, that is.
Because if the answer is "yes, I do take public transport in a city that struggles to pay its bills and I still don't care that the chairs have weird organic substances on them" then fair-play to you, but for me personally at some point I had to purchase a personal car (when I was already approaching my mid-30s) because I just couldn't convince myself anymore that it is ok to not want to sit down inside of a train ("better stand up by the window, that seat is too dirty").
And these robo-taxis will be worse than public transport, for the main reason that there's no-one "standing guard" inside of them (and, no, Big Brother cameras placed inside of them, which should be a dedicated topic all by itself, btw, really won't change a thing in that respect).
That's just one tiny example out of sea of examples.
Before making such bold claims, you've got to know a little tiny bit of the topic or at least look it up to make sure.
I can almost guarantee that if they don’t already, the robo-taxis will eventually start asking for tips.
This is already the case at self-checkout in some stores for example.
As long as the companies can get away with it, they will tack on any number of extra fees and charges even if those fees and charges really don’t make any sense.
Hell, even tipping people does not really make sense the way it works in some places. A person working for a company should receive enough pay from the company itself that they don’t have to actually rely on tips in order to make enough money to survive. Tips should be a nice extra that customers willingly add because of good service. Not a forced extra percentage that they have to pay on every transaction just so that the company can pay less to their employees.
If people are willing to pay base charge + tip for an Uber, then that is what robo-taxies will charge too. Especially if one company is allowed to keep a monopoly on the technology.
The only differentiating factor is that Tesla's approach has theoretical benefits, if they can prove it works. If they could hit the same disengagements/mile rate as Waymo, they would likely be ahead since they don't need pre-mapping.
That being said, theoretical benefits are worth about the same as monopoly money. Until they can demonstrate that they can get the same performance without pre-mapping and LIDAR, it's all just conjecture. There are no points awarded for "yeah, well if mine had worked it would have been better!" They might get there eventually, but the current signs don't seem promising. At least not for getting there before Waymo completely eats the market.
Also, at least Waymo's disengagements are known. They're defined by the DMV, it's basically "any time control needs to be taken away from the autopilot". If a human in the car manually takes over, that's a disengagement. If the car decides it can't drive safely and prompts a human to take over, that's a disengagement. If someone has to log in to the car with a joystick to get it somewhere, that's a disengagement.
The only confusing situation I'm aware of is when remote techs give the car waypoints, like if it gets confused in a parking lot. I don't believe that counts as a disengagement because the car is still driving itself, it's basically just failing at pathfinding. That seems reasonable to me, because it's not a safety risk at all, just an annoyance to the business.
I suspect an underlying issue with socialising is faith in humanity. It's hard to have faith in humanity in the modern world when every front appears to be telling you otherwise. If you don't have faith in humanity, then you're limited to only interacting with those: "you have to" and "are vetted".
At least I personally used it a lot, and knew several people that did.
I'm not sure why you are trying to justify such a large shortcoming - for most people having your car be randomly unusable for 1/3 of the year would be a big deal.
Hmm, I used them this weekend and was comparing pricing. Waymo was cheaper. That said, I wasn’t riding during peak traffic. And in peak traffic, I’d vastly prefer a Waymo. So I get the premium pricing.
Btw, the definition of disengagement is up to the companies, not the DMV https://thelastdriverlicenseholder.com/2024/02/03/2023-disen...
They can't fudge them too hard without the CA DMV yanking their license.
> People are doing 2-3hour drives with FSD 12 with no disengagement. You should check out the recent progress.
2 or 3 hours is impressive in a vacuum, but not all that much compared to the other players. It's an accomplishment for sure, and would have been overall impressive in 2018 or 2020. In 2024, it's only impressive because it can do it without lidar or radar, and even then that may only be because Tesla is the only one doing that so there's really not a comparison point.
It's certainly still an accomplishment, but I'm not convinced that it's economically relevant at this point. Best of luck to them.
To be clear, I dislike Tesla, but I do hope this works out because the alternatives are much too expensive for consumers to actually own. They'll have to use an Uber/Lyft kind of system because the lidar sensors alone cost more than a Lamborghini last I heard.
They do appear to be having basically the exact issues people thought they would have without lidar, though. Namely that different types of sensors are vulnerable to different types of phantom objects, and it's difficult to eliminate those without either having a different sensor that doesn't see those phantoms to sanity check against, or creating issues with failing to detect real objects.
A couple of examples: Video is prone to detecting the stick cyclist on road signs as an actual cyclist, where it's obvious on lidar that it's a road sign. Inversely, lidar can read steam (like that coming out of a sewer manhole) as a physical object where video can tell that it isn't.
It's difficult to use a single signal to fix either of those without having knock on effects that cause failures to detect real objects. It's much easier to use a combination of signals to disintermediate. Cyclists are not flat, so lidar tells you it's not a cyclist.
> Meanwhile, Waymo is somehow going to map out in full detail entire countries in order to ever scale up.
They're owned by the preeminent digital maps provider, who likely has other uses for that data and will be able to package and re-sell it to the other vendors that require mapping data.
It's also worth pointing out that population, travel routes and income are not equally distributed. It's almost certainly an 80/20 problem where mapping 20% of a country lets you handle 80% of the rides. They can reach ~10% of the US population by mapping a number of cities you can count on your fingers. Highways are incredibly easy to map, so it wouldn't be difficult to interconnect those cities.
Do you avoid handrails and sanitize the doorknobs and gas pumps you interact with? Those are going to be far worse than subway chairs. It's probably not a bad idea, but it's a bit beyond what I suspect most people consider normal. I had a friend who lived like that, and he ended up being diagnosed with OCD.
If you soil a waymo taxi, they can ban you from ever booking another one, the same cannot be said for public transport
While I consider this somewhat dystopian, I do think it's pretty clear that they will be much cleaner than public transport for that reason
You reached a situation where everyday things are "too dirty" and rather than realising that's a mental health problem and you might need to fix that, you... found an expensive and elaborate coping strategy which necessitates further crazy beliefs.
The lines near me have freight on them (I live in a port city, a noticeable fraction of the country's imports and exports go via intermodal containers on trains) and still run like two services to London and two to other big cities per hour. The freight has to fit in between passenger services, that's a policy decision and the US just picked wrong.
My rando observation is the few times something has gone really wrong on BART traffic is bad enough that it'd be better to take the day off. Ditto if a truck overturns on one of the bridges. And the latter happens way more often than the former. Reminds me after the Loma Prieta earthquake the bay bridge was out for a month but BART was running 12 hours later.
In places with big rail networks you're going to hear "I didn't make it into work this morning because of signal failures" really often. And when railways are repaired they will be totally closed for months. They're just way more brittle and unable to degrade gracefully.
Plenty of rail cars have room for luggage too or even baggage cars. Usually we are talking about people going to/from work however hence most people are not carrying lots.
Are you really saying your roads don't have shutdowns for maintenance? Or other problems? Also it is pretty rare for either roads nor railways to be fully utilised overnight. Also note that plenty of rail systems do run overnight.