> GM said in a statement that “... an increasingly competitive robotaxi market” were the reasons for the change.
Isn't there basically Google/Waymo and then, seemingly much further behind, Tesla Cybertaxi, Amazon/Zoox, and Uber/Yandex? Cruise allegedly has one of the most sophisticated autonomous driving platforms, and GM's Super Cruise (if they share any tech) is comparable to Tesla FSD. Strange that they would bow out.
Small anecdote: I visited a GM dealership this week and the salesperson told me Super Cruise was not enabled for test drives. The excuse was pretty weak, like the dealership would have to pay for the service or something. GM might have the technology but they are completely bungling the strategy.
Ford just lowered the cost of its BlueCruise subscription by 1/3rd. In an earnings call eight months prior they remarked they made a 70% margin on the service. It seems like drivers did not find the feature compelling and were not renewing. Interest in autonomous driving appears to be cooling across the board.
I drive (and have driven) a fair number of rental cars due to travel, and I have to say I feel that way about many new vehicles in their entirety. So much vehicle tech is at best unimpressive and at worst positively in my way as a driver.
> It's wickedly expensive
We own two cars, 6 and 10 years old respectively, yet I've never felt less motivated to look at new cars than I do now.
Guess our part towards saving the planet may well turn out to be 'driving that existing ICE vehicle for just a little bit longer'...
In the book Heat Monboit argues the stats show that when someone "greens" their consumption they also increase it. For example switching to "green" energy meant people stopped thinking about conserving.
In the specific case of electric vehicles (and many plug in hybrids) luckily they're such an immense improvement that my guess is that they still have a net benefit. But far from the simple efficiency math, and more likely the user will increase their consumption by some meaningful percent.
I wonder though if the more significant limitation for most people is the value placed on their time currently, so they seek to minimize travel and commute times (rather than minimizing for fuel costs). And if so, does will the rise of self-driving + EV paradoxically increase the number of miles someone is willing to sit behind the wheel in the coming decade?
EVs have a way higher initial CO2 footprint. The crossover depends on car and where you get the energy from. I think it's often around the 100 000 km mark where an EV gets the smaller footprint. But like I said, many variables factor in.
Production of EVs and batteries generate more CO2 before the first wheel turns, however, the total carbon footprint of ICE vehicles quickly overtake that of the EVs after 15,000 miles (24,140 km) of driving.
- It takes a typical EV about one year in operation to achieve "carbon parity" with an ICE vehicle.
- If the EV draws electricity from a coal/fired grid, however, the catchup period stretches to more than five years.
- If the grid is powered by carbon/free hydroelectricity, the catchup period is about six months.
> the total carbon footprint of ICE vehicles quickly overtake that of the EVs after 15,000 miles (24,140 km) of driving
Shouldn't we be comparing an old ICE which was already built years ago but is still operating vs building a new EV to replace it?
The carbon footprint of building the ICE is almost irrelevant, it's just the ongoing damage it is doing. Compare that with footprint of building and operating the EV that could replace it.
That would be comparing apples to oranges, you want apples to apples.
Everyone replaces their car at some point. If you are comparing a new EV, its best to compare its impact compared to a new ICE. We all know its probably best in most cases to keep any vehicle, even one that is inefficient, on the road for its useful life.
> We all know its probably best in most cases to keep any vehicle, even one that is inefficient, on the road for its useful life.
I didn’t know this until this thread. And I feel like I’ve heard people say things like they are going to improve the environment by ditching their current car for an EV, not because they’re planning on getting rid of their car anyways.
> But what if your existing petrol or diesel car is perfectly satisfactory? Obviously if you cause one less vehicle – of any kind – to be manufactured, you’re saving CO2 in the short term. But if you drive a lot of miles or your car is thirsty, then sell it to someone who drives less. Getting an EV would after a very few years move you into credit. If it’s efficient and you drive little, probably hold on to it for a while.
> In most areas of life, the greenest thing is simply to buy less stuff and keep it for longer. But with ICE cars, because they emit so much CO2 in use, it’s not always so simple.
(Apologies for the snark but...) mythbusting articles without any experimental data or references to back up claims - or claims they say they've debunked - aren't really my thing. Hopefully not really a HN thing either :/
I agree that there are shades of gray though. My counterpoint is that I am not so sure the non-commercial vehicle market is that efficient that when you sell a vehicle that it is going to the best candidate best on the miles driven. Your vehicle is still in the fleet and being used. I think the simpler argument is that for any modern vehicle, there is not a significant enough gain and you are better holding on to the vehicle until it reaches your preferred EOL.
Well pretty much everybody at some point replaces their cars. There are not many still driving around their Ford T's. So latest at that point, you will have to compare production+cost vs production+cost.
Is battery manufacturing that bad for the environment? I mean, ICE cars still have wheels, drivetrains, chassis, doors, windows, chairs, electronics, paint and many other things, same as the EVs. The electric motor I can't imagine its worse to produce than a combustion motor. So that leaves batteries as the cause of this huge disparity on day one, that takes 100,000km of fuel consumption to overtake.
Is this with the battery tech of today, december 2024? So for EVs built in 2020 let's say, it was worse? I think there is some advancement in battery making, isn't there?
> The electric motor I can't imagine its worse to produce than a combustion motor
I assumed we're attempting to weigh up keeping an ICE vehicle that already exists - like my 10 year old diesel - vs building building a new EV vehicle and scrapping the old ICE.
Nobody is crushing their ICE and dumping it in the lake when they replace their existing and working ICE with an EV. They're selling/trading in their ICE which someone else will buy and use until it's not worth servicing anymore.
Current recycling tech is pretty terrific (IIRC ~98% recovery of lithium, misc metals). Enabling the long fabled "circular" economy.
At some point we'll hit a (near) equilibrium. IIRC, ~20 years out. (Assuming Li-ion battery tech continues to improve at current rate. Thereby balancing out the lithium lost during recycling.) So we'll hit something like "peak lithium", greatly reducing mining (extraction) of new lithium.
Recycling is much cheaper than mining (both $ and CO2).
(I believe, but obliviously can't prove, that sodium (et al) batteries will enable new use cases and markets, complimenting lithium rather than replacing it.)
My guess: Most of the CO2 emissions for making cars comes from the body of the car. It is basically identical between ICE and EV. I guess a modern car body is a combo of steel and aluminium. (Is that correct?)
It suggests that making an ICE car is roughly equal to the running in terms of carbon.
In reality manufacturing is about 10%, disposal 5% and the other 85% from running it e.g. fuel and maintenance.
It then suggests that keeping an old car is better because if you keep the car for twice as long then the emissions are half per mile, but you don't scrap functional cars you sell them onto someone else. This causes a cascade with the oldest and most polluting cars being scrapped.
> In reality manufacturing is about 10%, disposal 5% and the other 85% from running it e.g. fuel and maintenance
I imagine this depends hugely on the kind of vehicle?
Our small [petrol] car cost the equivalent of $10k on the road, including all taxes. It has a tiny engine, four seats, and limited luggage space. It's fairly fuel-efficient (at least when I'm driving).
I've also recently driven a BMW 5 series (booked the cheapest "compact manual" as an airport rental vehicle, got that instead - win!) List price around 10x that of our little car. Nicer ride, of course. Bigger luggage space. Five seats.
Presumably in terms of carbon the BMW's manufacturing costs vastly exceed that of small cars, but is the carbon manufacturing cost linear with either its fuel efficiency or with its price?
That seems like an overestimate. A quick search is showing up numbers more in the 15-20,000 km region, and that's comparing a large EV running off of a mostly fossil fuel grid against a much smaller ICE (i.e. quite a conservative estimate).
The upfront is about 50% higher for an EV, so if you catch that 50% up in 1 year then you'd pay back an entire new EV in 3 years if you scrapped the old ICE one. (figures for simplicity depending upon grid and how much you drive it could be 6 months or 18 months).
But generally we don't scrap old ones.
If you instead sell the old one to someone who drives an ICE car that gets worse mpg then that's a further benefit.
In particular, grid decarbonisation means that distance to break even moves after the car is manufactured.
If you live somewhere that mostly burns coal to make electricity, your break even is determined based on the higher efficiency of electric motors and of the (energy efficient but filthy) big stack coal generators compared to the relatively clean but inefficient gasoline ICE.
But if they start building wind turbines and solar farms, because those are just cheaper and easier - suddenly that distance shrinks rapidly (maybe half) even though for you as an end user nothing changed, because electricity is fungible so you charge a car from the solar farm just the same as from a coal power plant.
> If you live somewhere that mostly burns coal to make electricity,
Which is almost everywhere in the world. Fossil fuels make up around 80% of energy production.
> But if they start building wind turbines and solar farms, because those are just cheaper and easier -
If they were "just cheaper and easier" they wouldn't need huge clean energy investments and subsidies.
I love solar, but don't let anyone sell you on the fiction that your EV is avoiding fossil fuels at any time in the near future unless you have installed enough solar on your personal residence to charge your car every day.
And if you ask the people who have done that "hey, was it cheap and easy to move your car's energy consumption to renewables?" and they reply "Yes!", please bring their story back here and share with the class.
> don't let anyone sell you on the fiction that your EV is avoiding fossil fuels at any time in the near future unless you have installed enough solar on your personal residence to charge your car every day.
Let's say doing things one way uses 10 units of something, and doing something another way then uses 5 units of that same something. Didn't you then avoid using 5 units of that thing?
Yeah, no. If this was 1984 you'd be right on the money. In 2024 "We mostly burn coal to make electricity" either means you're subsidising coal or you have no infrastructure investment to move to a better generation.
> Fossil fuels make up around 80% of energy production.
Where are you still seeing 80%? Last I saw was closer to 60%. A lot of that isn't coal, and the gasoline is not electricity production†. Gasoline is what we're replacing in the discussion, so "Yeah, but what about gasoline?" is a total misfire. Which mostly leaves methane. And that's a very different comparison to coal. There are indeed new methane electricity plants in lots of places.
A methane power plant might emit half or even a third of the CO2 of the equivalent coal plant. It's still a fossil fuel, but in terms of the ratio we're talking about for break-even on an EV, that's a huge difference.
> If they were "just cheaper and easier" they wouldn't need huge clean energy investments and subsidies.
You're seeing huge investments because they're expensive and make money, that's what an investment is. They're not investing in coal plants because that's a money loser. Yes, it's often subsidised, in my country the renewable energy schemes are subsidised via "Contracts for Difference" which have the effect of insuring the price paid for energy, leaving the problem of delivering the energy very much in the hands of the bidders, the government is only the hook for the agreed price of energy when it's delivered, this has the effect of making investment less risky - if you can make 1000GWh of energy over the lifetime of the project and the subsidy says you're definitely getting £50 per MWh, that's £50M, without CfD you can't be sure if you get paid £80 per MWh (we're £30M richer) or £20 per MWh (we're bankrupt) until the power auctions years after you've constructed the plant.
† Yes there are a handful of power plants running on this fuel, but they're insignificant at a global scale.
By that logic, also the CO2 footprint of manufacturing and transporting the fuel for an equivalent ICE vehicle, which is always conveniently left out of such calculations.
And also the footprint of building more generation stations (big) or building “renewable energy” generation (even bigger), which also conveniently gets left out.
The greenest choice most of us can make is an old, used car with reasonable emissions and that’s fuel efficient. Like a 15 year old Civic or Corolla. Or do what a colleague of mine did - he revived a first generation Prius, flashed newer software onto it, and salvaged a battery pack from a newer, wrecked Prius.
As pointed out by other nearby comments, that is not the greenest choice (which is to not drive), nor even the second greenest choice (which is to replace your ICE with a refurbished Prius like your colleague), or even the third greenest choice (which is to buy a new EV), but only barely beats out the worst non-green choices (buying a new ICE) and then only if you pay to make sure the car emissions controls and engine aren't becoming rusty and inefficient. Here's an article explaining: https://www.topgear.com/car-news/electric/mythbusting-world-...
There are also the caveats you mentioned that the analysis does assume that an EV is actually the greener choice, which is itself a function of a lot of other choices being made green also, such as whether it was constructed and powered with the most environmentally favorable choices of mining and manufacturing, most of which isn't really in the consumers direct control.
The actual scientist doing the calculations in proper LCAs do include this and production of the fuel is a notable chunk of the CO2 running cost of an ICE car (about 25%).
Relevant phrases are "well-to-tank" and "tank-to-wheel" which combine to give "well-to-wheel" numbers.
I think that analysis might have the “sunk cost fallacy” assumption? Just because we already have the ICE doesn’t make it more efficient to keep using it, over immediately replacing it with something more efficient, as long as you then keep the more efficient thing operating until it becomes inefficient to keep that operational. The crossover point seems likely it would occur much sooner if you replace the ICE now than if you wait for its end of life, regardless of the age of the ICE now
And I think this analysis suffers from ignoring the increased environmental cost of creating a new electric vehicle. It takes 15,000 to 20,000 miles of driving for the average EV to "break even" with new ICE vehicles due to the much worse environmental impact of creation.
Reduce, reuse, repair, recycle. Vehicles should be reduced first (drive less), reused second (keep using existing one), repaired (keep using existing ones), and finally you can recycle as best as possible.
All of this should happen prior to replacement. If you replace an existing ICE with an EV, the EV not only has to catch up with a new ICE, it actually has to "catch up" with your existing already made working ICE that has no new cost for construction. That's much worse than 20k miles because the cost of building your existing car is sunk. It could take as many as 50,000 miles of driving to break even against your existing used car.
Consumerism and early replacement of working goods has always been the mortal enemy of environmentalism.
Good news: in the car market, almost everyone follows 'reuse' and 'repair'. Not many people take a car and crush it for recycled parts unless it's truly worthless or super old. They sell it to someone else. For the type of person who buys a new car every 3-4 years, you are selling that car to someone else who will continue using it, likely replacing their less efficient car with yours.
This is unlike most other consumer goods which tend to be scrapped much earlier. If you're scrapping a car before its ~10-15 years old, chances are there's either something quite wrong with it, or you just drove it way too much and its gonna fall apart (i.e., something wrong with it).
> All of this should happen prior to replacement... it actually has to "catch up" with your existing already made working ICE that has no new cost for construction
That is precisely the sunk cost fallacy though: the principle is that continuing an endeavor simply because it already has been a cost paid shouldn't be done, unless the total continuing cost (including the eventual replacement) is less than the cost of immediate replacement (plus all continuing costs after initial replacement, including the eventual replacements of those in turn). Otherwise the principle says it is a waste of resources.
The 4 R's assume that the replacement is no better than the original at the job, which is why I described that analysis as the sunk cost fallacy. We don't have to take this assumption (personally, I prefer to bike over driving my ICE, so this analysis doesn't apply), but if we take the assumption that the EV does less environmental damage over its lifetime, then this assumed "environmental damage" function is minimized by discarding the ICE immediately, as any further use simply increases the total "environmental damage" caused by the choice of which car to use.
This can be seen in computers too, as newer computers are sometimes so much more power efficient then their replacement that they can very quickly save on resources by throwing out a perfectly working computer to replace it with a newer one.
As long as your old cars are like Toyota corolla/camry’s and not like giant gas guzzling SUV’s or trucks keeping them on the road is better than buying a new EV IMO. I think the current EV gen is kind of doomed because of the fact the next EV gen is going to have massive QOL features the current one doesn’t and waiting for a more compelling car is worth doing IMO
Even vehicles with poor fuel efficiency are better to keep on the road until EOL. This is why cash for clunkers was such a terrible program.
I don't follow your take on the current EV gen either. A car at the end of the day derives most of its value from taking you one place to another. The current gen of EVs do this and do it well. "current EV gen is kind of doomed" that is just hyperbole, nothing is changing in the near future that would doom current EV vehicles on the road.
Define massively. I think a 2024 M3 has ~20% better range than a 2018 M3. 2018 was the second year of the M3. The 2018 LR version still had a ~300mile range. No doubt future generations will be better but again, I don't think we will have a case that "dooms" the current generation of EVs, they will serve their primary purpose well until EOL.
I think proper V2H/V2G (vehicle to home/grid) is a big one that is very rare at the moment but is probably making a lot of people who could buy an EV at the moment (i.e. have a garage/driveway in a city) hold off until it's more common, because it makes the economics a lot more attractive.
(In general, because of cratering manufacturing costs, improving tech, and a still low trust for battery health, being an EV early adopter is an expensive option at the moment because even if an early EV is still at 80% functionality compared to new, the resale value is terrible, and I think this plus the lack of charging infrastructure is a big impediment to adoption growing to become dominant in the mainstream)
I'm not super convinced about the V2H stuff, tbh. I only need backup power at home to keep the heat on when I'm not there. If I'm not there almost certainly my car isn't either. When I am there I can light a candle and run the wood stove to keep the pipes from freezing. So a stationary battery bank for my solar system makes a lot more sense than trying to use my car for that purpose.
And probably a backup generator with a big propane tank is a better solution.
I've thought about it over the years and if I was regularly away for long periods in the winter I'd probably have spent the money. As it is, I cross my fingers and generally try not to have long periods away from my house in the winter.
Ah. I think lots of other people would like to keep their refrigerator and freezer and appliances running in the event of multi-day power outages when they are at home, even when it is not freezing (like in Florida, for example).
Sure, but if you're already going to make the investment in the grid cutoff, inverter, etc then the fact the car could in principle act as a battery/generator is a little gimmicky. Might as well also get a generator (or battery if you've already got solar). Then it works automatically whether you're there or not.
The vehicle already has the inverter, using the same active rectifier hardware in the on-board charger. People want to have electricity when they are at home. And they are home a lot of the time. So I agree that it is a very niche case to want to have a backup system to provide electricity to your house during a natural disaster when you are not even there.
Whatever you call it based on exact details, my point is that using your car for power is not "gimmicky" like they said. Based on budget you can get a transfer switch or you can get an extension cord. Either way you're making effective use out of the car you already own, and the idea that you're wasting time and should really just get a generator or fixed battery is incorrect.
Fast recharge and high capacity long life batteries.
Reliably getting 300-400 mile range (even in cold), getting 250 miles range in 10-15 minutes of fast charging, and seeing batteries comfortably make it to 150k miles with small degradation.
Unfortunately people seems to really care about edge cases (What if I have to unexpectedly drive 400 miles on a frozen January evening?), so until EVs can easily cover those bases, people are really trigger shy.
The biggest one I can think of is really stupid actually which is a NACS port and proper supercharger network access for any non Tesla (and I personally would recommend against buying a Tesla for build quality and lack of repairability). Either way ICE cars are basically about as good as they can get while EV’s have a ton of room for improvement
> So much vehicle tech is at best unimpressive and at worst positively in my way as a driver.
Honestly, I'm not surprised that new cars aren't breathtakingly awesome. It's been a really rough 5 years for everybody and we're not seeing much genuine light at the end of the tunnel. Everyone's just trying to survive. Hard to do your best work under those circumstances.
Its not what the public think of 'self driving'. Its not like you can tell the autopilot where you want to go and then sit back, sleep, turn around, do something else. You have to monitor it.
So you are basically paying to take legal responsibility for the crashes of a not-quite-ready autopilot whilst still having to keep alert.
The moment there is a true self-driving capability, where you can sit back and let the car drive even if you're not sober etc, then it would be wildly popular.
Exactly. My car has adaptive cruise control. In slow traffic, if the car in front comes to a stop, it'll also stop. But what happens if it's turned off? I honestly don't know and don't have the balls to find out, because so far it just keeps getting closer to the car in front. So if I accidentally pause it then drive the same way I would with it enabled, I might(?) rear-end someone.
It'd be the height of stupidity if it does just keep rolling forward, but who knows...
When I bought my last car, I was sort of excited about adaptive cruise control and lane keeping. But maybe because most driving I do (which isn't a huge amount) is on pretty busy roads I've sort of avoided really using it.
I’ve had a Yukon with supercruise for 18 months. It is fantastic for highway driving, particularly, we’ve driven it west from Tennessee to Colorado and Florida. Other comments claiming it’s the same as other standard driver-assist packages are incorrect. Totally hands free, no “ping pong” with the lanes.
That said, when trying to BUY a supercruise vehicle, the sales guys were clueless, I had to review the stat sheets of each car to see which ones did or did not have it. GM is treating this technology as “surveys say 2% of the market wants a self driving vehicle” (incorrect question) compared to “install on every vehicle as an available subscription, let the sales team earn their way.” I’m sure the tech is expensive, but it can’t be that much more at the OEM level.
I also have a Tesla with FSD. FSD is truly amazing but still struggles with edge cases like, my office has two entrances, one is blocked by barrels, it tries to turn into the barrels every time.
From my understanding, super cruise only works on pre-approved roads/road sections, whereas FSD works just about anywhere. When I last checked my commute to work, the main highway I take becomes unsupported half way to work, despite it being the same bland typical highway layout.
I really don't want to buy a Tesla, but from what I can tell, nobody has anything close to what FSD offers.
> Down to the point of unwanted veering towards road dividers and other vehicles.
Our 2024 Kia does this every 10-20 miles on average. You don't even have to enable lane assist. It also randomly decides to weaken the regenerative braking, or switch out of single pedal mode.
If you disable enough of the safety features it stops doing these things, but, for instance, that means you can't have blind spot detection that works when you're going over 5mph and a working steering column at the same time.
Yeah, people act like some Tesla problems are bad, but I've certainly used systems that are worse. Just in a few miles, I've had a Nissan accelerate towards stopped traffic aggressively enough to trigger its own collision warning. It'll also fly try to drive off the road in relatively shallow curves, triggering its own lane departure warnings.
The Tesla system isn't perfect, but it is far from unique in being imperfect.
That's true. I took a jab specifically at Tesla, because it's generally framed as the "leader of the pack and problem free to the point that you can sleep in it".
FSD is a very hard problem, and the systems deployed are at best alpha level. These things should be "hyped responsibly".
Not paying attention while FSD drives for you ended with a motorcyclist crushed to death under a Tesla, and a manslaughter charge for the driver near me this summer.
SuperCruise does what it claims to do, keep it up the middle on listed corridors. Most highways and interstates are eligible. I love both systems.
FSD is truly almost there for unsupervised, point to point driving. People want to virtue signal and complain about Tesla but it is an amazing piece of hardware being mass produced today. My car is 3 years old.
I worked in this industry. No, FSD is not almost there. Not even close. What matters is the long tails of events.
You might "feel" it is almost there because it gets it right 99.9% of the time but that is still way too many accidents and injuries in the long run. And the work to go from 99.9% to 99.9999% is 1000x more complicated.
We need to compare against the real human error rate (including drunk drivers, sleepy drivers etc). What is that error rate?
Also, fault tolerance error rates don’t work that way - difficulty increases exponentially as you increase fault. In other words, it’s much more difficult than three orders of magnitude to go from 3 9s to 6 9s - it’s easily 5-6 orders of magnitude.
When you're playing candy crush and FSD kills someone, whether or not it's drives better than the average driver is not something the judge and prosecutor are going to consider.
First, that’s because FSD is legally only L2 and thus legally you are required to pay attention. It has nothing to do with the engineering realities of the impact on improving safety.
When L5 becomes available then this becomes a different calculus. And it’s honestly questionable about how good the characterization actually is since Tesla’s L2 FSD seems to outperform Mercedes’s L3 and it’s more a matter of the liability the manufacturer is willing to take on vs objective measurements of quality.
yes, the long tail of difficult event is exponentially more difficult to handle. That's why I said above that people "feel" it is ready but it is nowhere to be even close to ready.
The average crash rate for human is one every 500k miles.
100 Miles literally don't matter. Even 1000. On average accidents happens every 500k miles.
Are you ready to have your tesla drive FSD with you sleeping in it for 500k?
You are letting your feeling dictate that FSD is ready. The math is more complicated.
Got it. differing expectations. Im not expecting 'unsupervised' out of the system, I dont think I would trust any system to transport me or loved ones un-piloted. Even when being driven in a taxi, I am still supervising despite not having much recourse beyond barking at the driver.
That said, based on my experience with FSD, I'd be tempted to take a nap if it let me, reinforcing my initial statement that it's very darn close.
If you drove with a very very bad driver that crashes every 50k miles (average is every 500k miles), you would have exactly the same feeling and you would also be tempted to take a nap.
Indeed, and I've held an anti-Elon position since well before he bought Twitter and went all MAGA. Personally, I don't want to support someone who continuously lies and breaks securities laws with impunity. The MAGA stuff is just icing on the cake.
This i what is so tricky with Self Driving. People "feel" it is almost there because most of their rides are mostly ok. However to make a system truly driverless you need to master the long tails of difficult events and FSD is nowhere even close to do that.
Going from 99.9% to 99.99999% is what makes a system truly driverless and where most of the work is. Waymo is way way ahead of FSD for this.
I know essentially nothing about machine learning, nor about what approaches to ML that Tesla or Waymo are using. Is there a Metcalfe-type effect possible here? Where the better FSD becomes, the more people use it, finding more edge-cases, so the driving get better, so more people start to use it. With the end result that the learning starts to get better/faster with time, in a positive-feedback like mechanism?
Your comment/question seems to assume that somehow accurate classification of performance is possible, which is demonstrably false assumption.
First, if somehow system-internal watchdog was able to detect erroneous outputs for training, it would also be able to stop those outputs propagating to control system live, leading to zero errors. Such a watchdog requires self driving to be quantified analytically, leaving implementation entirely testable and therefore control system not passing the testsuite (i.e. exhibiting erroneous behavior, i.e. still having "edge cases") not deployable in public. Many words to say that that in practice with driving being somewhat loosely defined you need humans in the training loop.
Second, vehicle operators are not only unqualified to accurately monitor behavior of these autonomous systems, in part due to not being formally trained on safety systems in general and the system they are monitoring in particular, but on top of that incentivized to underreport erroneous output.
My prediction has been that we are going to see increased number of accidents with ADAS deployments, not less, before the number can start dwindling.
> Small anecdote: I visited a GM dealership this week and the salesperson told me Super Cruise was not enabled for test drives. The excuse was pretty weak, like the dealership would have to pay for the service or something. GM might have the technology but they are completely bungling the strategy.
Sadly this is believable, I've asked to check out things like remote start and control of heating/cooling, and the sales people cannot show those features off because they require an app + subscription tied to the car.
Based on my anecdotal experience, another issue is that salespeople are not trained on the tech
I worked on a feature for new vehicles and the company failed in part because buyers simply didn’t know those features were part of the vehicle. Dealers never set it up for the buyer and it wasn’t something many people would think to do on their own
A salesperson isn’t going to jeopardize an easy sale by bungling some fancy new feature they can’t control
I’d guess this is at least half the features on my iPhone. I’m sure it can do things that I’m not even thinking a phone can do, but nobody set anything up, and it’s not very obvious or discoverable.
At least when you purchase an iPhone new, Apple does push a "Personal Setup" session with their staff either in-store or online https://www.apple.com/shop/personal-setup (click the "Feature Focus" tab)
What tesla salespeople? They're pretty well known for flaunting the dealer model. I've never bought a tesla, but from the friends that have bought one one of the big plusses is you don't talk to a salesperson at all, you buy the car at a fixed retail price online. Has this changed recently?
Tesla has lots of showrooms with cars, test drives, and sales people. If you're there and want to buy a car, they direct you to a computer with an Internet connection.
They have a couple people to manage test drives and they will ask if you have any questions and make sure your aware of the current interest rate or other sales incentives but they aren’t sales people in the sense that you would typically be used to.
I don't want to be a jerk but the salesperson couldn't pronounce "autonomous" which tells me they aren't being trained in selling the feature at all. I can't even remember him referring to it by the marketing name.
I've bought a bunch of new cars in my life. I've never had a sales person who knew basically anything about the cars they were selling me. I have however had a variety of very good salespeople, their value has never been in explaining the car to me but in making my purchase transaction straight forward and reasonably quick.
All of the sales people could tell me how many doors each car had and what color it was when we stood next to it, but that was about it. Many of them could look up if other near-by dealers had the exact combination of options I wanted.
If sales person training about the actual cars exists, my experience indicates very few new car sales people take advantage of it.
Concur. A Toyota salesperson was once unable to tell me what the highest trim level of a Highlander was and how it differed from the model they had. Seen staggering lack of knowledge in all things, not just tech.
Older example but when I got a second hand Mazda in ~2017, the salesperson I bought it from set up Bluetooth pairing for voice calls via my phone.
As it turned out, the feature was still pretty janky and I only tried it a few times before reverting to just regular speakerphone. And when the battery died and the head unit lost all its config, I never bothered figuring out how to set it up a second time.
The whole dealership system is parasitic, rentseeking, friction. Tesla might have a point here.
Instead of transparently selling a product for fixed price, the dealer system appears based on information asymmetry, haggling, upcharges, finance bullshit, warranty bullshit, subscription bullshit, and many decades of entrenched psyops culture against customers.
On top of that, salespeople are often poorly trained on the products and dealerships seem to have an adversarial relationship with corporate, especially around the corporate website differing from the local story.
And then the dealerships steer you for whatever benefits them. In 2017 I tried to test drive a Chevy Bolt: Motortrend's car of that year. One dealer hadn't even heard of it. Another said he couldn't get one. Another tried to dissuade me from looking at it by dissing the product. Finally I found a knowledgeable dealer that knew the product, had some, and revealed they were easy for all dealers to get in our area: the others were just being obstreperous. Suck.
It's really annoying to me because I would love to just pick a dealership and go there for the rest of my life for all things automotive.
I don't like being ripped off but I'm not particularly price sensitive, so I'd love to just show up and pay a reasonable sticker price that was just The Price Everyone Paid and not have this vague haggling system expected, and not have to worry whether I was being ripped because I wanted a particular option. I hate having to think about decisions I don't particularly care about, so I would prefer to constrain my choices to "maker Y's automobiles", be it GM, Ford or Toyota for the rest of my life, so that once I have the capabilities I need out of an automobile there's only one or two choices. I would love the simplicity of the dealer being the default maintainer of my automobile and I just show up once a year or so and they take care of it in a pleasant experience.
It's irksome instead that it feels like they just want to trick you out of as much money as possible up front and that they don't want to have a maintenance department, but are legally required to.
I think service departments are actually pretty profitable for at least ICE dealerships. They certainly put the old college try into introducing you to the service manager even when you've purchased your car in a situation where you're unlikely to use that dealership for service.
My last purchase during the supply chain shortages was actually remarkably stress-free other than not being able to get the car sooner rather than later. I was paying cash. Knew there was almost certainly very little negotiating leverage on the list price. Did a bit of haggling on some "factory-installed" add-ons with a car that hadn't been built yet against my trade-in which I got IMO a good deal on.
The whole thing took me maybe an hour which is hard to complain about for a large purchase.
One of the keys is being ready and able to sever the actual purchase from the financing and the trade-in. It's when people really are stretching their financial envelope (which is common) that conflating various aspects of the transaction starts causing a lot of stress.
Of course, it also helps if you've been around the block a few times. OK, if I MUST I'll watch the stupid video about some coating product but know that the answer is no. It's annoying but no more than lots of other things.
This is definitely something that Tesla nails imo. No haggling, a single price listed on the website that changes with market conditions. A list of options available with again explicitly listed prices. No slimy car salesperson trying to upsell the scotchguard on your carpets bs. No hidden prices where the websites lists price $x but when you get to the store they tell you its $x + y. A price, a delivery and its over.
Brad Templeton's Robotaxi Timeline[1] (October 2024) shows a dozen milestones that Waymo has achieved over nearly 15 years from "Make a nice video" to "100,000 Rides per week." Waymo needs two more milestones to achieve "Production!" Waymo's competitors are behind on the timeline, but may move faster thanks to improved technology and a more welcoming social environment.
GM hasn’t been a car company since the 1980s. It’s a finance operation. Leadership only cares about developing products that are good enough to sell financing.
They really don’t know how to develop a compelling product, even though they have an enormous amount of experience and engineering talent.
It’s very similar to the failure of Boeing but GM doesn’t make anything that can kill 200 people at once so we don’t notice.
I’m one of the people that did not renew BlueCruise and the reason is kind of weird
BlueCruise vehicles were updated with the most sophisticated driver attention system ive ever seen. It will complain in less time than it takes me to change a radio station.
Before that update, I was enjoying the automated driving a little bit too much, and using it to take liberties with my attention in the car. That was awesome as a user experience.
After the update, I am paying more attention to the road than I did when I was driving a manual. There is no benefit to being able to take my hands off the steering wheel. That’s a good thing for safety, but it means that “level 2.5” driving (or whatever) adds zero marginal utility.
Ford brags that BlueCruise gets top marks from Consumer Reports but I think it's mostly because of because of this driver monitoring, rather than actual driving performance.
> Starting with 2024 model-year vehicles, we will deduct points if an [active driving assistance] system doesn’t have adequate [direct driver monitoring systems]. Right now, only Ford and GM’s systems meet our criteria for earning additional points, but others could be available soon. ...
> Ford’s BlueCruise sets a high standard among [active driving assistance] systems, aided by an infrared camera that monitors the driver’s eyes to determine whether they are looking at the road. If the driver glances away from the road for more than about 5 second... the system will give the driver a visual warning and an audible chime.
Supercruise might be comparable to Tesla's "Autosteer" product but its incomparable to Tesla's paid "FSD Supervised" product. The latter is closer to a Waymo or Cruise than it is to a Mercedes or GM driver assist.
> The latter is closer to a Waymo or Cruise than it is to a Mercedes or GM driver assist.
It depends how you look at it, but in the (super limited) cases where you can use Mercedes Drive Pilot you can legally and safely read a book, watch a movie, or work while in the driver's seat. [1] That's not the case with any Tesla product.
Yes there are actual "legal" notions of driver assistance "levels," but out of those 3 companies Tesla seems to be taking it the most seriously. Elon has a high conviction in the product and has continued to pour billions of dollars in GPUs, talent, supply chains, etc to make it happen. Other manufacturers don't take it nearly as seriously, unfortunately.
Sure you might not be able to legally claim that a Tesla is totally "autonomous," but the fact of the matter (to me, at least) is that they are putting in way more effort to solve the problem than legacy auto manufacturers are.
I can step into a Tesla today, press a destination, and go there without touching the wheel or pedals. Sure it won't be flawless but the fact is, I can. I can't do the same in any other consumer car, and the closest thing is a Waymo. The effort is there, I think its just a matter of time before we start seeing the legal stuff play out.
I think there's some chance that Tesla's approach will work out for them, but I'm not optimistic. Getting self-driving software to "impressively good" happened pretty quickly, and even systems from 2016 would usually pass your "go there without touching the wheel or pedals" test. But from there to "as safe as a human" let alone "as safe as we require human-replacing machines to be" turns out to be quite a hard gap to cross.
When you say "what was available" do you mean what a Tesla was capable of? Because I was trying to talk about things like Cruise driving for 90min at night in a city with no intervention: https://youtu.be/KSRPmng1cmA
If you let me pick the roads, I could easily drive 90 minutes at night with no intervention in the LA area: driving from Ventura, down the 126 to the 5 to LAX can be done basically with no interventions no problem today. It could be done with EAP two years ago too. But, FSD works well on essentially arbitrary roads and it didn't four years ago.
> driving from Ventura, down the 126 to the 5 to LAX
That's a lot simpler than what Cruise showed in their video. Dealing well with pedestrians, stopped cars, cross streets, etc are a significant challenge.
Well, there's consistently and there's consistently.
In 2016 Waymo reported their safety drivers intervening once for every 5,100 miles driven [1] - which implies to me that 99% of journeys nobody touched the wheel or pedals.
The problem is 99% isn't enough, as there are tremendous numbers of cars out there, and a busy bit of freeway would get a disengagement per minute.
It might be missing a few logging roads, private driveways and carpark lanes - but it shows Google is entirely capable of mapping streets in detail all over the world.
> and even systems from 2016 would usually pass your "go there without touching the wheel or pedals" test
This is complete bullshit. The systems in 2016 were restricted to pre-mapped areas with lots of training. (i.e. the waymo/cruise approach today) So if you weren’t in a tiny slice of San Francisco or a few other training areas, this didn’t work.
Even if an area is pre-mapped, if you're operating on public streets you need to handle all sorts of unusual things. Here's a Cruise video recorded in 2016 on SF streets, showing a tricky interactions with buses, stopped delivery vehicles, pedestrians, etc: https://www.youtube.com/watch?v=1Tp6Ubf6mE4
Yeah, I’m not saying they weren’t doing impressive things. I’m saying they were still severely limited and would not be capable of a “go around without touching the wheel/brakes test” like a Tesla can in the entire US now.
> of those 3 companies Tesla seems to be taking it the most seriously.
You call lying about it endlessly, misleading marketing, beta-testing on public roads, and not even having L3 properly in production "taking it the most seriously"?
No, I'm mostly referring to buying talent and GPUs.
I can't point to a single Andrej Karpathy or Chris Lattner or Jim Keller working at Mercedes, BMW, GM, etc (not to mention the not so big named people who are still very very good). And I also don't see many people crossing over from legacy auto autonomous orgs to openai, anthropic, deepmind, etc.
Other manufacturers don't have custom in car inference chips either, or spend billions in R&D for custom training silicon. This is clearly not a side project for Tesla, whereas with other manufacturers, its an obvious afterthought.
My guess is other manufacturers will just license some AV product from whoever is most successful and try to sell products like they always have - through design language and brand feel, not through breakthrough technological innovation.
So yes, I don't think any other consumer car manufacturer is taking it seriously.
From what I've seen, Tesla FSD would fail a driving test here in the EU within the 1st minute. Even the latest and greatest version behaves like a drunk teenager.
I'm not in EU so I can't comment if they have a special circumstance that makes there driving test different than elsewhere in the world, but "behaving like a drunk teenager" is not a descriptor I would use. I make use of Tesla FSD every day, 2 -3 times a day. Over the last few months I've had to engage (of my own choosing) twice, otherwise it is completely hands and foot off experience.
I travel hundred miles a day on average on a mix of local and highway, but all major roads in city and suburb.
Driving tests in Europe, certainly in northern Europe and Scandinavia, are considerably stricter than in the US and many other places in the world. In Norway the test is also not enough, there is also obligatory practice with a qualified instructor including driving on motorways and on a skid pan to simulate driving on ice.
Would definitely pay more to avoid FSD pilots, because the texters sense their surroundings more often than the drivers lulled by false advertisement and (at-best) beta grade software.
The snarky posters can stay, doesn't hurt my life or property.
Or course it would fail a driving test designed for humans.
Humans would fail a driving test designed for autonomous driving too. We don't have the reaction time comparable to computers to ..say.. avoid oncoming traffic or an object/animal on the road within a few milliseconds. Or identify dark objects easier to see in infrared. Or maintain an exact speed, to a few significant digits. Or manually do a good job with traction control.
> Yes there are actual "legal" notions of driver assistance "levels," but out of those 3 companies Tesla seems to be taking it the most seriously.
..and yet Tesla is quick to blame the drivers when accidents happen - and has disclaimers for supervision in its TOS. Mercedes on the other hand, takes all responsibility for eventualities that happen while self-driving is engaged (DrivePilot - not ADAS)
And in those same cases, you could probably safely do the same with FSD (s), but because FSD isn't as limited and the Mercedes product is not a serious competitor, there's no compelling reason for allowing it.
The MB Drive Pilot only works at ridiculously low speeds (under 40 mph), on select highways, no sharp turns, no weather, no construction zones, lanes clearly visible, etc. It's laughable to compare it against FSD which works everywhere. FSD has been flawless when driving on such easy conditions for years.
Are you assuming MB Drive Pilot couldn’t also work pretty well outside of its design domain?
This argument is always so silly because it’s just an argument about the correct way to build a safe control system (i.e. with guaranteed performance characteristics in a defined operational domain) versus the incorrect way to do it (just kinda work over an ambiguously defined set of conditions and hope you don’t kill too many innocent people to cripple your business).
MB Drive Pilot is designed not to attempt to operate outside of a well-defined domain.
It will, by design, not operate in a way that is unsafe or in a context in which it's unsafe.
You are arguing the system would perform worse in contexts in which you have not seen it perform (because its designers decided to prevent it from such attempts).
What evidence do you have for your argument?
Restated: You are assuming that if Mercedes engineers had a similar appetite for putting millions of innocent people at risk by operating in unreliable contexts, that their system would perform far worse than FSD does. On what basis do you make that assumption?
The fact that their software is written only to just follow the car in front, and completely and immediately give up if there's no car in front or one cannot be seen due to it being dark or bad weather, on premapped roads.
Meanwhile FSD has repeatedly shown that's it's capable of driving without a car in front.
You can watch YouTube videos of FSD driving without a car in front for thousands of miles on non-premapped roads.
What makes you think Mercedes would be better at doing that?
What makes you think that MB doesn't have the same capability, but they aren't willing to ship a half-arsed highly dangerous, people killing feature like some other companies do? Unlike any of their competitors, they assume full legal liability for their car's behavior while Drive Pilot is enabled.
MB also has a luxury brand reputation to maintain. When they ship a major feature like that, it absolutely has to work 100% as advertised.
> Unlike any of their competitors, they assume full legal liability for their car's behavior while Drive Pilot is enabled.
Source? If someone gets killed when the system is driving, there is no legal framework for you as a driver to escape criminal penalties. Who at MB are they going to jail for it?
Can UPS indemnify their drivers from involuntary manslaughter charges?
I've worked on competitors to the mb system from legacy auto; generally the way we work is make the best possible system for the long term, then limit the odd to meet safety, regulatory, and brand risk requirements.
So there will be a system happily running producing target driving paths, but some monitoring system disables in the case with no car in front.
I would guess that the mb system is in fact ok without a car in front, but didn't hit their appetite for risk yet.
> You are arguing the system would perform worse in contexts in which you have not seen it perform (because its designers decided to prevent it from such attempts).
Should we also assume that Mercedes cars can fly better than a plane, because we have not yet seen them fly?
What kind of logic is that?
In other news, OpenAI has solved AGI last year, but is keeping it a secret because it's too dangerous.
I have solved self driving, how can you say I didn't if you haven't seen it perform?
No, but generally you should assume a system with near-100% performance within a well-defined ODD doesn't fall off to near-0% performance immediately adjacent to that ODD.
Here's a better example: let's say by law Mercedes has to restrict their cars from exceeding 80mph. They have a car that can reach 80mph in 2 seconds. Is it therefore credible to claim that the car is actually incapable of driving at 81mph?
Or is it more credible to say, "we don't know much about its performance beyond 80mph, but it can probably achieve something outside of that."
A car not being in front means more data for cameras because they can see more of the road ahead if the system can actually do proper road navigation.
The fact that it shuts off immediately shows that they're just copying what the car ahead does.
The system is unable to function on a pre-mapped highway on a clear day with lane markings if there is no other traffic on the road. What does that tell you?
What kind of messed-up logic is that? How many products do you think you own that contain hidden features that you're not supposed to know about? Just about everything with at least a microcontroller inside has extra software modules that you never learn about as the end user (e.g. service modes and factory calibration routines).
FSD isn't driving though, you're driving. Mercedes is driving itself and you have 10 seconds to takeover. Huge difference between level 3 and level 2. One is self driving and one is just a drivers aid.
> Mercedes is driving itself and you have 10 seconds to takeover. Huge difference between level 3 and level 2. One is self driving and one is just a drivers aid.
None of that is true, it just follows the car in front in slow moving traffic during daytime in good weather on premapped roads and cannot even change lanes. No car in front anymore? It completely fails to "self"-drive.
These limitations are taboo to talk about because the Mercedes system is used as an anti-Tesla talking point, you can get permanently banned from Reddit and BlueSky for bringing it up. Thats why so many people think its a great system.
Edit: Downvotes for bringing up inconvenient facts.
I'm very much aware of how Drive Pilot doesn't handle most cases: in my comment I described it as "super limited" and linked to where Mercedes describes the details.
But within the range of cases where the Mercedes system is applicable (which recently went up to 59mph in Germany [1]) it's solidly better than the Tesla system because Mercedes did the engineering to make it reliable enough that you don't have to supervise the car's driving. If I had a highway commute in stop-and go traffic I'd be comfortable reading in a Mercedes, but not in a Tesla.
If Tesla were to artificially set up limitations like highway only, daytime, clearly visible lanes, good weather, car in front, no construction zones, no lane changes, etc, then they would easily have been able to offer full L3 and take full liability for it. Tesla FSD is practically perfected, with no accidents whatsoever, under such conditions already. But Elon wouldn't want to release this kind of gimped system as he is aiming for full self driving everywhere.
Absolutely untrue. It's not even close to perfect. I use it a few times a week and while it's seen some significant improvements over the last year, it still makes pretty dangerous mistakes, especially on left turns, right-turn slip lanes, intersections with flashing yellow lights, streets with very worn road lines or where the positioning of the road lines shift from one side of the intersection to the other. I can think of a dozen other similar situations so I'm very sure there are many others I've never even encountered.
It remains to be seen how much juice Tesla can squeeze out of a transformer approach to autonomous driving, but it's by no means a sure thing.
edit: I misunderstood the comment. I see now that "practically perfected" is a reference to the ideal conditions mentioned in your first sentence.
Thank you to post your personal experience. One pattern that I have noticed about Tesla FSD stories on HN: it is black and white, with little grey area. Either people live an area where the roads are easy for FSD to navigate, and they come here to say "It is perfect!". And vice versa: People live in a place with a bunch of complex roads and intersections where FSD does not perform well... And they come to HN to share their experience. (Privately, I cannot wait to see self-driving models try driving on Jakarta or Napoli. It will so much fun to watch those on YouTube!) I do think that Elon/Tesla is taking a crazy gamble to release FSD early to gain billions of hours of training data! I can understand where this would make some people uncomfortable, due to the safety concerns. Except Waymo, who else has the training data that Tesla has at this point? Wiki tells me that FSD has been driving on public roads since 2016. That must an astonishing amount of training data accumulated in the last 8 years! I assume he will be tasking X.ai with improving the FSD model using this enormous training data.
This is so absurd. If they could accurately predict when their system is so reliable, they could get it certified for Level 3 use in these conditions. That's not a gimped version, that's strictly better than the product they offer.
They can't though. Mercedes-Benz can. To pretend that Tesla doesn't offer a better product because they don't want to is...
There are plenty of FSD videos on YouTube where the driver didn't have to intervene at all on long trips, where the car isn't just following the car in front.
You hit it on the nail, the driver didn't have to intervene but they had to be ready to take over at any time. If FSD could safely operate where you had 10 seconds to take over for level 3, why doesn't Tesla go get it certified for L3 driving. FSD isn't even approved for being a drivers aid(hands off, eyes on) in EU and China because Tesla has not demonstrated that it is safe. Even Blue cruise is approved in EU for comparison. And it isn't because Tesla isn't trying to get ADAS certification in these markets.
Are they? they haven't start certification for self driving any market? It seems FSD is more of way to pump the stock. Mercedes is testing their same cars for L4(point to point) in china.
For past decade, with meh results consistently so far. They were/are? selling FSD as premium package and never delivered on promises, that's outright fraud in plain sight.
Technologically Tesla is far behind since one man's ego wants to trump physics and computing limitations and its failing. Still no Lidar.
I don't care about 90 youtube videos of long drives with no intervention. I care about a million out of a million (and maybe 10 on top). Either there is something I can trust with mine and my kids life with, or I am not interested with anything on top adaptive cruise control (since I don't spend my life in stop&go traffic).
I thought supercruise only worked on "qualified roads".
Tesla autopilot 1.0 has been around on tesla since 2015. It's actually pretty good on the freeway as traffic aware cruise control + autosteer. I think it's a pretty good balance of driver + assist.
Which tells approximately nothing about their relative capabilities. The fallback system is the driver (who needs to be ready to take over immediately) and the range of conditions for autonomous operation is better than Level 1.
Tesla can't self drive in any condition on any street. Hard to compare that to a level 4 company where the car can go hundreds of thousands of miles without disengagement. Being a owner FSD, its no where near being able to drive itself without supervision, it tries to run reds, misses people in cross walks and can't understand signs like "Do Not Enter".
> Hard to compare that to a level 4 company where the car can go hundreds of thousands of miles without disengagement
This is nowhere close to being true. The latest numbers I could find for 2022 are 17,060 miles per disengagement[1] for Waymo. And even then the definition is not quite clear, because these are safety disengagements from what I can tell, and not disengagements for things like getting stuck because the road is closed or something else.
> This is nowhere close to being true. The latest numbers I could find for 2022 are 17,060 miles per disengagement[1] for Waymo
Tesla doesn't reveal it's miles/disengagement stats, but according to crowdsourced data it's 690-828 miles between critical disengagement [1], which is 2 orders of magnitude worse than the Waymo number you posted, and far below their goal of besting human drivers' abilities.
Your 1 accident per 500k miles stat is an average across urban and rural areas. But accident rates in urban areas are likely significantly higher than in non-urban areas, due to higher vehicle density/congestion, despite the fact that vehicle accident fatality rates are higher in rural areas[1]. Urban areas tend to have a lot more low-speed no-injury or mild-injury collisions. When you account for all of those, AVs are in shooting distance of human driver accident rates in urban areas. Also, the accident rates are much higher for drunk, sleepy, or distracted drivers, so to the extent AVs reduce those types of driving, they bring the overall accident rates down much further.
Arguably, it would be a better comparison if the Tesla was restricted from highway driving. I understand the letting the car figure out its own route, but I’m more interested in seeing the Tesla navigate city streets.
Are we forgetting that this only happened because they got in huge trouble with regulators and had to withdraw their cars from the street for like a year? I wouldn't take their statement at complete face value and don't see any reason for pessimism for Waymo.
Yup, there is zero chance that the market is too small. The only issues are whether a company can the technology to work and obtain regulatory approval.
Size of the market is for all practical purposes function of price. The competition to robotaxis is first and foremost humans. The major question is if robotaxi companies can provide the service at the cost of ubers *profitably*.
"In the early 1980s AT&T asked McKinsey to estimate how many cellular phones would be in use in the world at the turn of the century. The consultancy noted all the problems with the new devices—the handsets were absurdly heavy, the batteries kept running out, the coverage was patchy and the cost per minute was exorbitant—and concluded that the total market would be about 900,000. At the time this persuaded AT&T to pull out of the market, although it changed its mind later. "
And in the more recent past, we asked Juicero if squeezing a small volume of bags by hand could be done more cheaply by machines, and they told us "absolutely!"
We know how big the taxi market is and it's growth rate. There is clearly room for a few businesses here alone. Then consider driverless will go beyond taxi to general transportation like trucking which is massive market. Also likely play a significant variable in what cars consumers choose.
I think the risk here is software tends to a winner (or small number of winners) gets all market.
That has to be a major risk/reward concern on the companies investing in this tech.
>Cruise was not enabled for test drives. The excuse was pretty weak, like the dealership would have to pay for the service or something. GM might have the technology but they are completely bungling the strategy
Before the iPhone came out, some pretty decent/popular smartphones existed. I had clients in the industry. They sold a lot of n95s and such.
But... data plans sucked. Apple did an exclusive deal and forced them to include 2gb of data. They could not sell phones with data-less plans. Customers could not buy iphones on data-less plans.
> Interest in autonomous driving appears to be cooling across the board.
I've had one Tesla for 6 years, another for 2: I've gotten to a point where I find Autopilot boring and turn it off just for something to do.
My older Tesla has enhanced Autopilot. As impressive as it is, it's very glitchy. I primarily use it so I don't speed.
Because enhanced autopilot was so glitchy, when we added a 2nd Tesla to the household, we didn't pay for FSD. (Enhanced autopilot was too expensive given how glitchy it was in my older car.) The free trials of FSD in the 2nd Tesla are impressive... And glitchy. I constantly need to take over on surface roads; but it is very nice on freeways.
The thing is, "self driving" as a feature just isn't worth the sticker price. If you're sitting in the driver's seat, and there's nothing to do, you can just save yourself a boatload of cash and drive the car yourself. Especially if you have to remain alert at all times, the best way to cut the boredom is to drive the car.
self driving makes total sense if its the sci-fi like dream of being able to read the news or a novel in the drivers seat while your car safely takes you to your destination.
But if I have to maintain a hand on the steering wheel and monitor the car I may as well be driving.
Any of the advanced lane keeping systems are immensely helpful for lowering fatigue, especially on longer drives. I think it's quite hard to recognize how fatiguing it is to constantly make sure your car is centered in the lane with small hand movements. In my experience, long car rides are substantially less fatiguing when I just have to 'oversee' the car, rather than fully drive it.
And when you go to Tesla for a test drive, they first thing they do is shove it into auto-cruise or whatever they call it whether you want to or not! At least that is what happened to me when I test drove one 10 years ago. Maybe they don't do that anymore.
I test-drove a Model S about that time - it was hair-raising because the system would correct the cars position about a half-second after I would have. I hovered my hands by the wheel the whole time, waiting for it to careen us into a bridge abutment.
While SuperCruise is a highway only system, riding in a friend's car with it was pretty much a non-event.
Realistically, how many chances do you think the average person is going to give the alg to drive them into an object at 70mph?
Has there been a significant customer side advancement in Tesla autopilot in 10 years? From what I can tell, all Elon has accomplished is becoming the new Prius- a sprinkling of Teslas blocking the left lane of our highways (ask any Autopilot aficionado, that’s the safest place to use it) while the human inside watches TikTok.
> Has there been a significant customer side advancement in Tesla autopilot in 10 years?
Not super interested in whether you're "impressed" or not. I think the video answers your question. The part of the video you chose to point out (poor behavior in parking lots at the end of a drive caused by bad cell signal) says a lot on its own, imo.
Autopilot of 10 years ago was lane keeping and changing on highways only. Now the car does a drive from point A to point B with some erroneous behavior in the end in the parking lot, and you're like "huh, what's the difference?"
Clearly you are not coming into this with an open mind.
The video does answer my question. It shows a Tesla vehicle that (still) requires human oversight and intervention to navigate point to point. At around 6m20s in the person behind the wheel says "OK it's going the wrong way into that one way entrance again" as he puts his hands on the wheel and takes over from the system.
Tesla engineers may have made engineering progress towards their goal. Maybe they make it N miles further before requiring intervention. Maybe they can successfully navigate an additional M% of some region without issue. That is progress! However that progress does not necessarily translate into something that customers will consider an advancement or valuable. Certainly not all customers.
There is no value to me in a semi-autonomous driving platform that requires my attention and feedback. So I would not consider the performance captured in the video to be a "significant customer side advancement". The fact that others might disagree and find value in the video is OK with me. We need only to be honest with ourselves and others when describing the system performance.
There is another company actively running an autonomous fleet every day. That is impressive to me.
> Clearly you are not coming into this with an open mind.
I think it's great that you want to cheerlead for Tesla. "FSD" and "Autopilot" just don't live up to their names yet. I'm sure eventually they'll ship something worthy of the trademark and make all your evangelism worth it.
Come experience it between my house and walmart before you go telling people they need to experience it. Good luck having good weather during your drive, here in louisiana.
Self driving probably works great on Interstates with numbers like 5, 10, and the feeders. Based on my experiences with subaru self driving and watching videos and watching tesla drivers around here, a self driving car would give up after maybe 3 minutes. There were more teslas and other expensive "self-driving" style cars a couple years ago, but now i rarely see them. I wonder why?
I can consistently get my wife's subaru self driving to swerve into another lane without warning, without jerking the wheel, with very minor control inputs. Subaru keeps updating the firmware, but they haven't fixed the suicide merge!
Are you from Houston originally, by chance, or from a part of western Louisiana? Anytime I hear the word “feeder” it’s a dead giveaway the person either grew up near Houston area or hung around there long enough to use this word in their lexicon
Nope! I'm from the state where i had to consciously not type "the 5, the 10" - California. I've lived in Louisiana for 12 years, total, now, though. I couldn't think of a better name for like the 405 or 710 than "feeders" - maybe tributary; i know there is a term of art for those freeways, though.
“Frontage” is the nomenclature I usually see in lieu of “feeder”. Though I agree that feeder is more descriptive of what the road actually does when we consider a road as being as a network or graph with flow.
Though the roads you reference would probably not be called “feeders” in the same way. I take feeder specifically in the Houstonian meaning to refer to a frontage road that is used for local access that runs parallel to a limited access highway. Notably it must run parallel to the highway and exist for the primary purpose of providing local ingress and egress while preserving the limited access nature of the highway.
Yeah 10 years ago it was a bit much. It was pretty much just lane keeping and distance keeping, but it didn't read signs or lights yet. So if you're coming up on a signal it would just keep going!
> Chrysler was the first manufacturer to implement the device in 1958. They called it “Auto-Pilot” and it appeared in their luxury model as an upgraded option. Soon after, General Motors installed it in their Cadillac vehicles, naming it “Cruise Control” which has stuck to this day.
> Small anecdote: I visited a GM dealership this week and the salesperson told me Super Cruise was not enabled for test drives. The excuse was pretty weak, like the dealership would have to pay for the service or something. GM might have the technology but they are completely bungling the strategy.
That was a your-dealership issue, not a GM issue. GM doesn't run the dealerships (though they sure have plenty of influence).
I test drove the Silverado EV a month ago and got to try out super cruise.
That's a "GM's business model" problem, which is again a GM problem.
The top-selling EV maker in the US doesn't have any dealerships, so clearly it's not an insurmountable problem. GM just hasn't prioritized surmounting it.
To be fair to GM, Tesla basically had to be a brand new auto company and work through a whole bunch of legal loopholes and battles just to have their direct sales model. GM and the rest of the legacy automakers are essentially bound to the dealership system and the dealerships represent a much larger lobbying base than the automakers themselves.
We are seeing this play out with VW’s attempt to direct-sell Scout vehicle's and immediately getting challenged/sued by dealerships over it. [1]
I could do a undergrad essay on how the dealership model is bad for the manufacturers, which would boil down to how Indian motorcycles failed (which was also due to mismanagement). However, the VW scouts are an absolute abomination that should be cleansed from the earth.
They aren't going to be any good as an off-road vehicle, the approach/departure angles are horrible. They will also be way too heavy, just as a function of being an EV. Plus, the styling is absolute garbage. I've owned a scout 80 for 18 years, it was my first vehicle. I owned an OH scout for longer than they have been collectible. This new scout doesn't have any styling cues to any prior vehicle, it just looks like if you asked an AI to generate an off-road truck. And they even had a better example, the new broncos absolutely hit it out of the park. Line I E if those up next to a mid 60s bronco and you can see the resemblance, which is why Ford did that in their marketing. These new VW scouts, if you park them next to either an 800 or a scout 2, don't even look like the same lineage. They are just an abomination, and the only car company I'd trust less with shepherding the scout brand would be stellantis2. We've seen how poor of a job they've done with the jeep.
And the stupid part is,most people don't and have never understood why the original scout was good. They were an overly heavy, underpowered, and expensive vehicle. But, they were a tremendously capable off-road vehicle, my stock scout 80 has gone places a modern Rubicon with lockers couldn't go. It's down to a very fortunate suspension design and stupidly low gearing.
Simple and rugged will beat "sophisticated" every day of the week. Underpowered is an asset because you don't break shit. Simple is an asset because you can keep it running. Steering box NLA? No problem fit one from a different vehicle. Axles? NBD just need some U-bolts, maybe some custom linkages, and any other solid axle of similar dimensions will do. Trying to make cars into some kind of disposable appliance is incredibly short sighted.
They're never going to make enthusiasts like you happy because the vehicle needs to sell to a mass market, not the kind of person who has owned a Scout 80 for 18 years.
I think you're not in a marketing mindset if you think that any significant amount of people are to buy a new Scout for off-roading. Premium-priced vehicles are more about image than the implied capabilities of that image.
Porsche sells more four door vehicles than two door sports cars. You might say that's not a real Porsche in the spirit of a Porsche, but it is what it is. It sells, and the average person doesn't need a two door car with no trunk or an off-road truck that is painful on the road. They usually need a family vehicle for mundane domestic life.
GM can’t match the lobbying power of local dealers in the state houses. Every state rep has a dealer or two in their district, and they employ staff who vote.
And, to be fair to the dealers, there is a long history of the makers trying to screw them over, and the public, too.
The dealership model is an anachronism from the days when the Big 3 US automakers were basically a cartel with a lock on the market, and predates the internet. For new sales it has no real advantages over what Tesla is doing, and for service the public would be better served by strong right-to-repair laws so people can have their vehicles serviced at any independent mechanic (or do it themselves).
Automakers started out with direct sales. Anti-trust issues quickly arose with that business model and that’s where we got dealership laws, as imperfect as they may be.
The document is rather one-sided and slanted, but I think it’s interesting to see how a dealership association explains their perspective:
Of particular interest is warranty issues. Tesla has an extra-strong interest in hiding recall-worthy issues and preventing warranty service from occurring since their service department is owned by the manufacturer.
A car dealership has the opposite incentive: every dealer-performed warranty service is a payment from the manufacturer to the dealer.
Finally, even if GM is lobbying to do direct sales, dealerships employ more Americans than the carmakers and parts manufacturers. Their lobbying group is larger and more local.
> A car dealership has the opposite incentive: every dealer-performed warranty service is a payment from the manufacturer to the dealer.
Wouldn't this be served just as well by having independent repair shops certified by the manufacturer to do recall and warranty repairs, without prohibiting direct sales of vehicles?
I would say that laws covering right-to-repair are almost entirely a different subject than laws governing retail competition.
Being forced to buy a vehicle from the manufacturer would be exactly like being forced to buy Indiana Jones and the Great Circle only from Microsoft. But it's a lot better to be able to buy it from multiple game stores, many of which compete on price.
Overall I think it's totally fine that Tesla wants to sell cars direct, and any manufacturer should be able to. But they should also be obligated to sell to dealerships at a fair price (they shouldn't be able to just pretend that their own selling process has no cost associated with it).
Dealerships also make most of their profit from the service departments which is why they don’t really like EVs. No longer do you need to go in every 3-6 months to get an oil change, belt replacement, spark plugs, etc.
1. There are still service items on a Tesla that are relatively frequent. Tires rotated every 6,400 miles, brake calipers cleaned and lubricated yearly for northern climates with salted roads, cabin air filter changed every 2 years, and two other service items every 4 years.
2. This is technically a tangent to the issue at hand. If Tesla was a gasoline vehicle startup they would have the same incentives to avoid dealership franchises, perhaps an even stronger incentive.
3. The fact that vehicle service is the most profitable part of the business is a great argument for the dealership model. Tesla owning their own service centers means that they are directly incentivized not to fix reliability problems in their vehicles, while all the car companies who have to pay out dealerships to handle recalls and warranty claims have a direct incentive to produce a reliable vehicle. Lo and behold, Tesla is about halfway down the vehicle reliability list according to consumer reports. At least they're better than Chrysler?
No, not really. Safeway's advertising and store layout isn't directly a Coca Cola problem business model problem. They care, they have some influence, but if the one seller goes under for being crappy another will pop up to take its place.
This is just a car dealership tying its hands behind its back and not competing well. If it isn't widespread it isn't a GM problem.
vehicles driving into the back of emergency vehicles, slamming on the brakes in the middle of highways - both because the CEO decreed that RADAR/LIDAR isn't necessary
Issues with its "self driving" randomly swerving at objects, pedesrtians, cyclists, other vehicles...
Enormous service/parts/bodywork backlogs
vehicles blacklisted by insurance companies becaue of high crash rates and repair costs
windows that randomly shatter
drivetrains that fail because you drive them in too heavy a rainstorm
a truck that cannot be driven in the snow (snow reflects light off the DRL bar, blinding the driver, the headlights getting blocked by snow, and the vehicle is stymied by barely a few inches of snow):
GM doesn't have any of these problems (except for a small number of Bolts which had defective cells made by LG - all replaced now, and given extended ten year warranties.)
> drivetrains that fail because you drive them in too heavy a rainstorm
Not even that, colleague's Model 3 died because it was parked out in rather normal european heavy rain over 1 night. It just completely died.
Tesla service center just took it and replaced it without any questions - clear sign this wasn't an exception and it was well known issue at that point.
People incorrectly throw around the term "rent-seeking" all the time, yet this is a case where it would be correctly applied. The dealerships petition the government to enforce the franchise model of car sales to block potential competitors. Like Tesla, but also mega-chains like CarMax.
And for good reason. The whole point of the regulation is to assure that there is someone nearby for the owner to get a vehicle repaired and be able to buy parts.
ahem Guess what Tesla has a ton of problems with?
Edit: there are a whole lot of people replying who seem unaware that I was referring to new car customers who have a vehicle which is broken under government-mandated warranties, but also that:
- The majority of manufacturers have long required proprietary diagnostic tools (the OBD-II functions in your car are only useful for the most basic diagnostic info, cannot adjust parameters, and cannot trigger diagnostic procedures such as cycling a valve or motor.) That is why some states and countries have right-to-repair laws. Automakers responded by shifting all that stuff to work only over telematics, and then claimed that online websites were not "tools."
- At one point car companies were doing bullshit like shipping parts with electronics in them "uncoded" (not uncommon, VAG does this) and not flashed with any firmware. Volvo did this. The firmware could only be flashed by a dealer, and it required them to plug the car into their internet-connected diagnostic tool through which servers in Sweden would encrypt the firmware for that specific VIN number and module, which could only be decrypted by the main control units in that specific car - and transmit it back to the technician's diagnostic tool.
- Teslas require software/online tools only provided to Tesla service centers and technicians
- You cannot buy parts for a Tesla without providing them with the VIN number of a car you own and which they still consider roadworthy. Tesla has a very arbitrary process for declaring vehicles to not be roadworthy, with relatively minor collisions triggering it, whereas vehicles nearly totally destroyed often are not declared unroadworthy and their VIN can be used to purchase parts
- Tesla does not distribute parts via third party distributors such as Worldpac, and prohibits its OEMs from selling parts, which is extremely unusual. You (and your independent mechanic) can only get parts directly from them - if they'll sell you the part at all.
The laws requiring dealerships and that they be independently owned were created because in the early days of the auto industry, people were getting endlessly screwed by automotive manufacturers producing cars that would break within weeks or months, did not have parts available for them at all, and the company would dissolve right after a production run so that customers had nobody to go after for damages.
The reason for requiring it be an independent business is simple; the state government wants to see that a local business has judged the manufacturer to have their shit together sufficiently enough to want to do business with them, and they won't be left holding the bag.
> The whole point of the regulation is to assure that there is someone nearby for the owner to get a vehicle repaired and be able to buy parts.
> ahem Guess what Tesla has a ton of problems with?
If the goal is to allow people to have vehicles repaired, then mandate right-to-repair laws, mandate that all auto manufacturers must fully document how to service the vehicle.
The problem with Tesla isn't the lack of dealerships, it's that even if you go to a normal auto repair store, Tesla won't sell repair parts nor provide detailed enough repair manuals. Tesla will void your warranty for having someone else fix it.
If the regulation is meant to allow consumers to repair their vehicles, then it should say that.
It is widely known in America that dealerships are the worst places for repairs and that c independent mechanics are generally better. I discovered this myself slowly over time.
Yea generally it's a slew of other laws that mandate the opening up of vehicles to aftermarket components that makes it viable, not necessarily dealers who just installed overpriced OEM parts.
An reasonable alternative would be for GM to have a "dealer license" mode for qualified dealers (for example, training), where they can get it enabled until the sale is complete.
It does seem strange but when one considers that the CEO of one of these competitors spent ~$260M to help elect the next President and he will have his ear on matters of autonomous car regulation, maybe not so strange.
If future autonomous car regulations are influenced by Musk and end up being lax and favour the technological approach that Tesla has taken, then GM/Cruise may have made significant irrelevant investments to solve issues that are no longer relevant, and this puts them on poor footing in the new competitive landscape.
It's possible they're bowing out early on the assumption that this will no longer be a competition that they can win.
Maybe in some small parts of the country. Tesla is dominating most everywhere else in the US. I definitely have no access to a Waymo but have been driven by FSD in heavy construction. (Do not own a Tesla myself).
Tesla is not "dominating" Waymo, they provide completely different offerings. Waymo provides a taxi service (like Uber) with autonomous vehicles. Tesla sells cars that offer a driving assistant feature, but the system is not able to operate safely without human supervision.
I can safely and legally take a Waymo home after getting drunk at the bar, this is not true of a Tesla.
It's possible that Tesla will have a massive jump in capability of their system, but currently it doesn't do any self-driving anywhere (e.g. you can't sit in the back seat).
ASS. Actually Smart Summon. You can sit in the back seat (but there's no point to, as the maximal distance to the destination and speed are appropriate only for parking lots).
I think an actual realistic goal for this is closer to 2030 or 2035. But it certainly won’t be your personal vehicle or it’ll require subscribing to an overseer system (like Waymo uses)
Btw that kinda thing is one of the few use cases that really does require 5g’s latency and bandwidth… which Tesla currently doesn’t ship. More hardware updates are going to come, despite musks promises)
My dealer experiences have included hours to swap out a dome light under warranty, 1+ hours to do a software update, and a dealer trying to claim that the manufacturers automatic engine warranty extension didn't apply to my car. Also, misread a code and tried to charge >$1k for a repair that was actually part of the same warranty repair. These were separate brands and different ownership.
On top of that, one offered free oil changes for 12 months, IIRC. They'd regularly try to sell unneeded services on top of it (alignment after <12 months of ownership without even checking it first, for example).
My experiences with Tesla have been imperfect, but vastly better by comparison. There are issues, but I'd take them over the dealer experiences 100% of the time.
Their buying/test drive experience is top tier IMO but I know multiple people who waited literally 4+ months to get undrivable cars repaired. Those experiences were mostly pre-2023 so maybe it's better now
It was bad a couple years ago but has gotten a lot better. The reason for this was that Tesla scaled up manufacturing quickly, but took longer to scale up service centers and certify third party shops. For a while there was only one certified Tesla repair shop within 100 miles of me. Now there are at least ten, four of which are Tesla service centers.
No, there is a reason they are called stealerships.
A Subaru dealership tried to charge $200 to reset a passenger window like this (literally just pushing the button to roll down the window and back up):
But they tried to just charge me for it even though I knew how to reset the window and didn’t ask them to. They were thinking I didn’t know how to do it myself, so wouldn’t know any better that they didn’t actually do anything.
I love the tesla service. Bought my car without talking to a single sales person, they come to my home to rotate my tires on my schedule. Only complaint is cosmetic service tends to have a long appointment wait time, but things like a cracked windshield were immediately serviced.
I don't know how I will switch to another car brand after this.
Similarly, there's a lot of pro-Tesla propaganda by those who own shares in Tesla. I avoid this by not considering Tesla when considering a car, as the truth is too murky.
It's not a regulatory issue. There is no federal rule requiring automakers to sell through franchised dealers. Some states have such laws. In other states the automakers have voluntarily entered into contractual agreements that prevent them from competing with their franchised dealers.
It's weird to me to see so many people thinking Tesla is "far behind", when in reality, they're far, far ahead of the competition for a general purpose and scalable self driving solution. Seems like political biases just cloud judgement so much. Youtube now has videos comparing Tesla FSD V13 with Waymo rides on the same routes.
I tried both extensively (have the full FSD package on my Model 3). FSD still feels suicidal in a good day, plain incapable on average.
I can’t imagine how someone would compare it to a multi-sensor setup - a single camera can’t match the level of prescience a Waymo or Cruise have. But also as the overall experience is just absurdly different. Waymo was the first tech thing I felt was truly magical in _decades_.
Meanwhile the best I can get out of Tesla FSD is straight highways. And even there, there’s random glitches (the forever annoying ghost braking being the one that makes me feel particularly unsafe with it)
Certain routes and models are infamous for this. I recently got a 12.5 demo from a friend that was both impressive when it worked, and truly horrendous when it didn’t. And yes, we had a phantom break event - which appeared to be due to windy road conditions. That’s inexcusable tbqh
The problem is that at the end of the day, enough people still have serious issues and Tesla wildly oversells the capability of their system.
I think they’ll “get there” in terms of functionality and reliability but it’s nowhere near what waymo is reporting
There are (crappy, dangerous-seeming) genuine full autonomous-assuming-you-don't-count-remote-safety-drivers Waymos on the street where I live, in volume, for a while now.
Tesla's equivalent system appears to be 100% vaporware. I don't understand the case for them not being far behind.
That means you live in 1 of like 6 cities in the world. Kind of hard to compare Waymo's approach and Tesla's approach and accurately judge how far ahead Waymo is. Come to my city and you will see exactly 0 Waymo's. And it isn't clear to me if it's much easier for Waymo to scale from 6 cities to tens of thousands of cities or for Tesla to scale from tens of thousands of cities to tens of thousands of cities but better.
My gut tells me Waymo is ahead, probably by a good amount. But calling Telsa FSD "vaporware" is absurd, and leads me to the same conclusion that GP already called out.
The equivalent system is no driver in car at all, and unless there's some trial I don't know about, Tesla's product that does that is announced but not existing anywhere, 0 cities, 0 cars. the definition of vaporware.
I don’t think FSD is vaporware, but I believe it has been massively overpromised and underdelivered. And also, I have reason to believe its going to require yet another major hardware revision - despite teslas protestations
Have a look at V13.2 and tell me they "underdelivered". This is the most advanced AI driving system in the world. It's essentially DriveGPT, trained on absolutely massive amounts of training data on the largest GPU super cluster in the world.
First of all, I notice that you didn't take issue with my saying another major hardware revision is due - despite Tesla's constant promises that "we totally can do it with this one bro" for the last 4 revisions.
I have been following closely. 13.2 is a major leap forward. It is VERY impressive.
It's also five years late.
It's got some serious issues, such as intervention rates (which, despite being massively lower, are still too high).
It has trouble in California-style adverse weather conditions. This is largely due to it's vision based sensors.
And by the way, I'm not saying they're not going to "pull it off". I think they're one of the few that will succeed, to be honest. I'm saying that it is nowhere close to where it needs to be to match Tesla's promises.
> This is the most advanced AI driving system in the world.
That is a bold claim, and one that I don't think is true. Tied for first or just second most advanced, I would absolutely believe. And certainly one of the five major players.
It's also missing some key customer service and realtime service components, that took Waymo over 4 years to roll out and fine tune. And Waymo is still having edge case issues!
> Just because Elon was optimistic in 2017 doesn't make this any less impressive.
That understates his repeated broken and aggressive promises, often made during investor meetings. It makes it hard to believe his claims about the future.
Well, you seem to have your head screwed on right, but here's my big disconnect: I get that you don't trust what Elon says, he's been wildly optimistic and sometimes misleading. But what does that have to do with objective reality? Would the moon landing had been less impressive if NASA had promised to land on Mars? I don't get why we have to anchor what we find impressive on what Elon promises or predicts. To me that's irrelevant.
As for the most advanced AI driving system, yes you could argue Waymo is more advanced, it's certainly more complex. By most advanced I mean that it's fully end to end, and this is a very new development in AI and robotics (which I work with). It's actually a much simpler architecture than other systems. But as we've seen repeatedly, what tends to win is simple models with a ton of compute and data.
As for V13.2, it won't be perfect, but it's also not the final version. It's approaching feature complete, and we've seen it handle snow covered roads at the very least. Rain hasn't really been a problem for recent versions. In my view, weather is mostly a matter of data collection at this point.
Maybe they should then, I was blaming humans for the unsafe/illegal right turns and deranged maneuvering in the presence of lights and sirens firetrucks.
"I visited a GM dealership this week and the salesperson told me Super Cruise was not enabled for test drives."
Seems like
a) the salesperson might not know his vehicles and perhaps it just wasn't a Super Cruise enabled vehicle (meaning, it wasn't added-on to that specific vehicle).
b) the actual Super-Cruise module was broken in that vehicle, and that was his excuse
c) he didn't know how or wasn't comfortable with telling you how to enable it. (some people are still quite afraid of giving full control over to computers)
I have tested numerous vehicles with Super Cruise; I quite enjoyed it.
I could be wrong, but unless it's a new GM rule, afaik, and with my understanding of how dealerships work (as far as the ones I've visited), dealerships try to avoid messing with the vehicle programming as much as they can. It's rather sad, as aside from the manufactures, dealerships have the most tools and access to the vehicles programming, but rarely-if-ever stray from the service manual.
Custom programming kind of is a market for that reason.
Living in a state with a very large automotive industry, there are a lot of people who have the tools + knowledge that offer custom programming services like: Programming the key fob that if one was to hold the unlock-door button for 5 seconds, the driver side window opens. And if the lock button is pressed once, then held for 5 seconds, all the windows are rolled up. Enabling a bunch of different video codecs/formats and enabling video-playback while the car is in motion if the passenger seat detects something siting there, etc, etc...
The vehicle was equipped for Super Cruise. The feature is tied to a subscription account, and GM/the dealership do not provide subscriptions for showroom cars. After I wrote my comment, the dealership contacted me to say they had set up a demo subscription with GM for the vehicle if I wanted another test drive.
This seems like an industry wide admission that this stuff doesn't actually work, tbh. I've heard similar things from Farley at Ford--advanced lane keeping cruise control systems work, robotaxis don't. Only the "tech" companies are so committed as to double down and try to wish this sci-fi into reality. The car companies are a little more grounded in reality.
But Waymo actually works? It's limited by geography and speed but they have robotaxis carrying paying customers in multiple cities. This is reality today, not sci-fi.
When will it be profitable? The carmakers are betting that L3 autonomy is a profitable investment, but robotaxis aren't. Regardless of what Waymo can or can't demonstrate technically. So either the carmakers are wrong or Waymo is. I guess we'll wait and see.
I would think that getting to a positive cashflow shouldn't be that hard, most of the cost are fixed. Recouping the billions of R&D will take many years, but those cost are sunk. The fact that they're adding cities probably means they already have positive cashflow, or are near it.
I have absolutely not shifted a goal post. What I believe is immaterial, what's important is that the companies that really know this industry are behaving as if there's no future in robotaxis. That should give you pause.
Well GM is in the automobile industry, so maybe the automobile industry?
I’m trying to understand how this would be them determining autonomous driving is not viable, which appears to be your premise, rather than them and others just not having a path to compete.
Well now you're asking a different question. Robotaxis clearly work today and are rapidly improving. Waymo is past the technical demonstration stage and has paying customers. Alphabet doesn't report profitability for that unit but it will probably take a number of years until they can drive down costs and achieve economies of scale. The senior leadership seems to be willing to fund this initiative indefinitely.
Several other carmakers beyond GM are still pursuing L4+ autonomous technology. So apparently they don't think Waymo is wrong, although they're way behind.
It's not a different question. Passion projects that aren't profitable don't last. It doesn't matter what a technology can do if there's no reason to use it.
No. The big automakers are abandoning this idea. Why? Because they don't see a path to profit. They've been at this for 120+ years. Their judgment is probably better than some ads company with an automotive side project.
> the salesperson told me Super Cruise was not enabled for test drives
I wonder if this is a liability thing (driver would be responsible).
Or it could suck and the chances of buying the car go down.
I remember walking into an apple store and seeing the apple vision pro headset. I asked to try it, and they said "you can't. You have to make an appointment and blah blah". If someone adds friction to the buying process, there must be some reason - they must be hiding something (poor performance, complexity, price, etc)
Super cruise started out as high definition maps of US highways that GM made themselves. At the time (around 2018/19) they were using mobile eye’s technology/SOCs and integrating it into their vehicles. Cruise was operating as a completely separate unit and nothing was shared.
GM later had a PR release for UltraCruise which was supposedly developed in house and maybe that uses Cruise technology. However, AFAIK it was never released and is rumored to be shelved.
Reminds me of my father in law's experience with Toyota (I think). The car's nav system had an update available on a CD you put into the car stereo. The dealership made him order one, instead of the much more reasonable loaning him one.
Cruise and GM Supercruise/ultracruise share nothing except the names and a VP somewhere inside GM. Entirely separate organizations and tech, at least the last time I had any knowledge of the matter.
Yeah and neither of them is close to FSD in capabilities lol. Super Cruise/Ultra Cruise is a brittle highway-only thing, only available in a handful of $200k+ GM cars, that works by localizing to a detailed lidar map of some interstate routes.
This was also available on the Chevy Bolt EUV as an add-on for $2,200. The Bolt EUV started at $28,795, so with the add-on it could be had for ~$31k before tax, title, etc. Not sure where you got the $200k+ number.
The parent post is talking about the Cruise vehicles made by the company this article is covering, a completely separate thing from the supercruise feature that consumers could actually buy. At one point many years ago, the Cruise vehicles cost >200k to manufacture.
Ultra Cruise was only available on the expensive Celestiq before it got cancelled and merged with Super Cruise. Apparently, the Celestiq starts at $340,000, even higher than my original estimate of $200k.
> Cruise allegedly has one of the most sophisticated autonomous driving platforms, and GM's Super Cruise (if they share any tech) is comparable to Tesla FSD.
Just edited to add it in. Wikipedia says they're not going into production until 2027 though, and unlike the others I don't think they're operating in any markets.
I am not sure I have all the facts straight but: Super Cruise and BlueCruise are features available on GM and Ford passenger cars, respectively. They aren't taxi services. I tried to make the distinction between companies participating in the "increasingly competitive robotaxi market" versus just the ones that have some level of autonomous driving capability; tweaked comment.
GM was majority stakeholder in Cruise, which was gearing up to a ride hailing service. Cruise's cars have instruments like LiDAR on the roof, unlike Super Cruise-equipped cars.
Ford wrote about an "initial launch of a commercial self-driving service in 2021" but I don't think it materialized. A photo of their robotaxi bears the logo of Argo AI, a startup that shut down in 2022.
The difference is that GM and Ford have no viable strategy for transitioning these features into taxi services, whereas Tesla does.
Not sure how it's suddenly so controversial to realize that Detroit is terrible at software strategy. This has been the conventional wisdom (and with good reason) for quite some time.
I briefly owned a Chevy Volt with the worst infotainment UX imaginable. I found a quote from a Chevy spokesperson describing it as an "Apple-like experience," presumably because the graphic designer on Fiverr found the Gradient tool in their pirated copy of Illustrator CS6.
True, but even that's just at the surface level. When you dig into the architecture / planning / business strategy that happens in Detroit, it's even worse than what the UI suggests.
"Creative destruction" is coming for them real soon.
That is a 10-year old UI based on their even older designs from their ICE vehicles. The newer UI for the Ultium is much more modern. Not as severely styled as some other companies not not as overdone as that old UI.
For reference:
"Ford BlueCruise is a hands-free driving assistance feature that allows drivers to take their hands off the wheel on certain preapproved roads"
"It seems like drivers did not find the feature compelling and were not renewing. Interest in autonomous driving appears to be cooling across the board."
No, because BlueCruise sounds like an asinine feature. To make me pay $2450 more for a car for some arbitrary selection roads for that feature to be available doesn't make sense to me.
Cruise really was on to something, well w/e GM's loss!
BlueCruise is a legitimate incremental improvement over more limited driver assistance systems. It clearly has some value to some drivers. But an extra $2450 is tough to swallow on a mainstream brand when middle-class consumers are already struggling to afford a new car. In a few years Ford will probably just bundle it in with common option packages.
> Interest in autonomous driving appears to be cooling across the board.
My pet theory is that it really took off because a bunch of investors were hoping to capture and capitalize on the baby boomers aging out of being able to easily drive themselves. They tried to speed up technological advancements in order to meet that window of a large more-affluent cohort of potential customers.
So it might be interesting to somehow plot total "self driving" investment against that kind of customer base projection over time.
Interesting theory but I think the investment issues are simpler, and non-revenue producing business models simply don't make sense outside of the window of the low interest rate years.
> were hoping to capture and capitalize on the baby boomers aging out of being able to easily drive themselves. They tried to speed up technological advancements in order to meet that window of a large more-affluent cohort of potential customers.
That's still a huge market. And the baby boomers are the most affluent demographic in the western world.
I just bought a new Mustang, and from what I can tell it's total junk.
Many others on forums report the same, and here's an example of a component being borked.
AEB, Automatic Emergency Braking, is on by default and always turns back on at vehicle start.
It's dangerous. An example? Even with a camera and a separate radar, it is borked. I've been behind another car, slowing for a light, and hit a downward slope of a large speedbump and SLAM the brakes come on.
Why?
The radar suddenly sees pavement, as the nose of the car is pointed down, but the camera sees car + brake light ahead.
This is at 30km/hr, and everyone is braking fine, the car ahead is not even close. And proof it's not me is the 5 times it happened in 2 weeks, it was always something like this.
I was almost hit 3 times.
It literally panic brakes. Horribly designed.
I finally had it and removed the radar sensor with forscan.
With behaviour like this in something as simple as AEB, by no means would I trust anytiong else. Forums show people complaining about:
* adaptive cruise slowing down, or speeding up because it sees speed limit signs on offramps, or even adjacent roads
* the lane assist pulling strongly because the car thinks an offramp is the main road.
* reports of swerving into other lanes when passing some vehicles, the car confused
It's just junk.
This beta testing on our roads needs to stop.
What cracks me up is this.
If you asked a skilled software dev, if they'd prefer to have software designed by a team, any team, responsible for tgeir life? They'd likely say no.
Because we've all seen the crappy code that gets produced. We know of bugs. We know of failures in testing, automated or not.
These cars ae all extremely buggy, just as any software project is. But beyond that, auto manufacturers are notorious for cutting corners.
I just can't imagine why anyone, this decade, would want to trust auto drive anything.
I cease to understand how the average HN commentator doesn’t see Tesla as the leader in this space, and my only guess is all the FUD media around them.
If you have a Tesla, you know that it’s far and above the best driver-assist software in the entire market, and next place isn’t even close.
Sure FSD isn’t Waymo, but it’s entirely different tech and it can drive infinitely more places than the two cities Waymo supports.
V13.2 is probably close to 1000 miles per critical intervention now. Geofence it, avoid stupidly dangerous UPLs and fix some map issues and they're already on par with Waymo. Next version has 3x the number of parameters and 3x context length. If you know anything about scaling laws you wouldn't be betting on Waymo right now.
Lidar obstacle detection algorithm from a Git repo leaked onto Tor
This is a drivable region mapping (obstacle detection) algorithm found in what appears to be a git repo leaked from an autonomous vehicle company in 2017. The repo was available through one or more Tor hidden services for several years.
The lidar code appears to be written for the Velodyne HDL-32E. It operates in a series of stages, each stage refining the output of the previous stage. This algorithm is in the second stage. It is the primary obstacle detection method, with the other methods making only small improvements.
The leaked code uses a column-major matrix of points and it explicitly handles NaNs (the no-return points). We've rewritten it to use a much more cache-efficient row-major matrix layout and a conditional that will ignore the NaN points without explicit testing.
This is an amazingly effective method of obstacle detection, considering its simplicity.
Is there any indication this object detection algo was originally from Cruise? The linked code has been rewritten and has no attribution beyond "an autonomous vehicle company".
Fun fact - Cruise did actually have a source code leak a few years back, some engineer decided to embrace the 4 hour work week and outsource their role to a contractor, who uploaded their work to a public github repo (I assume this was accidental and they intended to make a private repo). But it was not core AV code.
Waymo needs competition. Who else is even close to being able to provide it? I'd hate to think Tesla is next in line because they haven't even started a commercial pilot program yet.
I think it will be Tesla. Regulatory pressures for them just lifted, and they move faster than all the other driverless companies, courtesy of Elon's fairly hardcore mode of operation.
It may not be a perfect driving solution (probably 1:1 teleoperator to car in the beginning), but their go to market speed is undeniably better than Waymo et al. Waymo, Cruise, etc have to map out an entire region and have long rollout timelines. In theory Tesla could say they will deploy in X city, talk with the local government to get necessary approvals, and get it going after a few test drives. Thats the tail, anyway. With that kind of go to market speed you can start acquiring customers pretty easily and start scaling up the business. Its basically the typical SaaS setup - cut some corners for market share and become the defacto player in it after some time.
I wouldn't write them out purely out of spite. Their solution is the most conducive to scaling laws, on the ML, operations, and business side.
Tesla is a veeeeery distant second. Elon is full gung ho on vision only. They even removed ultrasonic sensors for parking. Tesla even removed radar. Vision only is Tesla’s bet.
Meanwhile Waymo bets on jamming the car full of sensors to ensure it has superhuman 360 perception in all conditions with multiple sensor redundancy. Waymo takes safety insanely seriously. Tesla doesn’t care if 100s die in their car while on FSD. Waymo doesn’t want to make cars or manage fleets. They want to purely focus on automating the driver end to end.
Unless Tesla has a magical perception algorithm breakthrough, Waymo may expand all US states that allow it, while Tesla won’t even get an operational robotaxi with 100+ rides per year.
Elon has tremendous success in SpaceX with falcon & starlink, that’s a fact. But Tesla FSD is so much hype compared to substance.
Karpathy said in a recent podcast that they pre-train on LIDAR data so no, Tesla is not gung ho on no LIDAR. They "just" do transfer learning with inference on vision-only, which might be a great idea in the end.
Having been in a Tesla in FSD mode recently, it felt quite a lot like Waymo's car did about a decade ago. That's a purely subjective opinion taken off the cuff.
Omar is a fanboy yes, but AIDRIVR and DirtyTesla have been openly criticizing FSD for years and show all the problems openly, now waxing lyric over the latest version. Omar, while sometimes cherrypicking videos, has reportedly had 1 safety critical intervention in over 1000 miles of driving in SF and LA.
Most states and communities aren't ok with unproven AVs on their streets causing accidents and injuries, even if they are superficially politically aligned with the AV manufacturer's CEO.
That's a good thing, we shouldn't be lowering the bar for AVs.
> to become proven the AV needs to drive somewhere for a while.
There's a straightforward solution to this that doesn't compromise public safety:
Employ paid training drivers who monitor the AV while it is collecting training data in environments where it does not yet meet the safety bar. These drivers are paid to be observant and intervene to avoid accidents. Furthermore, in their interventions, these training drivers provide critical training data used to improve the system.
Uber did exactly that and got excoriated when the paid driver was on her phone and the car killed someone on the freeway. I think many people raising are concern trolling this because they either don’t want private cars in favor of public transportation or because they are luddites towards automation of driving jobs.
They exceed the minimum viable human driver. They're some combination of a teenager and grandma. They generally get where they're going safely but anyone who's not a minimum viable driver is cringing pretty often.
Nobody wants to let a minimum viable driver chauffeur them around every day. Accepting the reduced quality of performance and increased risk a short duration cab ride is another story.
The only federal requirements are FMVSS, which have nothing to say on the subject of AVs. Some states have their own testing programs with paperwork requirements.
Nov 18 (Reuters) - Tesla (TSLA.O), opens new tab shares rose more than 5% on Monday after Bloomberg News reported that President-elect Donald Trump's transition team was planning to set up federal regulations for autonomous vehicles.
The report comes days after Trump named the automaker's CEO, Elon Musk, as a co-head of the incoming administration's government efficiency department.
Trump's team is looking for policy leaders for the transport department to develop a federal regulatory framework, the report said, citing people familiar with the matter.
Last month, Musk criticized the state-by-state approval process, required for self-driving vehicles, as "incredibly painful", weeks after unveiling a two-seat "Cybercab" robotaxi without a steering wheel and foot pedals, set to go into production in 2026.
Interesting... so Musk may be angling to get a federal law passed that would force all states to accept autonomous vehicles with less testing or data sharing than say, California currently requires.
If each car has someone piloting it because the tech can't do it, maybe the tech is the problem? I don't see how cameras can succeed. Sunlight and fog are constant and not practical to work around with visual data. I'm glad there's regulation keeping those lithium incendiaries piloted by people. People are bad enough at driving.
If you simply search for things like "tesla direct sunlight", "tesla camera condensation", "tesla camera bad weather", you can see thousands of instances of vehicle cameras not working.
The entire investigation was based on Tesla's data if you're unfamiliar. I don't care what you believe. Seems like you're one who consumes the marketing rather than data.
As a Tesla customer I’m inclined to agree. I paid in full for FSD six years ago, but never received anything even remotely usable in my region. They just took my money and kept lying. The feature is actively dangerous.
There’s no evidence that this is even going to be a viable business model, commercially or technically. There’s no “moat” around it or obvious first mover advantage. It seems very early to worry about competition.
a) Early technology adopters are more forgiving of minor problems and are your free word-of-mouth marketers. The company benefits from lowered GTM costs.
b) Fast, reliable service depends on density; there needs to be enough cars on the road that there's always one available within a few minutes when you open the app; users will develop a habit of choosing whichever provider is reliably faster and that habit will become ingrained, making market entry for later movers more costly. First movers also have the advantage of their early adopters helping pay for their fleet as it's built up; later movers have to flood new markets with subsidized capacity to get to equivalent density. That works (Uber did it for rideshare) but it's expensive. So again: lower GTM costs.
Regarding whether this will be a viable market - Lyft/Uber have already demonstrated there's tons of demand for on-demand transportation services, and Waymo in SF is already very popular. One can back-of-the-envelope whether the economics work but generally speaking any time you replace labor with automation in a large market, it does.
In the long run though you're right that there isn't a moat that can't be solved with massive capital investment. But you could say the same about many markets; turns out one can have a successful company within a competitive market the old fashioned way.. by just being really good. It's harder, but it works.
I mean Waymo already has competitors, Uber, Lyft, and regular taxis. Right now Uber and Lyft are typically strategically priced to be slightly cheaper than taxis are but I think Uber has enough margin to cut costs to compete with Waymo as they enter more markets, and Waymo needs to get people on board to make robotaxis a real thing people use and to offset their dev costs.
Waymo is also in competition with non AV ride sharing companies. They'll need to make it more cost effective and/or (continue to) deliver higher quality rides
> Who else is even close to being able to provide it?
Baidu/Apollo, maybe. There are deployed robo-taxis in China. Reviews of driving quality indicate "meh", they are sometimes directly teleoperated to get out of problems, and difficult intersections have fixed cameras the cars can see through.
From what I've read China is forging ahead very successfully, but I assume there would be a lot of friction if they tried to enter the US market, to say the least.
I’ve taken a few dozen Waymo rides, but it doesn’t matger because this is not something you can really test as a person. The entire issue is whether you have, say, 7 or 9 nines of reliability (e.g., 10x or 0.1x times more dangerous than a human), and you can’t extrapolate that from how smooth it feels or whatever. FSD doesn’t have that reliability, which is why they don’t let you sleep.
Yes, I have, it's been than anything else out there and it's not even close. There was never a moment in time that Waymo was not a distant leader despite how competitors wanted to appear.
Tesla hasn't even launched a pilot program. Cruise has actual commercial road hours logged and Waymo has a fully operational commercial service that keeps expanding. Tesla has a LOT of catching up to do. Cities won't just let them launch a service on a whim. It's going to take years to catch up.
That was 6 years ago. I find it odd for a CEO to talk so cautiously (odd but perhaps refreshing), but I'm certainly not holding anyone accountable for what they said in an interview 6 years ago when it relates to a bleeding edge industry.
Well yes of course accountability clearly isn't a thing at play here, Musk said self driving cars were coming in two years back when I was still studying, that was ~10 years ago...
We're still not even close to anything remotely capable of being autonomous, even on sunny straight californian roads... Bring these to switzerland snowy mountain passes or serbian dirt roads lmao
> a bleeding edge industry.
The only thing that this industry is bleeding is money so far
Five cities in CA/AZ/TX, aka 90 degrees angles grid, 0 inclines, sunshine 300+ days a year, no snow, no rain, perfect visibility, &c. Musk was 100% talking about self driving cars for the average joe, not robotaxis in gentrified city centers. I also have a fully autonomous vehicles, my vacuum cleaner navigates my living room flawlessly...
And again that's exactly what waymo's ceo said, it won't work everywhere / in all conditions, and as it turns out: outside of deserts and the west coasts these conditions are far from edge cases, they're the norm in most of the world
Without even talking about edge cases like mountains, put one of these in Paris or Rome and I give it 15 minutes before it gets stuck/crashes, people underestimate how hard the last 20% will be. 100 000 cars go through that place every single day: https://youtu.be/JgWhagB4d_g?t=79
I don't see how Waymo is going to survive the onslaught that is CyberCab. Cheaper to build and operate, more efficient, much faster to build, and deployable across all cities without any manual configuration. The clock is ticking on Waymo, they have 2 years to figure out how to compete otherwise they're cooked.
It's very similar to legacy space and internet, the existing players saw SpaceX coming years out, did nothing and got crushed.
Isn’t it kinda the other way around? I feel like the clock starts the first time a CyberCab gives a paid passenger ride, and that seems like quite a ways out — potentially much more than 2 years.
And Waymo has that whole time to bring down their costs of construction, learn how to solve problems of scale, and so on.
I agree that a CyberCab would have structural advantages if the service were at parity w/ Waymo, but that is a very big hypothetical!
More than that, even -- Waymo is already transitioning to having partners that do the operations & maintenance work while they focus on the tech. I read this as a confidence in the tech timeline, strategy, and maturity. Meanwhile, Tesla's strategy is "hype, ???, deal with it."
Confidence in the timeline is one read. Lack of confidence and finding partners to shoulder the financial burden at the cost of lower upside is another.
Yeah, I could be wrong and they're just looking for the lowest risk profile. But, riders seem to be pleased and accident rate has been impressively low.
Not an insurmountable problem for sure. If Level 4 self driving requires more compute in the car, someone has to physically put more compute in the car, which may require redesigning something to do retrofits, etc. A "FSD" car today may not actually have Full Level 4 Self Driving when Tesla finally rolls it out.
You pose your comment like the answer is obvious, but I don’t know that it is.
If Tesla can solve FSD with only cameras, then yes, their ability to scale is much better than Waymo’s.
But that is a stupendously large “if”!
I’ve seen the Karpathy quote before, and I think it’s kind of hilarious to compare one company which does X0,000 robotaxi rides per day, to another company which has a robotaxi maybe entering production in two years, and say, “well look, they both have problems!”
Bit of a misleading simplification, no? Tesla boosters assume they will crack FSD, it’s just a matter of time. That seems like a big leap to me.
Andrejs point is that software problems are easier to scale your way out of. I think we can both agree there's plenty of precedent for this line of thinking.
A company cannot as easily scale their way out of lowering material costs by an order of magnitude or go from 1000 cars a year to 1M cars a year.
But the implicit assumption is that Tesla’s problems can be solved by scaling laws. That seems to me like a big assumption, no?
I’ve read The Bitter Lesson, I understand the argument “the human eyes are just cameras, and we do just fine”; I’m not totally unsympathetic to the idea that Tesla is better-positioned than some people think.
But I still think it’s a big leap to suggest Tesla will definitely solve FSD by scaling
/ software, and Waymo is fundamentally screwed and has no way to innovate out of its own scale problems.
Companies lower material costs by an order of magnitude all the time! Look at battery/solar over the last 15 years — in many areas we see two OOMs of cost reduction. It’s a challenge; but one that seems on par with creating software to replicate the optical capabilities of the human brain.
Waymo (and my lived experience with FSD) is proof the end goal is obtainable.
We're surrounded by AI miracles, this website is full of examples, yet people can't conceive of Tesla reaching this finish line. It's pretty odd.
I've clocked 20k miles on FSD. In the brief period I've used it I've seen step improvements. I guess in my mind it's a foregone conclusion with no breakthroughs required. All I need to believe is on the Tesla earning reports, which is Tesla data center build progress (training) and vehicles sold (data)
I agree completely with everything you've said, but it's worth keeping in mind that if legal/compliance requirements were to be loosened, a company could expand their offerings with a less robust safety record. And someone close to the incoming presidential administration would be in a position to influence what those requirements are.
Self driving cars don't have to be on par with human drivers, they have to be WAY WAY safer else public backlash is likely to ban them.
It's not like Waymo couldn't be rolling out faster, but it's obvious they're doing everything they can to maintain the public's trust in them because once that's lost it'll be incredibly hard to get it back again.
And ultimately the law is one thing, but if people legitimately see Cybercabs as dangerous to them personally they'll take things into their own hands. Someone assassinated a CEO in New York a few days ago, you think pissed off people won't molotov a cybertaxi?
Even supposing slanted regulatory rule making, Tesla’s autonomous capabilities are vastly inferior to Waymo. The rules can’t be that complicated, either you can drive a car autonomously on a given road or not.
FWIW I’m a Tesla owner who rides Waymos whenever they are available for a cab ride, and won’t turn on FSD despite a number of free trials
>either you can drive a car autonomously on a given road or not
This doesn't seem obvious to me. My impression is that the rate of improvement on autonomous systems is at least somewhat gradual. You've seen this in Waymo's slow roll out, moving from offering the service in initially less complicated settings (suburbs of Phoenix) and progressing to more challenging terrain (San Francisco). If that's the case, then a competitor could provide a worse product but, due to the fact that accidents are somewhat rare in general, have this go undetected for a while. In the intervening time, the worse performing company would benefit from collecting additional training data that they couldn't have been able to collect under a more strict regulatory regime.
Definitely possible that Waymo would benefit even more though! They definitely seem well prepared (except for the unfortunate hiccup of working with a Chinese OEM right before a bunch of EV tariffs got put in place)
My impression is that we’re regulations to soften, Waymo would quickly expand to freeways - the lack of their ability to do so now constrains their usefulness
Tesla has data from 5 million vehicles in real world conditions.
What does Waymo have, a few hundred cars in San Francisco and Phoenix? In a field where data matters more than anything else, Tesla has an insane advantage.
Waymo (and every other AV company) mostly use simulated data. The real world stuff is very repetitive and not that important; they also don't collect nearly as much as you think they do.
And it doesn't help with the main issue, which is Tesla uses cameras and cameras don't work reliably at night, in glare, when they're dirty, etc.
> a company could expand their offerings with a less robust safety record
But would people take it?
I mean, if one service is 25% cheaper but there are a bunch of horror stories of their vehicles crashing into highway barriers (like this past October), I'm not sure people are going to get in.
Driving is 100% cheaper, and there are a bunch of horror stories of people crashing their vehicles into highway barriers. We’re not starting at zero here.
Is driving really 100% cheaper? Owning a car has pretty high fixed costs.
Also, it's different when you crash a car yourself vs a robot doing it. If you are driving, there is clearer accountability than a robot, so the bar must be much higher for the robot.
I'm not up to speed on the area of autonomous driving at all. Does anyone know if Tesla is basing its next moves off just what the cameras see at the current moment that gets fed into the model? Or is there also a "mapped" component, where it looks up previously mapped information from other Teslas that have traversed this area before? That is what I understand Waymo and GM Supercruise do at least some of the time. There are only certain pre-mapped routes that Supercruise will work on.
They have non-heuristic training. Maybe the AI model is making those connections on the hyper graph if it’s part of the training data, but it would have a decay function as roads get constructed.
What's your basis for saying this? Waymo actually serves rides and has a safety record. If you work in robotics you know that eval is everything -- what's Tesla's L4 driving record? It doesn't even have one, so how can you possibly compare?
L2 driving record vs L4 driving record are insanely different. You can't fall asleep in an L2 car without risking death. You can fall asleep in an L4 car.
Am I living in a bubble or is much of the excitement behind autonomous cabs just hype? Like, sure, it will be popular, but the way some folks(and not just Tesla zealots) talk about it, it will be 25% of global GDP.
Beyond, idk, saving my relatives the task of driving me to the airport, I don't see my family using an autonomous taxi at all. I think I'm like many Americans in that I like my personal vehicle because it's mine. I'm comfy in it, it's filled with only my stank, and I have tailored it to my liking.
if you understand why Uber, Lyft and taxis are such a big industry than you understand why driverless rides will be such a big industry. It's the same industry. But driverless requires a fraction of the labor cost, hence the potential for cheaper rides that will only grow the industry even larger. And I don't have to tip a robot.
Further proof that General Motors is run the finance types, not car guys.
Wonder if they'd sell Cruise back to the founders for say a buck? Sure Cruise could then cut a deal with another auto maker. Maybe even get financing from one?
GM is is making 12 billion a year of profit though. It seems like madness to cut Cruise unless there is no hope of it ever working or making a profit… The only U.S. competitor besides Waymo that’s seemingly remotely in the ballpark is Zoox.
Oh believe me i agree. As a former Cruise employee (i left in 2019) it’s incredible to me how they went almost all the way only to get it gutted like that
GM’s current annual expenditure on Cruise amounted to about $2 billion, and the restructuring would cut that by more than half, CFO Paul Jacobson said.
GM braces for a $5 billion hit as it fights to keep up in China’s intensifying EV price war
SAIC-GM revealed in a regulatory filing on Wednesday (via The New York Times) that it expects to write down between $2.6 billion and $2.9 billion in the fourth quarter. The automaker is also expecting another $2.7 billion in restructuring expenses.
GM has a minority ownership at 49%, so I expect that SAIC is taking an even bigger hit? Reading the article, it seems like SAIC-GM is taking the hit, not GM, so divide those losses/expenses in half and that is the effect on GM
GM believed Kyle Vogt when he was hyping up their tech by having small, meaningless deployments in a dozen cities and promising $1B revenue by 2025. Then they had that incident and GM found out exactly how far behind Waymo they are. Waymo is now giving 175k rides every week and expanding to new markets every year. No other robotaxi company in the entire country giving rides to the public.
Now they've gone back to what I think GM's original plan was for acquiring Cruise — using their tech for passenger cars. Selling cars is GM's bread and butter. This wasn't going to end any other way.
Cruise started as "fake it til you make it", with a drive down the Great Highway in San Francisco. The ideal public road - several miles of straight road with no cross streets. They managed to parlay that into an acquisition by GM and board seats on GM's board.
Then they didn't deliver.
Waymo hasn't left blood on the pavement. Cruise and Tesla have.
Here it is, the original Cruise RP-1 video.[1] There's the car, driving down the Great Highway in the Sunset district of San Francisco. Perfectly straight, no cross streets, great visibility.
Nobody ever promised full autonomous with this kit to the best of my knowledge so your claim of "fake it till you make it" is a mischaracterization. It calls it "highway autopilot" in the video, shows highway, what is the problem?
On a normalized basis in the USA? 1 fatality per ~4000-5000 years of driving. 1 injury per ~60-70 years of driving.
Can you point to scientifically rigorous evidence, at least reaching the bare minimum of fit for publication, of the absolute level of safety of those systems and where they stand in comparison?
How many people am I allowed to kill or injure before I need to run scientific studies on unproven and empirically fatal technology?
This rebuttal is fine, i guess, but to answer your final question: At least 1645 deaths if your government is british. I forget the US numbers, but globally you can kill upwards of 30,000 people with unproven and empirically fatal technology.
Really, you are arguing that I get to kill 1645 in the UK before I have to show any scientifically rigorous evidence? I get to do 1645 drives and kill the passenger every time before anybody cares?
That is not only insane, but empirically false in the USA as people cared very greatly about Cruise injuring people during the development process.
You have to be answering the question: “Assuming we already know for a fact that the technology is safer than a human driver, then, if scaled to all drives, how many can die annually and still be morally justifiable?”
My question is: “How many do I get to kill before I need to figure out if I am killing fewer people?” That is the basic question to be asked during the development process of any unproven dangerous technology to minimize sacrifices to technological dead ends.
considering humans leave oceans of blood in their wake with their own driving, all of the autonomous driving companies are doing great by your measurement of success
There's always someone saying this, and they always don't seem to realize that random humans aren't a company with billions of dollars in investment. If your billion dollar investment kills someone and you are even possibly found liable, that can kill the whole company right there
I took cruise at least a dozen times, back when they were giving free rides from downtown SF late at night, originally advertised as a promotion for restaurant and hospitality workers going home. (I was just hacking late into the night ;) I lived on the other side of town, so it was about a 6 mile trip.
I can't speak to their technology, but generally speaking Cruise seemed too cautious, both in the aggressiveness of the driving, and in how quickly the company capitulated. I've taken Waymo a few times and what stands out to me is how aggressive Waymo cars are. As a passenger this aggressiveness makes them feel more sophisticated, not to mention faster, but the cars also spook pedestrians to a greater extent as they nose their way into traffic or through turns. Nonetheless, Waymo has been more adept at moving past bad publicity than Cruise, including in the face of accidents.
Above and beyond whatever technological shortcomings there may have been[1], I think there was a management and business development mismatch between GM and Cruise. Cruise manifestly worked well enough. With better management--more honest, but also more aggressive--it could have been a contender; heck, it was (is?) the onlyproven competitor to Waymo for autonomous public taxi service, at least in the U.S.
[1] In 3 Waymo trips remote intervention was required once, which is a comparable ratio to my Cruise trips.
That would require paying more than competitors and that’s unacceptable to anyone except for the successful companies that commonly do just this.
I’m serious about this point too. Eg. Intel pays less than half of what tsmc in Taiwan pays for similar jobs on a ppp basis after 20years of wage freezes and cuts at intel. They can’t poach any talent and investors want more layoffs and wage freezes next quarter to fix the decline they see. GM is similar. It’s not going to turn around and start competing in the high end labor market reasonably. It can’t. The big shareholders would sack the ceo in seconds. Think GM could ever compete with Google (Waymo) in a market where ai engineers can earn ) 1m/yr? Ha! The MBAs wouldn’t allow it for a second.
To follow up on above I’ll give a clear example. $77k/yr hardware engineer degree required job at intel in a high cost of living area for anyone wondering how dire the situation is at these uncompetitive companies.
They unsurprisingly rate ‘very easy interview’ on glassdoor since they’d hire no one without lowering the bar significantly. I would never begrudge anyone working at these broken companies either fwiw. You just have to be aware it’s a job you have while looking for a reasonably paying job. If anyone wants a lesson in the core place the rot starts consider maybe there’s a reason intel is struggling to achieve what it’s much higher paying competitors achieve.
A very similar story plays out for GM, Boeing and the other rotting companies. Short term focus leading to non-competitiveness in the labor market leading to poorer results in all aspects. GM is not going to win this. They can’t due to a focus on production lines using minimum wage staff and applying that to knowledge work.
My Cruise offer was the highest of the 3 AV companies I got offers from. Cruise was known for great comp and worse working conditions/tech when I was searching
What Waymo is doing is very localized and really only suitable for self-driving taxis and not normal cars. Not clear if that would help GM to sell more cars in the short to medium term (and obviously their execs have no reason to care about anything else).
there's a significantly lower volume of taxis sold than personal cars sold, and if taxis got bigger in the way the self driving utopians foresaw, it would significantly cut into their personal car sales.
If taxis got bigger it would cut into personal sales, yes, but it creates a new revenue stream whereby they take a cut of every ride. The price anchor in that new revenue stream traditionally accounts for the cost of having a human driver, which a self-driving car doesn't need, so they would likely be off to a really good start in terms of profitability. Especially considering they have already paid for a significant amount of the R&D.
that's a sunk cost fallacy and that assumes that the software spending on Cruise is salvageable. They never really restarted Cruise robotaxis after suspending the testing last year.
That's true. Their technology direction is probably not efficient.
You can see something similar happening with Waymo. If it were actually scalable & in sight of a promising ROI, why wouldn't Alphabet fund it themselves rather than bringing in outside investment?
As long as passenger miles stay equivalent, switching from personal cars to taxis would not significantly reduce car sales.
Every car sold does about 100,000 to 250,000 passenger miles. Every taxi does about 200,000 to 500,000 miles. Due to deadheading, a taxi does almost 2 miles for every passenger mile.
Maybe the meant sales per capita? The US population at the end of that data was 1.5 times what it was at the beginning, but the sales haven't gone up anywhere near 1.5 times.
He is right if you look at cars, vs vehicles, but that's due to sales moving to larger vehicles that don't count as cars. But when it comes to self driving or electrification, the difference between car and truck isn't significant. Therefore, vehicles sales being up is what matters.
It does still show that Americans are buying significantly less cars per capita than they did back in e.g. mid to late 80s. So the market is only stagnant because of population growth. Add Tesla and other companies that are making EVs and it doesn't look that great for GM anymore.
That's an entirely new market for them. They are struggling even in their core area of expertise and trying to compete against Waymo, Uber, Lyft etc. would be like flying to the moon for them.
Cruise had extremely competitive base, and you're in a thread trying to partially attribute their downfall to a lack of comp: so how much that equity is worth after a downfall is a bit of a non-sequitur
Good connections: "In its announcement this week, the White House said Levandowski's full pardon was supported by the billionaire pro-Trump tech businessman Peter Thiel."[1]
There was so much noise about autonomous vehicles in SF from first responders and other officials. There was a segment that dismissed this as histrionics about self-driving cars.
As soon as Cruise closed up shop, the complaints went way down.
My fav was when a bus driver killed a pedestrian but Muni blamed Cruise for his death because it was allegedly blocking responders. They managed to completely shift media blame from bus drivers to evil robocars.
Some it could also be cold math of expected gains vs expected losses. The expected gains are growing slowly, but the major blunder of last year has made the expected losses skyrocket.
Ad hominems aside, it's easy to cheerlead Waymo since they stand alone as the nation's only commercial robotaxi service. They're expanding, their track record is exceptional, and having used them a handful of times they definitely leave an impression.
Waymo's strategy traded "do robotaxi early" in exchange for big penalties in cost-per-car (LIDAR) and cost-per-city (requirement for millimeter mapping).
Yes it gets them early service in a small number of cities, but it dramatically kneecaps their ability to scale to the entire planet.
No doubt the one-time costs are staggering, but I've read the cost per car is only around $100k. That's a lot more than the $20-$50k for a new Uber or taxi, but the potential savings in labor are massive and create the potential to severely undercut the (manual) competition.
If they don't someone else will. And then they'll risk being pushed out even of the cities that they pioneered.
They might have a plan to nail that specific niche, and there's plenty of money in that. But I am certain they have their eyes on the largest possible scale.
Completely agree, but I'm very glad they exist. It's always good to have multiple companies offering a service, and they've opened the regulatory door for Tesla to eventually launch driverless robotaxis that operate in more cities.
If by "kneecap" you mean it lets them scale appropriately as their systems and processes are able to handle the increased scope, then yes.
Because the alternative is self-driving cars inexplicably speeding into stationary objects, killing everyone inside. 5 people have died from that in just the past month.
Pretty bad news for the space as a whole. A decade and a half ago there were a dozen+ companies and tens of billions of VC dollars in the space. People were proudly proclaiming the end of driving. Now all but one of them have folded.
My thoughts today are the same as they were back then. Self driving will only be solved city by city and street by street. Local governments will need to create dedicated closed-off autonomous lanes with sensors. Auto manufacturers will need to abide by a common standard to talk to local infrastructure and to each other.
Instead we chose to solve the problem in the most complex and inefficient way possible, and are now seeing the results.
If you're going to block off lanes and install sensors why not just lay track and use streetcars? Rail is way more efficient. This has been a solved problem for the better part of two centuries.
That's like asking why do cars and roads exist at all when everyone can just use public transit.
The answer is that people want cars. They want to drive. They want to be taken directly from point A to point B without needing to walk to transit stops and switch buses/trains and face unpredictable schedules and delays.
Autonomous road infrastructure is a sensible, incremental improvement. And in fact city buses will tend to benefit from it the most.
Ok so your problem with fixed infrastructure is that it is fixed, and your solution is to create a slightly different fixed infrastructe that has many of the same problems, but instead its more expensive, less tested, less safe and less efficent?
> The answer is that people want cars.
Actually in many cities in the world, only a minor % of people own a car. And even those that do own a car don't use if for that many trips.
> They want to drive.
Driving in cities is almost universally hated and is proven to cause lots of stress.
> and face unpredictable schedules and delays.
Because in cars its impossible to have unpredictable delays, roads are always completely open and there is never any concention?
In most cities in the world, public transport has less unpredictable delays compared to driving.
> Autonomous road infrastructure is a sensible, incremental improvement. And in fact city buses will tend to benefit from it the most.
A sensible incredmental improvment are regular bus lanes with regular trolley buses.
I almost agree. I'd say people want the most convenient thing. We can see people do not want to drive by the fact people who owns cars take public transit, or take the high speed train when driving is an option.
How are you going to keep me in my non-autonomous car out of your special "autonomous only and also maybe busses" lane?
EDIT: and yes, it's a completely reasonable question to ask why invest in special new kinds of roads when more people would be better served by public transit.
Those things are easy to detect. It's much less easy to prevent a simple lane change. Some cities have "bus only" lanes, which cars use anyway, but it's easy to detect them since they don't look like busses. How are you going to detect that my car isn't autonomous?
These are pointless concerns. It is a technically trivial problem. We are literally in a thread about self driving cars. If there is a will to enforce it then it will be enforced.
That's what I'm trying to get at--why would there be any will? Why should we all decide to give up a lane to people who are rich (or stupid) enough to buy a brand new car?
As much as I love urban rail, its usefulness scales in direct proportion to population density. Europe has >2x the population density of the US. Japan has 10x the population density of the US. Rail for transit just doesn't make as much sense for the US. It works in a few of our densest cities, but nowhere else. Buses are cheaper, simpler, and easier than rail, but also are useful in proportion to density.
This has always been a bullshit claim. The US HAD good urban rail infrastructure back when we were less dense and had half as many people total.
Even if the claim (you need Europe density) wasn't trivially false, several parts of the US DO meet that level of density and could easily support good public infrastructure.
Or do you not know that in the early 1900s nearly every city had copious public transportation that was relied on by literally everyone in the city?
Japan has pretty good rail in the parts of the country that aren't dense. The way you get it is (among other things) transit-oriented development, which the US has the exact opposite of when they surround suburban train stations with giant parking lots.
Streetcar rails are terrible for bikes and pedestrians. Dedicated bus lanes are a lot more flexible. A rail car is probably more efficient than a bus, but a bus is a lot less expensive.
I've seen people wreck super hard trying to cross the streetcar rails at too shallow an angle in pdx multiple times. But that's an optional behavior. Just the other day I saw a moose crossing the road over in Twin Mountain, and I chose to slow down and not hit it. If I'd made another choice that resulted in a collision, and I'd survived, I wouldn't expect to get some settlement from the town of Twin Mountain, NH. Similarly if you ride your bike into a train track that one's pretty much on you lol.. being aware of one's surroundings is one of those basic life skills.
Installing a huge hazard in the road is optional too, though. If [insert city here] had saved money and used a bus instead of a street car, all of the problems that come from having a rail in the road would have been avoided.
For heavy rail, they don't often run in the road, parallel with the road, in the lane of travel. And when they do, it's usually in an industrial area that people aren't walking and cycling through.
Earlier than that, and IIRC what actually happened was cities decided the streetcars were greedy, put price caps on the fares, and they went out of business.
That could also mean that one company just won, right? Ten companies try to do a really hard thing, nine fail, one succeeds, the nine failures close down.
When was this peak, which cities and how many cars? Waymo has a few hundred cars, and is serving the public in Phoenix, LA and San Francisco, as far as I can tell Cruise had a kinda shaky San Francisco service, with restricted hours, now discontinued, and they have a tourist zone Dubai service and... some test areas.
So I must be missing a lot of cars and a lot of extra cities with service. Or you've confused ambitions (Cruise wanted to have thousands of vehicles and wanted to be in dozens of US cities by 2027) with reality (Neither of those things happened)
With world models like Genie 2 and other advances in LMMs (large multimodal models), the aspiration of human-level autonomous vehicles might be realized relatively soon.
Ford did the same thing a few years ago when they shut down Argo and brought it in house as Latitude to work on ADAS tech instead of actual self-driving cars. Does anyone know if Latitude has shipped anything?
They operated a robo taxi fleet for a time (until they got involved in an accident and mislead the following investigation), what other results would you like to see?
GM stock up 2.5% since the announcement, so it doesn't look like investors attributed that much of GM's future value to Cruise (or that shutting down the robotaxi development was already assumed)
QUOTE:
Cruise and GM’s technical teams will be combined into a single effort focused on developing autonomous technology to offer in future models sold by GM, according to a statement
Tuesday. GM said it will no longer fund robotaxi development work “given the considerable time and resources that would be needed to scale the business, along with an increasingly competitive robotaxi market.”
It’s a big retrench for GM and Cruise, which survived a shakeout among autonomous-driving companies and restarted operations after one of its cars dragged a pedestrian last year.
...
The move has significant implications for GM. Chief Executive Officer Mary Barra wanted to transform the automaker into a transportation technology company and double GM’s revenue by 2030 in part by generating $50 billion from Cruise. Without a robotaxi business to bring in fares, that goal looks remote.
/QUOTE
The business of robotaxis has always been a bit suspect, given the unit economics of running 1000s of taxis, each with ~$100K of equipment (even as the COGS were coming down each generation).
My 2c: the only reason Waymo is succeeding (so far) is that its parent Alphabet has deeep pockets, it has a a 6th gen (?) technology of its AV driver, and they already exited the other AV trucking segment... implying they need to succeed in robotaxis, so it's all hands on deck.
It also helps that they have, so far, NOT blundered their way (=> a significant pedestrian collision) the way Uber ATG did or Cruise did.
This is a loong game to win, far longer than anything the tech bros have promised their their investors... We are still in the initial innings
I wouldn't underestimate the advantage of not having to bootstrap a planet scale backend/ML platform. Look how much R&D Uber spends on software and they're not even doing the same level of ML. Waymo can just start using excess capacity across the google fleet whenever it wants for training and simulation. That's a huge advantage.
Which cities in which countries and climates have robotaxis (or have had robotaxis) now?
It felt like there was a lot of buzz around it a while back but then it only happened in a couple of cities which were just US cities with favorable climate. Did it ever move beyond that? You'd think that if the market is really competitive, then operators would bring it to cities where it doesn't exist, rather than just all competing in the same small handful of cities?
I actually think robotaxis will never really catch on widely in the US outside of big cities or maybe very poor people with no other option or misc use cases like if you own a car and drop it off at the dealership or if you’re out with friends and want to go home early. Most of the US is designed for you to own a car and drive it everywhere, just because cars can drive themselves now isn’t going to make self driving taxis more convenient than driving yourself without changes to how cities are already set up. Eg if you use a taxi you have to drag all your stuff with you and it’s less convenient if you live in a place where parking is effectively guaranteed. I think it will displace the existing app based taxi market and regular taxis by offering lower prices but greed will prevent companies from pricing them low enough for them to truly get people to stop driving personal vehicles. They’re really going to do well in parts of the world where taxis are already very popular because parking is limited or expensive like Singapore, Japan, China, and Europe.
I was thinking that elderly passengers could be one use-case where robotaxis catch on. Although I guess you'd still be poor in this circumstance, since otherwise you could just own a self-driving car. Or have the self-driving car from one of your adult children drive you around when they are at work. And so I just talked myself out of robotaxis being a very big market for the elderly.
About time. Cruise was an absolute joke. The number of close calls I have had with Cruise vehicles either has a pedestrian or automobile is way too high.
Cruise was garbage when I rode them. Sat in the car for ten minutes waiting for it to go. Someone over the car speakers said they were doing something to get my ride ready. I could have walked to my next destination but wanted to give them a chance.
Waymo has been a little less expensive than Uber/Lyft for me. I just saw a Zoox autonomous vehicle today. That weird looking box with no drivers seat.
Frankly, I think Tesla is going to win this. The new self-driving is remarkably good. Based on this alone, I estimate (actual) level 5 self driving in 2-3 years. I'm convinced that the lidar sensors are essentially unnecessary and the vision-only strategy is basically going to work and be much cheaper in the process.
How about we wait till they have given a single public ride before crowning them winners? At the moment Tesla's effort is nothing more than a marketing campaign.
Tesla FSD version 13 (new version) videos are starting to trickle out on YouTube and while they could be edited, it does seem to handle some crazy stuff fairly well.
If Tesla can do level 5 in 2-3 years as you say (and that might be a pretty big “if”), that places them 5+ years behind Waymo.
What leads to the win here, then? Waymo constrained by the cost of LIDAR? Is it truly such a massive % of build cost that they can’t succeed? Is it that Tesla is vertically integrated?
Apparently they have been surprised at how few photons are required to see for these sensors. They are skipping the image computer vision step and going from photons to car control in as few layers as possible.
It's not an event camera, so it's very much taking images, which are then being processed by computer vision algorithms.
Event cameras seem more viable than CMOS sensors for autonomous vehicle applications in the absence of LIDAR. CMOS dynamic range and response isn't as good as the human eye. LIDAR+CMOS is considerably better in many ways.
Next time you’re facing blinding direct sunlight, pull out your iPhone and take a picture/video. It’s a piece of cake for it. And it has to do far more post processing to make a compelling jpeg/heic for humans. Tesla can just dump the data from the sensor from short&long exposures straight into the neural net.
Humans can also decide they want to get a better look at something and move their head (or block out the sun with their hands) which cameras generally don't do.
Sad news IMO. "We're not cut out for the future. We're all in on the dying business." Dunder Mifflin level stuff.
Cruise cars were way more numerous around SF, although the service was worse than Waymo's. That stupid October 2023 accident really snowballed, and it wasn't even their (primary) fault.
Agreed -- I think things would have gone much better for them if they had been upfront and transparent about the incident, instead of the sketchy cover-up.
Yeah the robotaxi is a pipe dream grift. Humans are gross and nasty, I think some of these people need to do a ride along in a cab sometime in the winter in a busy poorly marked city on a Friday or Saturday night. Pray tell how the robot kicks passengers out or navigates the inevitable unexpected weather/human activity.
Sounds like a dystopian hell to me. Best of luck contacting the robo help support for the robo taxi trying to explain your unique human scenario that it’s never seen I guess.
You bled all over the car because you were shot and then crawled out looking for help because the AI had no clue wtf to do? Ok good luck sir, you owe us 10 days of metered fairs. No there’s no human to get help from.
Tesla Driver assistance as it exists today makings driving far easier. You can relax and once you learn where it needs help very rarely take over. It only needs to continue the rate of improvement for another 9 months and you have robot taxi.
Lots of people are ignoring this because of politics.
Right. I'm looking at demonstrated competence in operating a Robotaxi service in the real world, backed by statistical data. You are looking at a video of a Waymo taxi stuck in traffic next to a video of a Tesla that is not. I have no idea how many takes were required for either of those videos. Please respond to the substance of the milestones in the links I shared - why is 100k miles per incident "the wrong stuff" ?
It's so funny to see the mental gymnastics and just plain ignorance. We literally have videos of FSD V13 doing basically what Waymo does (and more - like highway driving), yet they all have their heads in the sand.
Its so funny to see Elmo stans fap to literally anything produced by the God Emperor rather than pay attention to the actual facts on the ground. Tesla has a track record of 10 years of broken promises, and frankly out-right lies about the capability of their product.
People getting to their destination safer, faster, insulated from weather and crackheads, not listening to showtimes, not breathing in subway tunnel air [1], is good for the individual
People consuming less energy to get to their destination [2] is good for society
I say "we" do - ie, if people want to buy autonomous cars, then that's what's going to happen.
Incidentally if there's a large number of autonomous cars available in a fluid and efficient market, that would probably mean fewer people need cars in total.
> I say "we" do - ie, if people want to buy autonomous cars, then that's what's going to happen.
Baked in to this is the assumption that the operation of markets are democratic, but they are not. A very small number of people decide where investments go, what tech to design, and what to buy. The alternative - public transit focused infrastructure, is not available to buy in small quantities. Huge investments have to be made before anyone can “vote with their wallet” on public transit. And similarly, billions have to go in to AV development before anyone can buy in to it. And those with more money always have more “vote” in a “vote with your wallet” scenario. Fifty percent of the people in the US have so little money as to make their votes nearly meaningless in such a scenario. That’s not “us” making a decision that is very much certain people making that decision.
> A very small number of people decide where investments go, what tech to design, and what to buy.
Baked into this is a certain kind of economic unrealism that you frequently see online. Rich people aren't endless fountains of capital. They fund things that work, ie, things that sell, ie, things the public likes. Rich people who have some other investment strategy than this go broke.
> Fifty percent of the people in the US have so little money as to make their votes nearly meaningless in such a scenario.
I think if you look up the median wealth and income of the US, and the purchasing power of the US middle class in aggregate, you'll be very surprised by how much sway these people have. Its true that the upper-middle class and the rich are even better customers, but many fortunes have been made catering to the whims of the median American. The US consumer rules the world.
And all this is just academic. Countless poor and middle class families buy cars every day because it's a way better option than living in an urban zone and taking a bus.
> Rich people aren't endless fountains of capital. They fund things that work, ie, things that sell, ie, things the public likes. Rich people who have some other investment strategy than this go broke.
Rich people fund a certain kind of thing, but that is a subset of the class of all things which could be built. Free universal healthcare is very popular in the US, but rich people aren't building it. They have perverse incentives to build a healthcare system which becomes a good long term investment while using government protection to exclude all competition.
There are things that can be build by collectives of regular people, like libraries, which are conceptually overlapping but also distinct concepts from "things rich people build to sell to regular people." Not all good things are products which are sold to us.
> The US consumer rules the world.
Yes to the detriment and suffering of a sizable portion of the world's population.
> And all this is just academic. Countless poor and middle class families buy cars every day because it's a way better option than living in an urban zone and taking a bus.
Yes, in a world where we vote with our dollars, a car and a bus are going to be the only options for poor and middle class families. But in Japan, China, Spain, France, and countless other places countless poor and middle class families ride bullet trains. How come we didn't vote with our dollars in to bullet trains in the USA?
It's true that there are very good things that aren't products. It's also true that public goods come at the expense of private goods. I've lived in several places with "free" healthcare, including Canada right now, and while it can be a great system, it can also come at the expense of a huge tax bill that eats into other spending and slows down the whole system.
> > The US consumer rules the world.
> Yes to the detriment and suffering of a sizable portion of the world's population.
Why is it then that real income & wealth (and with that, life outcomes) have shot up in all the "sweat-shop" countries that stock the shelves of Walmart?
> Yes, in a world where we vote with our dollars, a car and a bus are going to be the only options for poor and middle class families. But in Japan, China, Spain, France, and countless other places countless poor and middle class families ride bullet trains. How come we didn't vote with our dollars in to bullet trains in the USA?
Have you been to these kinds of places? I've lived in several, and spent significant time in many. Poor and middle class people aren't taking bullet trains to work anywhere in the world. Those things are hella expensive to build and run and ride. Even rich, dense London with its world class mass transit has a hard time delivering an acceptable experience compared to an F150 and a stretch of wide tarmac.
For most poor American families, living in a big house in a small town and owning one or more cars is the optimal solution, not some utopian pipe dream involving high-rises and buses.
> Why is it then that real income & wealth (and with that, life outcomes) have shot up in all the "sweat-shop" countries that stock the shelves of Walmart?
We have very different views on the labor conditions of foreign workers.
> For most poor American families, living in a big house in a small town and owning one or more cars is the optimal solution, not some utopian pipe dream involving high-rises and buses.
I grew up in a big house in a small town and worked doing property maintenance in high school to buy my own car. I have been driving cars for 25 years. Two years ago I moved to a dense city with halfway decent public transit (by US standards). Finally I can walk or ride my bike to my friend's houses. I still have a car and I use it about once a week. But now I bike to work, even in the rain. I use public transit more than I ever have in my life, because the transit is more useful here than in any of the car-centric cities I have lived in before. I am happier leaving my car parked and walking or biking to see my friends and loved ones. And when I bike, I wish less space was devoted to cars - roads and parking divide up a beautiful place. Everything is compartmentalized, and high speed danger is always nearby.
Even cities with the best public transit still have room for cars. But I do not think that relying on cars as the predominant method of transit is "optimal". When cities are built to make life without a car hell, then yes people in those places will find a car the optimal choice as individuals. But I have been to New York, London, Paris, and Brussels. I saw how public transit can be used to reduce the need for car ownership and car infrastructure. We could do that in the USA too.
Isn't there basically Google/Waymo and then, seemingly much further behind, Tesla Cybertaxi, Amazon/Zoox, and Uber/Yandex? Cruise allegedly has one of the most sophisticated autonomous driving platforms, and GM's Super Cruise (if they share any tech) is comparable to Tesla FSD. Strange that they would bow out.
Small anecdote: I visited a GM dealership this week and the salesperson told me Super Cruise was not enabled for test drives. The excuse was pretty weak, like the dealership would have to pay for the service or something. GM might have the technology but they are completely bungling the strategy.
Ford just lowered the cost of its BlueCruise subscription by 1/3rd. In an earnings call eight months prior they remarked they made a 70% margin on the service. It seems like drivers did not find the feature compelling and were not renewing. Interest in autonomous driving appears to be cooling across the board.