Of course it can't scale. Putting 10000 cars with 1 human in them into a city is hard and just creates traffic. Putting 100 buses with 100 people in them is much easier.
Also why don't you use something what has less degrees of freedom like a train and thus is much easier to automate and also much easier to scale by just connecting more cars. You got that Loop thing, expand onto it and turn it into a subway...
Elon is trying to reinvent public transportation without a shred of understanding why did we get into current state - buses, subways, trams. There is natural development behind it, it is not like somebody said that city people shall only sit next to a stranger - you literally can't fit these cars in the city if you want to transport people to work and from work every day. There is no physical capacity for that. Unless you will use existing mass transportation solutions.
whynotmaybe · 50m ago
> Elon is trying to reinvent public transportation without a shred of understanding why did we get into current state
No, Elon wants to cancel public transportation. Just like in the 1930's when car companies bought public transportation to shut them down.
He already delayed public transportation work with his Boring company.
The average people won't agree to spend money on public transportation when there are robotaxi available 24/7 on "already existing infrastructure".
rayiner · 2m ago
[delayed]
elromulous · 2m ago
Literally the plot of Who Framed Roger Rabbit.
Zacharias030 · 7m ago
Can you point to an article about the 1930s? This is interesting.
You clearly have a pet issue. Why do you think that it's in any way relevant to the conversation at hand though?
Do you seriously think that the main challenge Tesla is going to face when trying to scale Robotaxi up is that there isn't enough room on the roads for all the Teslas? In a world where there's currently a dozen Robotaxi Teslas per city?
oersted · 29m ago
I agree with your take on public transport, but I don't understand your point.
Tesla's robotaxis are not trying to replace public transport, they are trying to replace taxis, and eventually private cars. Uber made taxis more accessible and it scaled plenty, perhaps Tesla can do the same thing, in principle, with another degree of accessibility and cost-effectiveness. If anything, they may reduce the number of cars on the streets if more of them are shared.
I haven't yet formed a strong opinion on the viability of their vision, but that's beside the point.
Public transport is the last thing they will replace, and it is to a large degree complementary just like it has been for the last century. Public transport is of course far more efficient if you are willing to sacrifice a bit of immediacy and, well, interacting with the public. For the rest there's the private transport option, with various degrees of who drives and how much you own the car, that's where Tesla is aiming.
zhoujianfu · 2m ago
Also, I realized a side effect of a hypothetical world where everybody rode in robotaxis/waymos/even ubers is we’d effectively get congestion pricing everywhere (due to “surge” pricing), and the use of roads would actually fit into regular supply and demand market forces!
hammock · 25m ago
Why take a plane to Europe with 300 pax when you can take an oceanliner with 7000 pax , fewer degrees of freedom, more energy efficient and use ports and harbors that have existed for 1000s of years rather than building new runways and introduce noise pollution?
oersted · 23m ago
Indeed, individuals optimise for latency not throughput. They may be willing to go for high-throughput low-latency options if the cost-savings are significant, but that is often not the case.
johnfn · 36m ago
Have you used busses or the metro recently? My girlfriend refuses to ride them because they are dirty, uncomfortable, and most crucially, unsafe. And although I am an ardent supporter of public transportation, I tend to understand her side quite well when I see fare hoppers, people blasting loud music, shouting incoherently or doing drugs - and I probably see one of those every time I ride.
I use public transportation frequently. But I feel that 50 years of work on public transportation has created a system that fails a large percentage of the population. Perhaps another solution is needed.
simgt · 24m ago
> I see fare hoppers, people blasting loud music, shouting incoherently or doing drugs - and I probably see one of those every time I ride
It has very little to do with public transports and everything to do with our declining western societies. I've never witnessed any of these in any part of developed Asia I know of...
> Perhaps another solution is needed.
Reducing inequalities, funding education, local police, etc. all seem like a more sustainable way forward than shielding ourselves from the shit in a robotaxi...
seanmcdirmid · 6m ago
Just enforcing basic laws and rules on buses would go a far way. Also fare enforcement is talked about in Seattle not as a way to raise more revenue for the system (indeed, fare enforcement costs more than it brings in), but to make it safer (no one wants to ride Link and get stabbed at Capitol Hill station).
almosthere · 18m ago
Make crime crime again and get rid of the loud music and drug use?
As a society we forgot how to just stand up and say no. We relegated people doing so as Karens instead of supporting them.
oersted · 19m ago
Sure but this is not a fundamental issue with public transport, it's an issue with the implementation of public transport in the US. There are plenty of places where it's done well, and it is felt deeply.
Living in major city in the Netherlands I have never felt any need to own a car. I've been quite open to it a few times and I can afford it no problem, but it was just not worth the hassle, it's just an inferior option.
general1726 · 29m ago
That sounds like a self centered culture problem not a public transportation problem.
mafuy · 26m ago
For what it's worth, I prefer drugged up people in a bus rather than behind the wheel. It's better to be annoyed by them than to be dead from a car crash with them.
apparent · 18m ago
This seems like a false dichotomy. In general, the drugged up people on the bus probably don't have cars.
__alexs · 30m ago
None of those things particularly contribute to the traffic problem.
If the problem you are trying to solve is "interacting with random humans" then sure Robotaxi is good. If you are interested in "moving a lot of people to where they need to go efficiently" then it's not.
autobodie · 26m ago
In America or elsewhere? In America, they are dirty and unsafe because they are drastically underfunded, because of people like Elon Musk. This is the history of public transportation in America. America also has a despicable homelessness problem because of all the libertarian nonsense.
baron816 · 16m ago
I hope at some point, ride hailing companies will start offering something like a bus service. The trick will be normalizing prescheduling trips. Like, if I schedule a vehicle to pick me up at 8:15am to drive me to the office Tuesday, Wednesday, Thursday, and a return trip at 5:30, you'd be able to better optimize picking up more people on the way who are going to a similar location and to do so in an appropriately sized vehicle. That would bring down prices a whole lot, while mitigating congestion.
I'd be willing to pay a bit more vs a bus if it means I can have a reliable pickup time, a guaranteed seat, and I feel safe.
triceratops · 43m ago
> you literally can't fit these cars in the city if you want to transport people to work and from work every day. There is no physical capacity for that
There's no space for 4-seater sedans each carrying a single person. How about tiny 1-seater self-driving golf carts going no more than 30mph? They take up as much space as 2 bus seats. This is ok because you're saving the space from the bus's aisle and driver compartment, and because buses rarely run completely full so they already waste some space.
These vehicles don't exist today. But I bet you could design and build them much cheaper than subway systems in most North American cities. These cities tend to be less dense than European or Asian cities, where trams and subways are more economical.
general1726 · 39m ago
Well then you will find out that these vehicles needs their own infrastructure, because they don't exist in a vacuum, but in a city where lot of other elements are using the roads, like people crossing streets on lights.
And when you will get your own infrastructure, well you are converging back to a subway or overhead light railway.
alexisread · 18m ago
No you don't have to converge to that. Railways and trams tend to be defined by track switching. In-car switching creates a different class of system as the track is dumb, headways can be much tighter, and with a network of rails you don't have to go to a central hub, to go to another central hub, to then go to where you actually want to go.
Stations can be frequent and offline, individual transports can be smaller to transport a family / sleeper car / pallette-load of goods, land use under the track becomes available ie. new routes over farmland become viable, tracks can cross easily in 3D, and can as they are prefabbed, a line can be constructed and repaired quickly.
> you will find out that these vehicles needs their own infrastructure, because... people [cross] streets on lights.
Isn't handling that part of the definition of "self-driving"?
general1726 · 32m ago
People repeatedly crossing lights will create traffic in the city, self driving or not.
Using your own infrastructure negates what self driving tech bros are trying to achieve - pods on existing roads.
triceratops · 26m ago
Your vision for public transportation sounds like "trains, walk, bike, or gtfo".
There's 0 room for buses. If someone lives 2 miles from a train station and can't bike, too bad.
Maybe you don't mean that, but it sounds like that to me. Please explain how last-mile connectivity works in your world.
simgt · 14m ago
This edge case always comes up as counter argument to better public transports and cycling infrastructure. The few who can't combine walking or a bike with some form of public transport will simply rely on less efficient modes of transport on the now much less busy roads... be it a (small) car, a shuttle, a taxi, or whatever makes sense.
It'd be silly to suggest we don't need escalators because the elderly only use the elevator anyway.
triceratops · 9m ago
> The few who can't combine walking or a bike
Outside of dense metros, getting to the nearest bus stop or train station (note: nearest, and not necessarily the one you need for that trip) entails a minimum 10 minute walk. Most people don't live in NYC where you climb down from your apartment and board the bus/subway. Add on a 10 minute walk at the other end of the ride as well, and some waiting for the bus or train, and your trip is already 30 minutes long without going anywhere.
So then you resort to biking to save time. Even assuming 100% of the population can bike (not true because elders, small children, people with disabilities etc.) buses only have 2 bike racks. Trains are somewhat better but even they don't have space for everyone's bike.
Train-heavy solutions work best in dense urban environments. Most of North America is not like that. You need complementary modes of transportation. They aren't for "edge cases". They are a holistic solution that makes the entire system accessible. A train is useless if people can't get to it easily. If there are frequent, fast bus routes to get people around then train ridership goes up too. It's not a competition.
appreciatorBus · 26m ago
Most of the land consumed by vehicle operation is not consumed by the vehicle itself, but by the headway in front of the vehicle that’s necessary for safe operation at a given velocity.
For instance, at 0 mph, car might be 20 feet long and occupy a 10 foot wide lane, for a total land consumption of 200 ft². However, at 60 mph using the two second rule for safe following distance it needs an extra 240 feet of distance for a total land consumption of 2600 ft.².
By limiting your one seat golf carts to 30 mph, you’ll certainly need less headway then something people expect to operate at 60 or 70 mph, but it will still be much larger footprint than simply the vehicle itself.
Of course, this is true of other vehicles, including buses and trains, but they get to divide that land consumption by the number of people inside, making the comparison much more favourable.
IMO Better than arguing about what type of vehicles people should and shouldn’t use, we should simply stop giving away public land for vehicles as if it’s free. Require the public, as owners of the land, be reimbursed for land consumed by private vehicles, including headway. If you’re Elon Musk and you’re terrified of strangers on transit, that’s fine. You just have to reimburse the public for the extra land required to travel alone in your robot car.
triceratops · 20m ago
I'm not opposed to congestion pricing. I think it is inevitable as self-driving technology becomes widespread. Robodrivers remove the pain of being stuck in traffic. So you need something else to prevent forever traffic jams.
> buses and trains, but they get to divide that land consumption by the number of people inside, making the comparison much more favourable
Buses and trains rarely run completely full. A one-seater golf cart is a "bus" that runs at 100% occupancy.
tossandthrow · 33m ago
Regardless, you'd need to get Americans out of their huge SUVs.
Whether it is into golf carts or trains, I bet you there will be resistence.
Grimblewald · 40m ago
"this is ok because..."
you might want to think a bit longer on the inefficiencies here.
triceratops · 38m ago
Do tell...
I see half-empty city buses running all the time and no one says boo about "inefficiencies" then. And even then city buses outside of dense metros run no more frequently than every 15 minutes, but only during rush hour, and only on popular routes. That's no way to get everyone loving public transport.
I want to take public transport. I don't want to spend hours waiting for the bus. I don't want to wait 20 years for a subway to be built either. What do I do?
The only answer is more frequent, smaller buses. But drivers are expensive. So then self-driving buses. And if the buses drive themselves, you really want to them to be completely full, all the time. What kind of "bus" is always completely full? A single-seater golf cart.
__alexs · 28m ago
Even a half empty bus is considerably more efficient than most forms of personal transportation.
triceratops · 20m ago
Is a half-empty bus more efficient than a 4-seater sedan carrying 3 people? Or an electric scooter? An electric trike? A golf cart?
WalterBright · 33m ago
> Elon is trying to reinvent public transportation without a shred of understanding
Given Elon's enviable track record of proving the doubters wrong, I'd be hesitant to make such a claim.
Disposal8433 · 40s ago
The lies about FSD, the webcams instead of Lidars, the fake robot controlled by a human being, the obsession with shitcoins, firing useful people, or being a literal Nazi in front of the cameras?
I'm wondering what track record you could talk about.
darkhorse222 · 30m ago
That's an appeal to authority. Either make an argument for why his reasoning is good or don't. But don't appeal to someone's reputation.
That being said, I do think what the parent commenter is missing is that those other modes require significant investment. With cars you get profit and can scale one by one. So less up front investment. Should do nicely in all sorts of cities.
WalterBright · 23m ago
An appeal to authority is a logical fallacy, unless the authority is an expert in a relevant field.
Elon is an expert at overturning conventional wisdom of what can be done in high risk enterprises. And transportation is one of his enterprises.
leptons · 11m ago
Please list some of the things he's "proved the doubters wrong" about.
maxlin · 28m ago
Your argument defeats itself. Robotaxi users won't be some sudden additional people in to the existing system, they'll instead replace that exact crowd who doesn't want to use public transit but live private rides with more freedom. Elon obviously knows more of this than your arrogant ass does.
And for the cities being filled? What cities are filled with are cars that are used 5-20% of the time. Tireless robotaxis instead, will multiply that number and thereby require a fraction of cars for the same ability, enabling greener, less contested cities, with _higher_ traffic throughput!
renewiltord · 25m ago
It’s actually not that much easier. Putting cars on the road with digital dispatch means that the cost to route is near zero. But getting efficient allocation for buses is very hard.
You need high frequency for urban transit to work. A single line at once per fifteen minutes is 48 bus visits at each stop in a 12 hr day.
Just try it. Put 100 buses in Austin in some sim. You won’t get 100 people in them. SF is 49 sq mi with 800k people. Second densest city in America. 1200 buses.
It’s a doomed proposition with 100 buses.
Mountain_Skies · 53m ago
Rarely have I been on a bus with 100 riders outside of a few cities like Vancouver. Even there, it's only during limited periods of time per day. Most buses in most cities run with only a small fraction of the bus's capacity. Transit systems can try to bump up utilization by making buses less frequent but then that lowers overall ridership. Simply creating more capacity doesn't itself create greater ridership. I'd argue for 500 buses with 20 people if you really want people to change their behaviors as reliability and frequency is far more important for attracting riders with options than larger buses that are less frequent.
toxik · 38m ago
How about smaller buses then? This is quite common in the Nordic region it seems.
CooCooCaCha · 33m ago
Elon is trying to reinvent public transportation in a way that he profits from.
WalterBright · 31m ago
Who is going to subsidize money-losing enterprises?
amanaplanacanal · 10m ago
The tax payers currently subsidize pretty much all of our transportation infrastructure. The question is what is the best way to spend that money.
babypuncher · 18m ago
Billionaires obsession with transporting people in tiny little private pods and calling it the future continually perplexes me.
I've seen billionaire-backed projects that are literally just 6 passenger pods that use existing rail infrastructure. Sounded neat, until you realize that just a handful of 6 person pods would clog up a small rail system just as much as a handful of small trains capable of carrying 60 passengers each.
It's like they're completely detatched from the needs of everyday people. All they can think of when they see public transit is "gross, I have to sit next to poor people!" and they try to solve only that problem and nothing else.
coffeecoders · 52m ago
I hear this argument often. Subways require billions in infrastructure and over 20 years of lead time. And they operate on fixed routes.
Additionally, robotaxis and Waymo-style systems are trying to replace the private car and on-demand ride-hailing, not mass transit like subways or buses.
Our country operates on 4-year political cycles, not 20+ years infrastructure visions. Hence long-term megaprojects are mostly pipe dreams.
general1726 · 46m ago
> Subways require billions in infrastructure and over 20 years of lead time
Of course they does, because purpose of subways is to transport hundred of millions to billions of people a year. You just can't fit that on a road with a unscalable thing like a robotaxi.
breadwinner · 2h ago
Having taken Waymo rides multiple times in San Francisco I can attest to how awesome Waymo is. I am worried Tesla will bring a bad name to the whole robotaxi industry. Waymo has never had an at-fault injury accident. Tesla FSD has killed many people.
tzs · 2h ago
In the US it is officially 2 deaths involving Teslas with FSD engaged, with a couple more under investigation but not yet verified that FSD was engaged.
Still way more (no pun intended!) than Waymo, which has had 1 Waymo involved in a 6 car crash that killed someone in one of the other cars. Besides the human fatality a dog was also killed, and 5 other people were injured, some seriously. The Waymo was empty at the time.
Ironically this crash was due to a Tesla.
The Waymo and the other cars were all waiting at a red light when the Tesla rear ended them at 98 mph.
The driver of the Tesla was not impaired at the time of the crash. He says he tried to stop but the brakes were not responding.
The driver was from Hawaii, and it was later discovered that there is someone in Hawaii with the same full name, Jia Lin Zheng, with a record of around 20 traffic crimes over the last 20 years, including excessive speeding and running red lights.
I don't know if it had been determined if the Jia Lin Zheng visiting from Hawaii who caused the San Francisco crash is the same Jia Lin Zheng as the Hawaiian Jia Lin Zheng who has the long record of unsafe driving.
I'm not familiar with the naming conventions of whatever country/culture that name comes from. Is Jia Lin Zheng the kind of name that probably many people have in Hawaii or is it one that is likely rare?
aqme28 · 1h ago
> In the US it is officially 2 deaths involving Teslas with FSD engaged, with a couple more under investigation but not yet verified that FSD was engaged.
That is not a useful metric for Tesla. They disengage FSD when they detect a potential accident.
darknavi · 1h ago
Well that, and most of the incidents I've seen haven't been using FSD but instead traditional Autopilot which hasn't received updates in years.
Kerbonut · 1h ago
It’s still counted as FSD enabled. Would you prefer FSD to remain active and potentially cause further collateral damage after an accident causing who knows what kind of damage to the vehicle? Safer to shut it down when systems are working and brace for impact. Seriously use your brain.
n8henrie · 53m ago
Can you elaborate on how things are counted?
Is it "counted" if FSD was engaged within a certain time frame prior to a crash? If so, do you know what time frame?
Or only if it was disabled automatically due to detecting a potential crash?
The latter would still be problematic, as a human driver noticing a problem just prior to the FSD disabling itself would potentially be missed (right?).
Do you know who does the counting and who makes the rules in this regard?
Asking as you seem to have more knowledge here than me.
maxlin · 22m ago
Your argument is ridiculous. Might as well attribute all accidents with any Tesla close in vicinity to FSD then?
The data is collected in all of these incidents, and most people have seen the clips of FSD avoiding otherwise potentially lethal accidents, so "They disengage FSD when they detect a potential accident" is also just patently untrue.
No comments yet
prng2021 · 1h ago
Tesla’s disengage their self driving just before a crash. That’s how the company can say FSD hardly ever causes crashes.
Kerbonut · 1h ago
Wrong. They disable it because it’s a safety thing. They still count it towards their metrics that FSD was enabled.
And the links don't even touch on things that are comparable lol. Waymo might keep all the data themselves as they own the cars, while with Teslas, the drivers can and will just grab the camera data themselves, many post it YouTube.
Jcampuzano2 · 26m ago
Okay, point us to the data Tesla publicly releases telling us they do or don't count it.
Why would you trust a word they say when Elon has lied out of his teeth at every single investor meeting for the last decade.
I say this as an owner of a Tesla myself.
computerex · 53m ago
Source ?
kindkang2024 · 1h ago
As for 'Jia Lin Zheng'—it's a typical Chinese name. I can confirm that, as I am Chinese myself.
renewiltord · 22m ago
Zheng Jia Lin is a common name. And there are many Chinese/Taiwanese in Hawaii. Source: wife’s family is Taiwanese.
I remember this incident. It happened a couple of blocks away. Unreasonable that they let him go.
sMarsIntruder · 1h ago
How many people saved by the tech?
sorcerer-mar · 1h ago
Tesla doesn't release the data required to assess this.
And from that information alone, you can get the gist of what that data says!
(inb4 you post the accidents per mile chart which is very obviously useless and designed to mislead midwits, as it is not controlled for age of automobile or driving conditions)
jeffbee · 1h ago
Probably none. Fatalities in Teslas grossly exceed those in comparably-priced vehicles, which is the only benchmark that matters. Unless your counterfactual is that Tesla would be even more dangerous without FSD, but I don't think that is a useful counterfactual.
tsimionescu · 2h ago
The good news so far is that Tesla doesn't have a robotaxi service at all, they have a plain taxi service. We'll see what happens if they ever release a self-driving car, but for now, as in the past 7-8 years, they are way behind Waymo in the self driving car arena.
hansvm · 2h ago
Waymo came close with me. There were two left turn lanes, and it migrated from the inner one to the outer one in the middle of the turn without a blinker while I was next to it. It got lucky that I'm young and wasn't too tired and also that it was a relatively safe place to run me off the road.
Marsymars · 1h ago
This is tough for people to get right too. Near my house we’ve got 2 left turn lanes that merge into a 3 lane road without any lane markers, and I’ve got to maintain constant vigilance for people drifting into my lane.
maxlin · 25m ago
With that logic might as well do away with all defensive armies because they've "killed so many people". Firefighters too.
FYI. FSD is safer than human drivers on large datasets. Accidents cause deaths of thousands every year. Arguing against FSD for "safety" has The Grim Reaper cackling.
renewiltord · 24m ago
It’s that Waymo is both more widespread and better. If there were no Waymo, then sure. But there is Waymo.
Joel_Mckay · 2h ago
Camera based guidance systems are unreliable, and the number of edge-case failure modes grow exponential in time.
LIDAR/LADAR based systems are not perfect, but do offer mm precision for guidance systems. SLAM based LIDAR systems can be very good, but are also not perfect when forced to guess where a platform is located.
Cheers, =3
maxlin · 15m ago
Human bicameral driven cars are unreliable too, but FSD is more reliable. And more reliable, with way more data than any competing technologies.
Using vision for driving is something that has worked for as long as cars have existed. Trying to push some "millimeter precision" solution with unproven feature set and prohibitive hardware accessibility is just asking for no real safety improvements and just more lives lost.
Cheers.
Bluestein · 2h ago
> edge-case failure modes grow exponential in time.
How so? Honestly asking.-
rightbyte · 1h ago
Failures in automotive programs are surely linear to time over a year with all seasons.
Joel_Mckay · 1h ago
Failure rates are a different area, and most engineers have proven it is not linear. =3
you should try your luck selling exotic snake oils.
Joel_Mckay · 1m ago
They were conflating computational feasibility with sensor device failures.
Ask questions if you don't know, as being rude is not constructive. =3
Joel_Mckay · 1h ago
That is a long explanation, but generally even human binocular disparity incorrectly guesses 3D structures and distances. Our brains automatically fill in a lot of missing information that computers just can't know a priori (example: you know a dog hidden behind a car doesn't actually vanish nor remain stationary.)
Most guidance platforms would use LIDAR/SLAM to describe the local road surface, and overlap camera vision data to extrapolate distant surfaces and objects. Note distant objects also have lower resolution, unknown non-distinctive features (speed bump, or open man-hole cover etc.), and increasing sparse data as velocity effectively lowers world-state sampling rates.
The world-state is constantly changing at every intersection, sampling constraints add latency, and the navigation way-point goals may reach contradiction with immediate path-planning due to ambiguous/expired information.
Cheers =3
peterfirefly · 1h ago
So, it doesn't "grow exponentially", you just wanted to express that there are "many".
> you know a dog hidden behind a car doesn't actually vanish nor remain stationary.
Computers can do that, too. It's not that different ... guessing missing words.
I think it was two years ago that Tesla told the public that they could do that.
Bluestein · 1h ago
I wonder if - at some point - AI might not do the "filling in" of information the brain does by generation (ie. filling in the gaps) or something.-
Joel_Mckay · 1h ago
Maybe, but first a real "AI" will need to be invented.
I suspect a Amish horse buggy is more practical. =3
Joel_Mckay · 1h ago
>Tesla told the public that they could do that.
Lol, really? They either developed superman x-ray vision, or just tracked object occlusion with a common re-acquisition mitigation (so worthless when physical inertia carries a vehicle into an object collision.)
>So, it doesn't "grow exponentially",
The further the object... the more possible choices will need to be made in the guidance system. Note, guidance and navigation are related, but different problem domains. Roughly, the possible choices (and errors) if I recall grew by:
...but that doesn't even cover the projected future risk(t). =3
samrus · 2h ago
the thing about waymo is that i suspect they're running the same ML fraud that tesla itself is running in the silicon valley in general, which is to overfit on the 20% of situations that occur 80% of times.
for waymo itself, you can overfit on 100% of the situations that will be encountered. 49 square miles isnt that large. its the real world outside that which im concerned about its efficacy in. i think if you put a waymo in a small town that no alphabet engineer has ever even heard of, then youll see it fail badly as well.
FSD is a reinforcement learning problem, and we have no good way of training non-simulation algos for that. and a real dynamical driving environment cant be simulated accurately enough
seanhunter · 2h ago
Waymo works in a remarkable range of situations. I took a waymo in LA and our route came through an awkward four-way intersection at the crest of a hill on a residential street. Another driver went through the stop when we had right of way, saw us and then just stopped, completely blocking our side of the road. The waymo just backed up a couple of yards and then slowly went round the wrong side and proceeded on its route. That is, in a weird situation it did exactly what a good, cautious human driver would do. Small sample, but it makes me think they are not doing what you say, they are just actually trying to approach the problem seriously rather than Tesla’s “full speed and damn the torpedoes” approach.
breadwinner · 2h ago
> if you put a waymo in a small town that no alphabet engineer has ever even heard of, then youll see it fail badly as well.
Which is why it is a non-goal for Waymo. It should be a non-goal for Tesla too, given the state of the art.
sorcerer-mar · 1h ago
It's hilarious to see Tesla fans try to act like designing for an undefined operational domain is somehow extra brilliant and not one of the stupidest fucking ideas anyone has ever come up with.
chronic0262 · 2h ago
> i think if you put a waymo in a small town that no alphabet engineer has ever even heard of, then youll see it fail badly as well.
Waymo is not claiming to work in small towns.
Tesla is. Soon™.
amscanne · 2h ago
San Francisco, Phoenix and LA represent a strong diversity of driving conditions. Certainly not all driving conditions, but no one is throwing a Waymo into a small town in the way you describe. Expanding slowly and cautious seems like the rational thing to do, I’m not clear what you are proposing as an alternative (or specifically what the alleged fraud is).
Marsymars · 1h ago
> San Francisco, Phoenix and LA represent a strong diversity of driving conditions.
This could very well be true, but if you’re looking at it from a perspective of someone who lives in a rural area with real winters, for driving purposes, those all look like pretty much equivalent large American cities without a winter.
tln · 58m ago
Waymo has done winter testing in Buffalo, Tahoe, Michigan FWIW.
findthewords · 2h ago
As a control engineer who knows something about sensor fusion - No LIDAR no ride. Musk can brute-force his unsafe robotaxis on the road but they won't be as safe as Waymo. Maybe people won't care if the price is slightly cheaper vs. competition, I don't know.
dazc · 2h ago
The non-technical answer is that people do not care so long as it's safer and less unpleasant than the current alternative.
My last taxi ride involved jumping red lights and speeding through residential streets at 60mph because, I assume, it was early morning and the driver had learned from experience that he could get away with driving like this.
The previous experience to this was a lecture about a certain religious ideology and how I should spend the next two weeks reading up about it.
gdbsjjdn · 2h ago
Remember that the robotaxi industry is currently doing everything they can to look good. They are being watched like hawks. Look at how everyone was horny for Uber rides, until Uber needed to actually make money so they jacked up fares, cut driver pay and let anyone with a piece of shit car join. Or how Google search has been enshittified.
In 10 years when robotaxi companies are short on cash and trying to IPO they will absolutely start speeding. They'll lock the doors and give you a paid presentation about Scientology during your ride. "Accidentally" drive you to a competing store that paid for sponsored traffic, instead of the .
madaxe_again · 2h ago
Over the years, I’ve had four human driven taxis crash, and two taxi drivers rob me.
scrollop · 2h ago
Never heard of those two occurrences happening for a taxi customer, ever.
Perhaps you live in a location that lends itself to such events...
amscanne · 1h ago
Before there was an alternative used to take taxis in Toronto occasionally, and the common refrain was that the card machine was broken. And sometimes no change was available. These kinds of soft scams were common.
So it’s not a hold-up, but definitely a form of robbery.
brianwawok · 2h ago
Chicago taxis were pretty bad. Had taxis driven by the guy not in the card. Broken seat belts. Mega speeding. It was terrible.
igor47 · 43m ago
I was in a Chicago taxi where the driver was basically passing out but (I think) chewing qat to keep awake
dazc · 1h ago
Well, you can't say that again.
madaxe_again · 1h ago
Crashes: New York, Istanbul, SW UK, Cairo.
Robbed: Bishkek, Astrakhan. Former nicked my SLR at a gas stop, I should have been more attentive, but wasn’t expecting him to loot my luggage. Latter delivered me to his buddies who threatened violence unless I turfed over every penny I had. Joke was on them as they thought I was a loaded oil exec when I was actually just a broke backpacker.
I’ve also had the old shake-down fare in Ljubljana, Bucharest, and Riga, off the top of my head - but I don’t count that as robbery, just assholes.
apparent · 15m ago
I'm curious to know how you would act as a driver, cyclist, or pedestrian if you saw a Robotaxi nearby. Would you be more cautious, and if so in what way? Or are you mostly worried about LIDAR-less vehicles running into white walls, which it can't identify as solid walls?
adastra22 · 2h ago
Removing LIDAR was the eye-opening moment for me re: Musk. The justification given at the time was technically bogus while it was so obviously in response to supply issues during COVID. If Tesla really wanted safe FSD they would prioritize sensor quality and multimodal input. They went for the bottom line instead.
bryanlarsen · 2h ago
Tesla's never had LIDAR to remove. They had short distance ultrasonic parking sensors that they removed.
tzs · 1h ago
They also had and then removed radar.
_giorgio_ · 32m ago
Not true, they had lidar in at least one model. It's written in his bio too.
bryanlarsen · 21m ago
They had lidar on some models they were using for internal testing, but they never sold them.
kamranjon · 2h ago
The part that I always found difficult to square was not that Elon Musk towed this “no need for LiDAR” line so hard, but that Andrej Karpathy, who I generally consider a very reliable voice in this space, was also in strong agreement that cameras were all that was needed. Does anyone know if he still believes cameras is all you need?
> but that Andrej Karpathy, who I generally consider a very reliable voice in this space, was also in strong agreement that cameras were all that was needed.
You really expect a Tesla employee to speak out against Elon?
Especially when $10M+ TC is on the line?
sgt101 · 1h ago
Andrej "train it on more data and the problem will go away" Karpathy
Grimblewald · 8m ago
Right? Im not arguing against the skills he obviously has, but if we're always just piece wise approximating the underlying manifold, then there will always be new problems. The amount of data required to reliably approxate reality, in the absence of an inductive bias, is infeasible to expect to collect. Not to mention how computationally inefficient it becomes as your model blows out in size/complexity.
Joel_Mckay · 2h ago
Many LIDARs are blinded by direct Sunlight, and often see highly reflective surfaces like mirrors as black hole or distant ranged areas. These still need tertiary safety sensors like mm RADAR for safety in dust/rain/sunlight.
In cities, high-speed rail and e-bikes make more sense than a honking traffic jam at 4am. lol =3
tlogan · 1h ago
I have a question:
- What are the main challenges in building software that relies solely on camera input?
- Which specific modules or tasks still require LiDAR to function reliably?
findthewords · 48m ago
Camera vision and LIDAR perfectly complement each other. Camera vision is no good detecting unknown/outlier obstacles quickly and accurately. LIDAR is great at detecting unknown obstacles quickly and accurately.
You can tune the camera obstacle detection to be hyper-sensitive, which results in phantom braking, causing Passengers to feel that the car is "unreliable" while it actually is safer.
Humans are better at braking the appropriate amount when they see something strange, dynamically tuning their sensitivity in a new situation.
You can lax the sensitivity, which will reduce false alarms, but will actually cause more crashes, deaths, and injuries. You don't want your customers to feel unsafe, so from a business perspective you will inevitably reduce the sensitivity.
samrus · 2h ago
can binocular cameras replace lidar you think? they should result in just as reliable distance estimation
djaychela · 2h ago
No, they don't. Look at what has happened when a tesla has mistaken a motorcycle with two small rear lights that is nearby for a car that is further away but with the same lighting configuration. Did not end well for the motorcyclists.
He's just wrong about this.
Grimblewald · 4m ago
I dont think it's wrong, but i do think models avaliable right now lack the inductive bias required to solve the task appropriately, and have architectural misalignments with the task at hand that mean for a properly reliable output you'll need impossibly large models and impossibly large/varied datasets. Same goes for transformers for language modelling. Extremely adaptable model, but ultimately not aligned with the task of understanding and learning language, so we need enormous piles of data and huge models to get decent output.
maxlin · 7m ago
You know quite little. No one has the kind of training data generation as Tesla does, and cases like what you describe were already discussed in 2017.
LIDAR will forever stay a niche technology. If it was so notably "better", Tesla would have just scaled that. But they went all in on vision because it has worked for 100+ years.
findthewords · 36m ago
They are complementary sensors. It's a much easier engineering feat to combine two (cheap) sensors that are good at different things and fusing this information than creating one perfect sensor that does everything.
Private moon landers (the Japanese being most recent one) keep crashing because they rely on a single high-quality altimeter and expect it to work perfectly, all the time. If they had a complementary low-quality backup altimeter that operated independently, they would have had a less failure prone distance estimation system.
typewithrhythm · 2h ago
Can cameras do it eventually is a bit of a tangent.
All the information is there in a video feed, but the amount of work to get reliable perception from it is not small. With LIDAR and radar you get to the end goal with less uncertainty.
rainsford · 1h ago
The real key is that things like LIDAR are designed to work well with the types of tasks computers are good at, like taking a bunch of precise measurements every second and performing complex calculations, while a binocular vision based understanding of the world is something humans are good at because we evolved that ability over millions of years.
You can probably eventually ("never" is a long time after all) get a computer to understand the world as well as a human purely through camera based sensors, but it's a much more difficult task than taking an approach that uses tools computers are already good at. Similarly, I suspect it would be an uphill battle to have a human drive using raw LIDAR input.
2OEH8eoCRo0 · 53m ago
I think you underestimate how many guesses, approximations, and filling in your brain does to what you think you're seeing.
findthewords · 46m ago
Absolutely. The goal of self-driving is to be better than human drivers. Even the best drivers struggle with the sun shining from low angles, or road reflections, or snow, and so on.
loloquwowndueo · 2h ago
Ever heard of optical illusions? If a brain can be fooled by input from its two cameras like this, what hope does a dumb (or worse, artificially “intelligent”) computer have?
doctorhandshake · 1h ago
I think optical illusions are a poor choice to illustrate this point. They are manifestations of the corner cases, peculiarities, and side effects of our visual processing system and neither cameras nor Lidar are without their own analogous issues.
loloquwowndueo · 53m ago
If you’re saying cameras have analogous issues I fail to see how the analogy is a poor choice - looks to me like you understood exactly the point I was making.
DennisP · 1h ago
Would Musk's argument about sensor fusion have made any sense back when they were doing lots of hand-coded C++?
I've been thinking maybe vision-only was a reasonable decision, back when lidar was expensive and the software was hand-coded. Now it doesn't, because lidar is cheaper and the software is and end-to-end neural net, and additional sensors are just more inputs to the network which will learn to use them. But Tesla is locked in because of the promises they made to early FSD buyers.
AlotOfReading · 1h ago
Sensor fusion is pretty straightforward. You can think of it like sorting algorithms in CS. There's a bunch of standard techniques simple enough to teach undergrads that work fine in production, and enough technical depth beyond them to last the rest of your career.
If you actually look back at the E2E tweet, musk only says that the NN replaced 300k lines of "control code". Control code usually doesn't encompass the entire AV software stack, but neither should it take 300k LOC. As far as I'm aware no one is 100% sure what they mean by E2E and if it's actually the standard meaning or something else that's been widely misinterpreted.
maxlin · 1m ago
Their engineers, who obviously are on the bleeding edge, going out of their way to avoid sensor fusion issues says something quite different. I could believe "Straightforward" could maybe apply for something many tiers below in complexity and safety requirements to what they're doing. But adding non-agreeing, non-uniform information sources to the most capable real-world ML vision system not driven by human-engineered code?
I don't care even if you said you had 70% of the experience their team has, what you say can't sound reasonable or caring for actually improving safety in numbers.
general1726 · 1h ago
> and the software is and end-to-end neural net,
Well considering Musk record for adherence to reality, question is if you can really believe that or if Musk thinks that this is happening or it is not happening at all and Musk is just making it up.
DennisP · 1h ago
Given what's happening in the rest of the AI field, it seems pretty likely that it's correct.
Plus various Tesla engineers have said the same thing, Tesla does have a very large AI training cluster, and FSD quality made a big jump when they claimed to deploy end-to-end.
Jyaif · 1h ago
You've got it exactly backwards.
When the software was hand-coded, having Lidar and a high-res map was vital.
But if you have a good enough AI, sensors that replicate the human senses are all what is needed.
The real question is: when will we get good enough AI that can be applied to all cars?
DennisP · 36m ago
The question isn't whether it can be done eventually, but whether it would get safer faster with extra sensors.
Back in the day, Elon's specific objection to lidar was that it was too hard to code the sensor fusion part. With end-to-end there's no coding to deal with.
And it's not like they've replicated the visual cortex. It's the same neural net technology everybody else is using. It can deal with any sort of sensors just fine.
CyberDildonics · 17m ago
> But if you have a good enough AI, sensors that replicate the human senses are all what is needed.
What is your example or evidence for this?
Jyaif · 6m ago
Concentrated and well-rested drivers with good eyesight not under the influence are able to drive very well.
dham · 1h ago
It's not "Musk's argument," it's Andrej Karpathy's.
Also, if you've ever done any ML you would note that more data isn't always better. Plus there's the piece about which thing to believe when you get conflicting data. It's a lot more to it than what random Hacker News people are saying in this thread.
DennisP · 25m ago
Conflicting data is an issue even when you just have lidar, and it's not hard to deal with.
The cofounder of Waymo taught one of the first Udacity courses on this subject. He went through a small Python project that processed lidar point clouds for self-driving. The data is noisy, you get conflicting information from different points, and the code aggregates all that into the most likely 3D model of the world.
Additional sensor inputs are just more of the same, and neural nets are pretty good at this sort of thing. They'd even learn which sensors are more reliable in different scenarios.
As for "more data isn't always better," I've mostly seen that applied to training, not inference in real-time control systems. Even for training, it turned out people had been fooled by a local maximum, and once past that, more data really was better.
tiahura · 2h ago
What version of Lidar does your head have?
general1726 · 1h ago
Counterquestions
- What living creature is using wheels to move around?
- What kind of birds come strapped with a jet engine?
Sometimes non-natural solutions are easier and often better than attempt to replicate nature for every cost. Imagine your logic applied on a plane - birds flap their wings, thus this 737 should spread it wings and flap away like a goose. Now take military goose and flap fast enough to get supersonic...
vidarh · 25m ago
> - What living creature is using wheels to move around?
A human driving a car.
I agree with you that sometimes non-natural situations are easier and can absolutely be better, but the point of bringing up humans is generally to show that it demonstrably is possible to do at least as well as humans, with humans as an existence proof.
But that it might well take too long and cost too much to get there, and that it might well in the end be cheaper and better to use additional types of sensors is a good point.
rainsford · 2h ago
There's no reason to think the best performing, or even adequately performing, technological solution to a problem would mirror how humans have solved it. Submarines don't swim like fish after all.
But more specifically to this case, human eyes are attached to brains with (generally) vastly better image recognition and reasoning abilities than any camera based self-driving car. Because of this, humans are better able to recognize visual input even in degraded or unusual conditions compared with a computer.
staplung · 1h ago
That's a bit like saying cars should use legs instead of wheels.
(Also, I don't know what star-system you grew up on by my lidar sensors are next to the tubular sheaths on my cephalothorax, right where Xoc'tlz'ik (the Creator) intended them to go.)
DennisP · 2h ago
My head contains intelligence hardware that far outstrips anything in these cars. Plus my cameras are a lot better.
Despite all this, a lot of modern cars are adding lidar to make things easier for me.
seanhunter · 2h ago
Your head has a general intelligence, so it can do sensor fusion in a way a car can only dream of.
4b11b4 · 1h ago
It's not that we lack lidar, but we what we have in addition to "cameras".
We possess a spatial intelligence (e.g. how your brain has an approximation of: It feels like I walked three blocks) that will never exist in this "photons-in" "controls-out" fantasy.
rainsford · 1h ago
There's a solid argument to be made that solving the self driving problem using only cameras might end up being roughly equivalent to solving the AGI problem because you will have essentially created a computer with a human understanding of the world around it using human senses.
4b11b4 · 1h ago
I have thought that Tesla will be forced to learn many things from the only camera approach.
And they may well end up with something "that works". But... many buts.
maxlin · 10m ago
*who wasn't able to get a job at Tesla and knows only of non-scaling, unproven technologies
Waymos which cost as much as an actual taxi and aren't scalable are nothing but a VC money carnival.
ArtTimeInvestor · 4m ago
The whole article is based on an incident a YouTuber talks about in one of their videos. The incident was not captured on video and was not even described as dangerous by the YouTuber himself.
I would say if that is all you have to discredit the Tesla robotaxi project, then the project seems to go pretty well.
codechicago277 · 2h ago
“On Monday, Musk thought it would be funny to expand the area covered by its three-week-old Austin robotaxi service to resemble a giant penis when seen on a map.
“Harder, better, faster, stronger,” the $1 trillion company wrote on Monday, a double entendre referencing the synth pop track of the same name by Daft Punk, a duo appropriately known for performing as robots. Musk approvingly reposted the phallus-shaped service map, adding the fare would now be hiked to $6.90 per ride from $4.20 previously, both numbers the 54-year old often employs for comical effect.”
It’s amazing that anyone still takes him seriously. One of the first riders almost hit a train, and this is what he’s spending his time on. It’s inexcusable, but his fans lap it up.
buran77 · 2h ago
A decade ago his antics were seen as whimsical and perhaps funny, especially to his fan base. With that fan base ever shrinking his antics today are disgusting and trying to appeal to the crowd which still appreciates something this low-brow.
whendiduassume · 2h ago
It's like a different dumb jackass assumed that role some time ago.
Why would they use what was a science role model that way?
dylan604 · 2h ago
Why the boards of directors of any of his companies do not ban him from making social media posts from his personal account is beyond me.
barbazoo · 2h ago
I doubt that the personality that leads to this behavior would accept being told not to do something by anyone.
ceejayoz · 1h ago
They’re his buddies and are set for life. Why rock the boat?
AtlasBarfed · 1h ago
Nazi salutes won't be forgotten.
On national TV.
In a presidential inauguration.
A permanent stain on US history.
xyst · 1h ago
All billionaires are permanent stains on US history. Musk is just the poster child for this behavior.
lpa22 · 1h ago
In your world view, what would you personally do if you start one, or multiple, incredibly successful companies and amass $ billion(s) of equity in the process?
ChicagoDave · 1h ago
Without subsidies from the Obama administration, Tesla never succeeds.
Without government contracts, SpaceX never succeeds.
Without a mysterious visa, Musk never stays legally in the U.S.
wat10000 · 1h ago
Shut the hell up and stay the hell out of politics.
madaxe_again · 1h ago
The Bellamy salute is as American as it gets. You’re meant to do it while you pledge allegiance.
If you actually read that link you posted you would've found that it was discontinued in 1942.
So no, you're not because its meaning became muddy at best
wahnfrieden · 19m ago
Does the Bellamy salute include a repeated motion from chest outward?
privatelypublic · 1h ago
Need more info on this, because its really easy for somebody to stick a camera at an angle that if somebody waves at the crowd it becomes iffy.
Was interesting to see actual video of a newish public figure wave at the crowd, wince and look in the direction of the camera. Clearly newly trained not to wave at the crowd.
fullshark · 1h ago
Weird they didn't include a photo, tweet including commentary confirming the juvenile justification:
Probably because major news outlets are reluctant to include pictures of penises, even stylized ones, in their articles.
vigilantpuma · 1h ago
Musk is a clearly brilliant man who does not know how to keep his mouth shut or not act like a teenager. I know he doesn't need my respect... But he's certainly lost all of it by now.
And for people who are like "yeah but he gets results," are you really saying he wouldn't be getting more results if he didn't spend the last 10 years being an idiot online?
jeffbee · 1h ago
Citation for the "clearly brilliant" part?
noqc · 1h ago
> Musk is a clearly brilliant man
proof left as an exercise for the reader.
samrus · 2h ago
hes always been like this. from faking video games to pushing memecoins.
his being cringe itself isnt the issue. its the lack of EQ that leads to someone being a slave to wanting to be liked. it doesnt inspire confidence in his ability to think rationally. like how hes perfectly fine grifting and contributing to the decrease of trust in society. its messed up
soared · 2h ago
He was definitely not always like that, at least loudly in public.
Zambyte · 1h ago
He made a space rocket company called Space Sex and your choices of Tesla cars are S3XY. Maybe he wasn't like that in the PayPal days, but ever since he has been loud in public, he has been like this loud in public.
wat10000 · 1h ago
The difference is that it used to be combined with some pretty cool stuff. He was a loudmouth jackass but he was (running companies that were) doing neat things with electric cars and rockets. Now the cars and rockets are stagnant, his newer ventures are jokes, he blew up Twitter, and is making a decent attempt at wrecking the country I call home. The loudmouth jackassery was kind of amusing when there was cool stuff to go along with it. Not so much anymore.
JKCalhoun · 2h ago
He lost me when he called (EDIT: REMOVED NAME) a respected cave diver a "pedo". (I guess 7 years ago now.)
j-bos · 2h ago
The man refrenced might prefer his name be continually indexed with said term.
immibis · 2h ago
If I were Vernon Unsworth I'd be fine with searches for "Vernon Unsworth pedo" coming up with discussions about how stupid Elon Musk is for calling me a pedo out of spite and how he really should be in jail for that absurd defamatory statement, but isn't because he's politically untouchable.
hengheng · 37m ago
That was the moment the veil was lifted.
There's a model Y with The Sticker on it in my street, introduced years after he showed his colors like that. No, dear neighbor, when you bought that car, you knew.
rightbyte · 2h ago
I think you can estimate the later bound of his meltdown to when he called the rescue diver a "pedo"?
Bluestein · 2h ago
Business practices are being enshittified to the point of 4chan. Profit, for the lulz.-
Not so long ago you'd gain traction through insight or depth. Now it's absurdity that wins the price, vanity upon vanity.-
At some point the idiogarchs are going to celebrate (celebrate) their accidents and dead as a win. Lol.-
fullshark · 1h ago
They are "the men in the arena" taking chances pushing humanity forward, we are on the sidelines, doing nothing but criticizing people greater than us. Death is necessary for humanity to progress. Eleven workers died building the golden gate bridge, etc.
wat10000 · 1h ago
Funny how people who live lives of unfathomable luxury are the ones “taking chances,” while the people who literally die in the process are just the price of progress.
Bluestein · 1h ago
I feared it sounded familiar. There it is.-
tiahura · 2h ago
“One of the first riders almost hit a train“
How many human drivers almost hit something?
JKCalhoun · 2h ago
I wonder: when a human does it "we" have someone to blame. I think as a species we like that.
When a machine kills someone do we go after … the company? (Maybe we should all be shorting robo-taxi companies.)
steveBK123 · 2h ago
In the 1000s of miles I have been a passenger in Ubers, taxis, etc.. I have never had one try to drive through a railroad crossing.
We need stats, not anecdotes. Last I saw, Tesla’s “full self-driving” required human intervention something like once every couple hundred miles. That’s absolutely atrocious. It’s probably better now, but still a long way from where it needs to be. For comparison, the average American driver goes around 100,000 miles between collisions.
steveBK123 · 1h ago
Right and if you remove elderly, drunks and teens, who are generally not your taxi driver of choice, the driving stats skew even better.
So average driver is a low, probably incorrect bar to meet.
tzs · 39m ago
(in progress...check back in a few minutes)
shcheklein · 6m ago
To be fair. I live at Mission Bay (SF) that has Caltrain railway nearby (and you have to cross it if you take particular ways in/out). I drive (and like it a lot!) Waymo. Waymo avoids crossing it (it takes a longer way to drive a bridge). So, they probably realized the risk and still to this are not willing taking it.
DustinBrett · 57m ago
These are like articles saying why Tesla won't make an electric car. Just more doubters to be proved wrong, again.
maxlin · 17s ago
Pretty much. Has the air of the oil&gas funded "journalist" piece where he ran circles in a parking lot then wrote about it suddenly running out of battery, and was disproven by logs lol
avidiax · 38m ago
Having worked at Uber, I can tell you that most people vastly overestimate how many Uber drivers are working at any one time. A medium sized city like Seattle might have a few hundred at peak hours.
There is a limited scale to any taxi service, at least until it becomes cheap enough that the company can afford to have lots of cars sitting parked and idle for much of the day. Otherwise, the rush hour peaks will be less convenient and affordable than having a private car.
itsoktocry · 47m ago
Or like the solar roof, battery, swap, Cybertruck best selling vehicle of all time...you know, things people said wouldn't work, and didn't.
Elon has a good track record, but he has some duds in there too.
chung8123 · 46m ago
Didn't Tesla start as an electric car maker?
jqpabc123 · 1h ago
“The fact that they’re hiding data should tell you everything you need to know. If you really want trust, you have to have full transparency.”
Anything Musk says is simply not trustworthy.
Not just my opinion but proven over and over again.
Hiding safety data is just icing on a cake that he has been baking for years. The real problem here is that those who choose to trust him and his companies are putting others at risk.
Those who own/drive a Tesla typically have insurance but what about those who hire robotaxis? Can they be held personally liable when their roboride hurts someone?
Closi · 2h ago
Tesla Robotaxi is a classic case of premature optimisation - with Waymo they decided to deploy whatever technology worked best and did not cost-optimise at the start. Each Waymo vehicle is much more expensive than a Tesla, and has much more expensive sensors and compute.
Tesla on the other hand started with a load of constraints - smaller amounts of compute, cheaper sensors, needs to hit a price point as it will be installed in every vehicle sold - can't be $50,000 of dedicated self driving hardware per vehicle.
There are some scenarios where constraints lead to breakthroughs, but often for true-moonshot projects it is the opposite (see: Space shuttle, System/360, the manhattan project, LHC etc)
redox99 · 10m ago
Tesla approach is so much better financially (leaving aside morality and false promises).
Waymo has sunk billions of dollars with almost 0 return yet.
Tesla has developed FSD basically for free, because people pay for software that isn't there yet, and Tesla doesn't waste money on expensive hardware.
dchftcs · 48m ago
FSD itself is not a failed bet as a driver-assist system. When it started Tesla was cash poor and often on the verge of collapse, so cutting costs made sense. They will never do as well as a mature Lidar-based solution, but over the years FSD added some value to some people, and might turn out to be a slight competitive advantage in selling cars.
I also think Tesla and their robotaxis are egregiously overhyped, but cutting Lidar was not a terrible business decision at the time. Though it's a terrible decision to have still stuck to vision-only (and using low-quality cameras apparently) when they could have at least got Radar and cheap Lidars.
ec109685 · 31m ago
If the key unlock for self-driving is maximizing real-world training data, then Tesla’s approach of using cameras make sense since it allows them to utilize their entire fleet to deploy incrementally better FSD to.
maxlin · 34m ago
This article will age so badly lol.
Tesla's architecture is the only one of the robotaxi providers which actually _can_ scale. Not being dependent on exact HD maps, having the same standard as humans do "if you can see enough to drive" means eventually there'll be scant roads Robotaxi can't serve. By this strength it'll serve in areas that would never work for LIDAR/pre-HD-mapping requiring solutions
renewiltord · 18m ago
Saying Elon Musk won’t succeed at something that is a hard technological problem is foolish, yes. His companies have done what everyone said was impossible for ages.
But I do wonder if rising nations will instead just build their infrastructure to line up with technology (e.g. HD mapping their streets as they build them, installing signed beacons to inform AVs). The cost of this tech is cratering. One possible future, though admittedly an unlikely one.
ChrisMarshallNY · 2h ago
I'm actually kind of stunned at the report by the guy that almost got turned into chum by a train.
The stan is strong, in that one...
jfengel · 3h ago
I recall that Teslas have a mode where it shows its camera and labels what it sees. I would love to see what these failures look like in that mode. What did it think was there? Nothing at all? Something far away? Insect on the lens?
Coeur · 2h ago
What that mode shows is that the Tesla thinks there is nothing there.
AFAIK what Tesla's software does is object detection (cars, people, bikes, road sides) and only uses that. Which of course is not at all safe since there can always be things on the road that are an obstacle but not recognized.
What it should do is model the entire world around it in 3d and consider that. But it doesn't. Cow on the road - problably not recognized. A child in an uncommon halloween costume chilling / lying in the middle of the road - pretty damn sure not recognized and the Tesla would just kill that child. Yep.
(I drove a Tesla Model 3 with the latest "self-driving" software for a bit a year back.)
rogerrogerr · 2h ago
> AFAIK what Tesla's software does is object detection (cars, people, bikes, road sides) and only uses that. Which of course is not at all safe since there can always be things on the road that are an obstacle but not recognized.
This is not true; in modern FSD the visualization is disconnected from the actual driving model. The visualization runs a lightweight model that labels some stuff and shows it; the actual heavy lifting is now a single model that takes pixels as input and outputs control commands. My car very clearly reacts to stuff that doesn’t appear in the visualization.
dylan604 · 1h ago
The sad thing is that based on your telling of it, if it detects a small child in a weird costume in the middle of the road but doesn't recognize it as a small child that it's not at least recognizing that something is in the road that should not be regardless of what it is. That should be the minimum in recognizing the road and then recognizing any object within the road.
Mountain_Skies · 36m ago
A while back electroboom did a test with his Tesla's self driving and in most cases had to slam on the brakes himself when it didn't recognize obstacles he had set up for it. Also have seen videos of Tesla's navigation screen having a difficult time with freight trains. It stopped for the trains but was very confused about what it was, showing it as various oddly shaped cars and trucks that often morphed into other oddly stretched vehicles.
EmilyJM · 3h ago
Idk if its just me but it seems the link to the article is not working, I believe this is the original https://fortune.com/2025/07/20/elon-musk-tesla-robotaxi-serv... . Honestly really excited in general for the expansion of driverless vehicle services like robotaxi/waymo, but I am a bit cautious on the robotaxi camera only approach (correct me if I’m wrong) compared to Waymo’s combination of LiDAR, radar and camera
tsimionescu · 2h ago
Robotaxi is not a driverless taxi service, though, it's a regular taxi service with a human driver in every car. It's laughable to even compare it to Waymo.
dham · 1h ago
Sounds like all the same things Waymo has when they launch in a new area?
i dont like musk but the driver it there during this initial phase to take over in emergencies right? they arent actually driving the whole time?
i do agree he should say its in early access or something
jmyeet · 1h ago
Tesla has to create new revenue sources. The Supercharger moat is evaporating. Elon is otherwise pretty successful in torching the brand loyalty. Regardless of one's personal opinion, Tesla owners as a whole are at odds with Elon politically. Tesla owners skewed educated, living in blue states and are environmentally conscious. I think this will go down as one of the biggest self-owns in history.
The only thing preventing Tesla from going bankrupt is trade barriers on Chinese EVs.
Elon has always been a snake oil salesman and I suspect this robotaxi venture will blow up in his face in a way that's potentially fatal for Tesla. All it will take is someone to be killed in or by a robotaxi. That'll bring in the authorities and likely reveal what a shitshow it is behind the scenes.
fullshark · 1h ago
It won't be fatal for Tesla, I'm not sure it will even be fatal for Musk as CEO. As long as the stock price goes up or even sideways Musk will be CEO of Tesla and they will sell cars and Musk will play whatever financial games / hype to keep the party going.
jmyeet · 51m ago
TSLA has defied gravity for years. Fundamentally, it makes no sense. It's been valued at between $500,000 and $1 million per car sold in a year. That's utterly insane.
But we know that stock prices are based on expectations for the future not fundamentals. So massive growth is built into the stock price but we have to start asking where this growth is coming from. That's why there's a robotaxi effort.
EVs are becoming a commodity. So over time Tesla's price direction is towards the Big Three. But with more risks, namely the Chinese supply chain for batteries.
You can view Tesla as an investment in Elon, which I think the market has. But that's on incredibly shaky ground. He's torched his relationship to the current administration going so far as to call the sitting president of the United States a pedophile [1]. Elon doesn't seem to appreciate that Trump could end Tesla with an executive order.
My point is that Elon is volatile and can so easily become a liability to Tesla. They're paying him billions so firing him can immediately and massively improve the bottom line.
As they say, the market can remain irrational longer than you can remain solvent. So I'm not predicting the imminent collapse of Tesla but I think it will come and it'll be sudden and dire.
Because there's always some new thing that will bring infinite wealth, and shareholders believe Tesla will be the leader in it.
Now it's robots, $10 trillion market.
xyst · 1h ago
I have the misfortune of living in this city and godforsaken state where they let any company with enough money use it as a testing ground.
Cruise was/is fucking awful. But these tesla taxis somehow 10x worse than Cruise. Going to end up live the governor of this state and without the ability to recover funds due to awful tort reform.
65 · 2h ago
Tesla self driving has always been doomed to fail without lidar.
lvl155 · 2h ago
OT, but we need smart roads before we get a reliable robotaxi. You need lidar + road sensors embedded into the road. You basically have multiple sensors covering a section of the road that constantly updates specific to that section. And you have robotaxis pinging these stations wirelessly. When disconnected, autonomous feature also disengages. You don’t need to install a ton of sensors on every cars.
trainsarebetter · 1h ago
Nonono. Externalizing the part of the compute is bad idea. increasing the cost and complexity of a socialized infrastructure is bad as now it relies on external reliability, we can barely maintain roads in there current form
lvl155 · 36m ago
Based on that view, maybe we shouldn’t have public roads at all. Go back to scavenging for food.
duxup · 1h ago
I recently ran across this weird world of the influencers who have been trotted out (invited) to post social media posts about their rides in the robotaxi. It is amazing how many are “other than the thing(s) that went wrong it was very safe and worked great”.
Some of those things that went wrong were dropping people off in the middle of an intersection, not recognizing railroad track guard rails… and simply not being able to get them to their destination at all.
ryandamm · 37m ago
Genuine questions for AI engineers, or self driving car people: is the Tesla approach of only using cameras inherently flawed? I've read that the AI is directly hooked up to the cameras, with no explicit intermediate 3D representation... everything is done in latent space. If true, this seems inherently hard to improve; throw more data at it, sure, but you can't necessarily understand how and why it fails when it does. That seems... non-optimal for safety-critical systems like self-driving cars.
Also why don't you use something what has less degrees of freedom like a train and thus is much easier to automate and also much easier to scale by just connecting more cars. You got that Loop thing, expand onto it and turn it into a subway...
Elon is trying to reinvent public transportation without a shred of understanding why did we get into current state - buses, subways, trams. There is natural development behind it, it is not like somebody said that city people shall only sit next to a stranger - you literally can't fit these cars in the city if you want to transport people to work and from work every day. There is no physical capacity for that. Unless you will use existing mass transportation solutions.
No, Elon wants to cancel public transportation. Just like in the 1930's when car companies bought public transportation to shut them down.
He already delayed public transportation work with his Boring company.
The average people won't agree to spend money on public transportation when there are robotaxi available 24/7 on "already existing infrastructure".
Do you seriously think that the main challenge Tesla is going to face when trying to scale Robotaxi up is that there isn't enough room on the roads for all the Teslas? In a world where there's currently a dozen Robotaxi Teslas per city?
Tesla's robotaxis are not trying to replace public transport, they are trying to replace taxis, and eventually private cars. Uber made taxis more accessible and it scaled plenty, perhaps Tesla can do the same thing, in principle, with another degree of accessibility and cost-effectiveness. If anything, they may reduce the number of cars on the streets if more of them are shared.
I haven't yet formed a strong opinion on the viability of their vision, but that's beside the point.
Public transport is the last thing they will replace, and it is to a large degree complementary just like it has been for the last century. Public transport is of course far more efficient if you are willing to sacrifice a bit of immediacy and, well, interacting with the public. For the rest there's the private transport option, with various degrees of who drives and how much you own the car, that's where Tesla is aiming.
I use public transportation frequently. But I feel that 50 years of work on public transportation has created a system that fails a large percentage of the population. Perhaps another solution is needed.
It has very little to do with public transports and everything to do with our declining western societies. I've never witnessed any of these in any part of developed Asia I know of...
> Perhaps another solution is needed.
Reducing inequalities, funding education, local police, etc. all seem like a more sustainable way forward than shielding ourselves from the shit in a robotaxi...
As a society we forgot how to just stand up and say no. We relegated people doing so as Karens instead of supporting them.
Living in major city in the Netherlands I have never felt any need to own a car. I've been quite open to it a few times and I can afford it no problem, but it was just not worth the hassle, it's just an inferior option.
If the problem you are trying to solve is "interacting with random humans" then sure Robotaxi is good. If you are interested in "moving a lot of people to where they need to go efficiently" then it's not.
I'd be willing to pay a bit more vs a bus if it means I can have a reliable pickup time, a guaranteed seat, and I feel safe.
There's no space for 4-seater sedans each carrying a single person. How about tiny 1-seater self-driving golf carts going no more than 30mph? They take up as much space as 2 bus seats. This is ok because you're saving the space from the bus's aisle and driver compartment, and because buses rarely run completely full so they already waste some space.
These vehicles don't exist today. But I bet you could design and build them much cheaper than subway systems in most North American cities. These cities tend to be less dense than European or Asian cities, where trams and subways are more economical.
And when you will get your own infrastructure, well you are converging back to a subway or overhead light railway.
Stations can be frequent and offline, individual transports can be smaller to transport a family / sleeper car / pallette-load of goods, land use under the track becomes available ie. new routes over farmland become viable, tracks can cross easily in 3D, and can as they are prefabbed, a line can be constructed and repaired quickly.
Although it's (IMHO) not ideal, https://openprtspecs.blogspot.com/2011/11/climbing-chain.htm... can give you an idea of the possibilities.
Isn't handling that part of the definition of "self-driving"?
Using your own infrastructure negates what self driving tech bros are trying to achieve - pods on existing roads.
There's 0 room for buses. If someone lives 2 miles from a train station and can't bike, too bad.
Maybe you don't mean that, but it sounds like that to me. Please explain how last-mile connectivity works in your world.
It'd be silly to suggest we don't need escalators because the elderly only use the elevator anyway.
Outside of dense metros, getting to the nearest bus stop or train station (note: nearest, and not necessarily the one you need for that trip) entails a minimum 10 minute walk. Most people don't live in NYC where you climb down from your apartment and board the bus/subway. Add on a 10 minute walk at the other end of the ride as well, and some waiting for the bus or train, and your trip is already 30 minutes long without going anywhere.
So then you resort to biking to save time. Even assuming 100% of the population can bike (not true because elders, small children, people with disabilities etc.) buses only have 2 bike racks. Trains are somewhat better but even they don't have space for everyone's bike.
Train-heavy solutions work best in dense urban environments. Most of North America is not like that. You need complementary modes of transportation. They aren't for "edge cases". They are a holistic solution that makes the entire system accessible. A train is useless if people can't get to it easily. If there are frequent, fast bus routes to get people around then train ridership goes up too. It's not a competition.
For instance, at 0 mph, car might be 20 feet long and occupy a 10 foot wide lane, for a total land consumption of 200 ft². However, at 60 mph using the two second rule for safe following distance it needs an extra 240 feet of distance for a total land consumption of 2600 ft.².
By limiting your one seat golf carts to 30 mph, you’ll certainly need less headway then something people expect to operate at 60 or 70 mph, but it will still be much larger footprint than simply the vehicle itself.
Of course, this is true of other vehicles, including buses and trains, but they get to divide that land consumption by the number of people inside, making the comparison much more favourable.
IMO Better than arguing about what type of vehicles people should and shouldn’t use, we should simply stop giving away public land for vehicles as if it’s free. Require the public, as owners of the land, be reimbursed for land consumed by private vehicles, including headway. If you’re Elon Musk and you’re terrified of strangers on transit, that’s fine. You just have to reimburse the public for the extra land required to travel alone in your robot car.
> buses and trains, but they get to divide that land consumption by the number of people inside, making the comparison much more favourable
Buses and trains rarely run completely full. A one-seater golf cart is a "bus" that runs at 100% occupancy.
Whether it is into golf carts or trains, I bet you there will be resistence.
you might want to think a bit longer on the inefficiencies here.
I see half-empty city buses running all the time and no one says boo about "inefficiencies" then. And even then city buses outside of dense metros run no more frequently than every 15 minutes, but only during rush hour, and only on popular routes. That's no way to get everyone loving public transport.
I want to take public transport. I don't want to spend hours waiting for the bus. I don't want to wait 20 years for a subway to be built either. What do I do?
The only answer is more frequent, smaller buses. But drivers are expensive. So then self-driving buses. And if the buses drive themselves, you really want to them to be completely full, all the time. What kind of "bus" is always completely full? A single-seater golf cart.
Given Elon's enviable track record of proving the doubters wrong, I'd be hesitant to make such a claim.
I'm wondering what track record you could talk about.
That being said, I do think what the parent commenter is missing is that those other modes require significant investment. With cars you get profit and can scale one by one. So less up front investment. Should do nicely in all sorts of cities.
Elon is an expert at overturning conventional wisdom of what can be done in high risk enterprises. And transportation is one of his enterprises.
And for the cities being filled? What cities are filled with are cars that are used 5-20% of the time. Tireless robotaxis instead, will multiply that number and thereby require a fraction of cars for the same ability, enabling greener, less contested cities, with _higher_ traffic throughput!
You need high frequency for urban transit to work. A single line at once per fifteen minutes is 48 bus visits at each stop in a 12 hr day.
Just try it. Put 100 buses in Austin in some sim. You won’t get 100 people in them. SF is 49 sq mi with 800k people. Second densest city in America. 1200 buses.
It’s a doomed proposition with 100 buses.
I've seen billionaire-backed projects that are literally just 6 passenger pods that use existing rail infrastructure. Sounded neat, until you realize that just a handful of 6 person pods would clog up a small rail system just as much as a handful of small trains capable of carrying 60 passengers each.
It's like they're completely detatched from the needs of everyday people. All they can think of when they see public transit is "gross, I have to sit next to poor people!" and they try to solve only that problem and nothing else.
Additionally, robotaxis and Waymo-style systems are trying to replace the private car and on-demand ride-hailing, not mass transit like subways or buses.
Our country operates on 4-year political cycles, not 20+ years infrastructure visions. Hence long-term megaprojects are mostly pipe dreams.
Of course they does, because purpose of subways is to transport hundred of millions to billions of people a year. You just can't fit that on a road with a unscalable thing like a robotaxi.
Still way more (no pun intended!) than Waymo, which has had 1 Waymo involved in a 6 car crash that killed someone in one of the other cars. Besides the human fatality a dog was also killed, and 5 other people were injured, some seriously. The Waymo was empty at the time.
Ironically this crash was due to a Tesla.
The Waymo and the other cars were all waiting at a red light when the Tesla rear ended them at 98 mph.
The driver of the Tesla was not impaired at the time of the crash. He says he tried to stop but the brakes were not responding.
The driver was from Hawaii, and it was later discovered that there is someone in Hawaii with the same full name, Jia Lin Zheng, with a record of around 20 traffic crimes over the last 20 years, including excessive speeding and running red lights.
I don't know if it had been determined if the Jia Lin Zheng visiting from Hawaii who caused the San Francisco crash is the same Jia Lin Zheng as the Hawaiian Jia Lin Zheng who has the long record of unsafe driving.
I'm not familiar with the naming conventions of whatever country/culture that name comes from. Is Jia Lin Zheng the kind of name that probably many people have in Hawaii or is it one that is likely rare?
That is not a useful metric for Tesla. They disengage FSD when they detect a potential accident.
Is it "counted" if FSD was engaged within a certain time frame prior to a crash? If so, do you know what time frame?
Or only if it was disabled automatically due to detecting a potential crash?
The latter would still be problematic, as a human driver noticing a problem just prior to the FSD disabling itself would potentially be missed (right?).
Do you know who does the counting and who makes the rules in this regard?
Asking as you seem to have more knowledge here than me.
The data is collected in all of these incidents, and most people have seen the clips of FSD avoiding otherwise potentially lethal accidents, so "They disengage FSD when they detect a potential accident" is also just patently untrue.
No comments yet
https://www.reuters.com/legal/government/musks-tesla-seeks-g...
Waymo publishes tons of safety metrics on their website. Here's an analysis/summary:
https://www.damfirm.com/waymo-accident-statistics.html
And the links don't even touch on things that are comparable lol. Waymo might keep all the data themselves as they own the cars, while with Teslas, the drivers can and will just grab the camera data themselves, many post it YouTube.
Why would you trust a word they say when Elon has lied out of his teeth at every single investor meeting for the last decade.
I say this as an owner of a Tesla myself.
I remember this incident. It happened a couple of blocks away. Unreasonable that they let him go.
And from that information alone, you can get the gist of what that data says!
(inb4 you post the accidents per mile chart which is very obviously useless and designed to mislead midwits, as it is not controlled for age of automobile or driving conditions)
FYI. FSD is safer than human drivers on large datasets. Accidents cause deaths of thousands every year. Arguing against FSD for "safety" has The Grim Reaper cackling.
LIDAR/LADAR based systems are not perfect, but do offer mm precision for guidance systems. SLAM based LIDAR systems can be very good, but are also not perfect when forced to guess where a platform is located.
Cheers, =3
Using vision for driving is something that has worked for as long as cars have existed. Trying to push some "millimeter precision" solution with unproven feature set and prohibitive hardware accessibility is just asking for no real safety improvements and just more lives lost.
Cheers.
How so? Honestly asking.-
https://en.wikipedia.org/wiki/Bathtub_curve
>Bathtub curve
you should try your luck selling exotic snake oils.
Ask questions if you don't know, as being rude is not constructive. =3
Most guidance platforms would use LIDAR/SLAM to describe the local road surface, and overlap camera vision data to extrapolate distant surfaces and objects. Note distant objects also have lower resolution, unknown non-distinctive features (speed bump, or open man-hole cover etc.), and increasing sparse data as velocity effectively lowers world-state sampling rates.
The world-state is constantly changing at every intersection, sampling constraints add latency, and the navigation way-point goals may reach contradiction with immediate path-planning due to ambiguous/expired information.
Cheers =3
> you know a dog hidden behind a car doesn't actually vanish nor remain stationary.
Computers can do that, too. It's not that different ... guessing missing words.
I think it was two years ago that Tesla told the public that they could do that.
I suspect a Amish horse buggy is more practical. =3
Lol, really? They either developed superman x-ray vision, or just tracked object occlusion with a common re-acquisition mitigation (so worthless when physical inertia carries a vehicle into an object collision.)
>So, it doesn't "grow exponentially",
The further the object... the more possible choices will need to be made in the guidance system. Note, guidance and navigation are related, but different problem domains. Roughly, the possible choices (and errors) if I recall grew by:
((m cars) * (k lanes ) * (r occlusions) * (s sign laws) * (1 + world_sate_delta_error(t)) ) ^ (n intersections + w way-points) = 1/hype_correction
...but that doesn't even cover the projected future risk(t). =3
for waymo itself, you can overfit on 100% of the situations that will be encountered. 49 square miles isnt that large. its the real world outside that which im concerned about its efficacy in. i think if you put a waymo in a small town that no alphabet engineer has ever even heard of, then youll see it fail badly as well.
FSD is a reinforcement learning problem, and we have no good way of training non-simulation algos for that. and a real dynamical driving environment cant be simulated accurately enough
Which is why it is a non-goal for Waymo. It should be a non-goal for Tesla too, given the state of the art.
Waymo is not claiming to work in small towns.
Tesla is. Soon™.
This could very well be true, but if you’re looking at it from a perspective of someone who lives in a rural area with real winters, for driving purposes, those all look like pretty much equivalent large American cities without a winter.
My last taxi ride involved jumping red lights and speeding through residential streets at 60mph because, I assume, it was early morning and the driver had learned from experience that he could get away with driving like this.
The previous experience to this was a lecture about a certain religious ideology and how I should spend the next two weeks reading up about it.
In 10 years when robotaxi companies are short on cash and trying to IPO they will absolutely start speeding. They'll lock the doors and give you a paid presentation about Scientology during your ride. "Accidentally" drive you to a competing store that paid for sponsored traffic, instead of the .
Perhaps you live in a location that lends itself to such events...
So it’s not a hold-up, but definitely a form of robbery.
Robbed: Bishkek, Astrakhan. Former nicked my SLR at a gas stop, I should have been more attentive, but wasn’t expecting him to loot my luggage. Latter delivered me to his buddies who threatened violence unless I turfed over every penny I had. Joke was on them as they thought I was a loaded oil exec when I was actually just a broke backpacker.
I’ve also had the old shake-down fare in Ljubljana, Bucharest, and Riga, off the top of my head - but I don’t count that as robbery, just assholes.
Edit: Here is a link to Karpathy discussing the trade-offs: https://www.youtube.com/watch?v=cdiD-9MMpb0&t=5276s
You really expect a Tesla employee to speak out against Elon?
Especially when $10M+ TC is on the line?
In cities, high-speed rail and e-bikes make more sense than a honking traffic jam at 4am. lol =3
- What are the main challenges in building software that relies solely on camera input?
- Which specific modules or tasks still require LiDAR to function reliably?
You can tune the camera obstacle detection to be hyper-sensitive, which results in phantom braking, causing Passengers to feel that the car is "unreliable" while it actually is safer. Humans are better at braking the appropriate amount when they see something strange, dynamically tuning their sensitivity in a new situation.
You can lax the sensitivity, which will reduce false alarms, but will actually cause more crashes, deaths, and injuries. You don't want your customers to feel unsafe, so from a business perspective you will inevitably reduce the sensitivity.
He's just wrong about this.
LIDAR will forever stay a niche technology. If it was so notably "better", Tesla would have just scaled that. But they went all in on vision because it has worked for 100+ years.
Private moon landers (the Japanese being most recent one) keep crashing because they rely on a single high-quality altimeter and expect it to work perfectly, all the time. If they had a complementary low-quality backup altimeter that operated independently, they would have had a less failure prone distance estimation system.
All the information is there in a video feed, but the amount of work to get reliable perception from it is not small. With LIDAR and radar you get to the end goal with less uncertainty.
You can probably eventually ("never" is a long time after all) get a computer to understand the world as well as a human purely through camera based sensors, but it's a much more difficult task than taking an approach that uses tools computers are already good at. Similarly, I suspect it would be an uphill battle to have a human drive using raw LIDAR input.
I've been thinking maybe vision-only was a reasonable decision, back when lidar was expensive and the software was hand-coded. Now it doesn't, because lidar is cheaper and the software is and end-to-end neural net, and additional sensors are just more inputs to the network which will learn to use them. But Tesla is locked in because of the promises they made to early FSD buyers.
If you actually look back at the E2E tweet, musk only says that the NN replaced 300k lines of "control code". Control code usually doesn't encompass the entire AV software stack, but neither should it take 300k LOC. As far as I'm aware no one is 100% sure what they mean by E2E and if it's actually the standard meaning or something else that's been widely misinterpreted.
I don't care even if you said you had 70% of the experience their team has, what you say can't sound reasonable or caring for actually improving safety in numbers.
Well considering Musk record for adherence to reality, question is if you can really believe that or if Musk thinks that this is happening or it is not happening at all and Musk is just making it up.
Plus various Tesla engineers have said the same thing, Tesla does have a very large AI training cluster, and FSD quality made a big jump when they claimed to deploy end-to-end.
When the software was hand-coded, having Lidar and a high-res map was vital.
But if you have a good enough AI, sensors that replicate the human senses are all what is needed.
The real question is: when will we get good enough AI that can be applied to all cars?
Back in the day, Elon's specific objection to lidar was that it was too hard to code the sensor fusion part. With end-to-end there's no coding to deal with.
And it's not like they've replicated the visual cortex. It's the same neural net technology everybody else is using. It can deal with any sort of sensors just fine.
What is your example or evidence for this?
Also, if you've ever done any ML you would note that more data isn't always better. Plus there's the piece about which thing to believe when you get conflicting data. It's a lot more to it than what random Hacker News people are saying in this thread.
The cofounder of Waymo taught one of the first Udacity courses on this subject. He went through a small Python project that processed lidar point clouds for self-driving. The data is noisy, you get conflicting information from different points, and the code aggregates all that into the most likely 3D model of the world.
Additional sensor inputs are just more of the same, and neural nets are pretty good at this sort of thing. They'd even learn which sensors are more reliable in different scenarios.
As for "more data isn't always better," I've mostly seen that applied to training, not inference in real-time control systems. Even for training, it turned out people had been fooled by a local maximum, and once past that, more data really was better.
- What living creature is using wheels to move around?
- What kind of birds come strapped with a jet engine?
Sometimes non-natural solutions are easier and often better than attempt to replicate nature for every cost. Imagine your logic applied on a plane - birds flap their wings, thus this 737 should spread it wings and flap away like a goose. Now take military goose and flap fast enough to get supersonic...
A human driving a car.
I agree with you that sometimes non-natural situations are easier and can absolutely be better, but the point of bringing up humans is generally to show that it demonstrably is possible to do at least as well as humans, with humans as an existence proof.
But that it might well take too long and cost too much to get there, and that it might well in the end be cheaper and better to use additional types of sensors is a good point.
But more specifically to this case, human eyes are attached to brains with (generally) vastly better image recognition and reasoning abilities than any camera based self-driving car. Because of this, humans are better able to recognize visual input even in degraded or unusual conditions compared with a computer.
(Also, I don't know what star-system you grew up on by my lidar sensors are next to the tubular sheaths on my cephalothorax, right where Xoc'tlz'ik (the Creator) intended them to go.)
Despite all this, a lot of modern cars are adding lidar to make things easier for me.
We possess a spatial intelligence (e.g. how your brain has an approximation of: It feels like I walked three blocks) that will never exist in this "photons-in" "controls-out" fantasy.
And they may well end up with something "that works". But... many buts.
Waymos which cost as much as an actual taxi and aren't scalable are nothing but a VC money carnival.
I would say if that is all you have to discredit the Tesla robotaxi project, then the project seems to go pretty well.
“Harder, better, faster, stronger,” the $1 trillion company wrote on Monday, a double entendre referencing the synth pop track of the same name by Daft Punk, a duo appropriately known for performing as robots. Musk approvingly reposted the phallus-shaped service map, adding the fare would now be hiked to $6.90 per ride from $4.20 previously, both numbers the 54-year old often employs for comical effect.”
It’s amazing that anyone still takes him seriously. One of the first riders almost hit a train, and this is what he’s spending his time on. It’s inexcusable, but his fans lap it up.
Why would they use what was a science role model that way?
On national TV.
In a presidential inauguration.
A permanent stain on US history.
Without government contracts, SpaceX never succeeds.
Without a mysterious visa, Musk never stays legally in the U.S.
https://en.m.wikipedia.org/wiki/Bellamy_salute
So no, you're not because its meaning became muddy at best
Was interesting to see actual video of a newish public figure wave at the crowd, wince and look in the direction of the camera. Clearly newly trained not to wave at the crowd.
https://x.com/elonmusk/status/1944688226037325868
And for people who are like "yeah but he gets results," are you really saying he wouldn't be getting more results if he didn't spend the last 10 years being an idiot online?
proof left as an exercise for the reader.
his being cringe itself isnt the issue. its the lack of EQ that leads to someone being a slave to wanting to be liked. it doesnt inspire confidence in his ability to think rationally. like how hes perfectly fine grifting and contributing to the decrease of trust in society. its messed up
There's a model Y with The Sticker on it in my street, introduced years after he showed his colors like that. No, dear neighbor, when you bought that car, you knew.
Not so long ago you'd gain traction through insight or depth. Now it's absurdity that wins the price, vanity upon vanity.-
At some point the idiogarchs are going to celebrate (celebrate) their accidents and dead as a win. Lol.-
How many human drivers almost hit something?
When a machine kills someone do we go after … the company? (Maybe we should all be shorting robo-taxi companies.)
So average driver is a low, probably incorrect bar to meet.
There is a limited scale to any taxi service, at least until it becomes cheap enough that the company can afford to have lots of cars sitting parked and idle for much of the day. Otherwise, the rush hour peaks will be less convenient and affordable than having a private car.
Elon has a good track record, but he has some duds in there too.
Anything Musk says is simply not trustworthy.
Not just my opinion but proven over and over again.
Hiding safety data is just icing on a cake that he has been baking for years. The real problem here is that those who choose to trust him and his companies are putting others at risk.
Those who own/drive a Tesla typically have insurance but what about those who hire robotaxis? Can they be held personally liable when their roboride hurts someone?
Tesla on the other hand started with a load of constraints - smaller amounts of compute, cheaper sensors, needs to hit a price point as it will be installed in every vehicle sold - can't be $50,000 of dedicated self driving hardware per vehicle.
There are some scenarios where constraints lead to breakthroughs, but often for true-moonshot projects it is the opposite (see: Space shuttle, System/360, the manhattan project, LHC etc)
Waymo has sunk billions of dollars with almost 0 return yet.
Tesla has developed FSD basically for free, because people pay for software that isn't there yet, and Tesla doesn't waste money on expensive hardware.
I also think Tesla and their robotaxis are egregiously overhyped, but cutting Lidar was not a terrible business decision at the time. Though it's a terrible decision to have still stuck to vision-only (and using low-quality cameras apparently) when they could have at least got Radar and cheap Lidars.
Tesla's architecture is the only one of the robotaxi providers which actually _can_ scale. Not being dependent on exact HD maps, having the same standard as humans do "if you can see enough to drive" means eventually there'll be scant roads Robotaxi can't serve. By this strength it'll serve in areas that would never work for LIDAR/pre-HD-mapping requiring solutions
But I do wonder if rising nations will instead just build their infrastructure to line up with technology (e.g. HD mapping their streets as they build them, installing signed beacons to inform AVs). The cost of this tech is cratering. One possible future, though admittedly an unlikely one.
The stan is strong, in that one...
AFAIK what Tesla's software does is object detection (cars, people, bikes, road sides) and only uses that. Which of course is not at all safe since there can always be things on the road that are an obstacle but not recognized.
What it should do is model the entire world around it in 3d and consider that. But it doesn't. Cow on the road - problably not recognized. A child in an uncommon halloween costume chilling / lying in the middle of the road - pretty damn sure not recognized and the Tesla would just kill that child. Yep.
(I drove a Tesla Model 3 with the latest "self-driving" software for a bit a year back.)
This is not true; in modern FSD the visualization is disconnected from the actual driving model. The visualization runs a lightweight model that labels some stuff and shows it; the actual heavy lifting is now a single model that takes pixels as input and outputs control commands. My car very clearly reacts to stuff that doesn’t appear in the visualization.
https://www.cnbc.com/2025/06/18/waymo-cars-are-coming-to-new...
I own a Tesla, I use FSD all the time. It is nowhere near viable enough yet to release it unsupervised to the masses with no driver at the wheel.
I have also ridden in Waymos and felt way more comfortable with how it drives. It's in a different league.
i do agree he should say its in early access or something
The only thing preventing Tesla from going bankrupt is trade barriers on Chinese EVs.
Elon has always been a snake oil salesman and I suspect this robotaxi venture will blow up in his face in a way that's potentially fatal for Tesla. All it will take is someone to be killed in or by a robotaxi. That'll bring in the authorities and likely reveal what a shitshow it is behind the scenes.
But we know that stock prices are based on expectations for the future not fundamentals. So massive growth is built into the stock price but we have to start asking where this growth is coming from. That's why there's a robotaxi effort.
EVs are becoming a commodity. So over time Tesla's price direction is towards the Big Three. But with more risks, namely the Chinese supply chain for batteries.
You can view Tesla as an investment in Elon, which I think the market has. But that's on incredibly shaky ground. He's torched his relationship to the current administration going so far as to call the sitting president of the United States a pedophile [1]. Elon doesn't seem to appreciate that Trump could end Tesla with an executive order.
My point is that Elon is volatile and can so easily become a liability to Tesla. They're paying him billions so firing him can immediately and massively improve the bottom line.
As they say, the market can remain irrational longer than you can remain solvent. So I'm not predicting the imminent collapse of Tesla but I think it will come and it'll be sudden and dire.
[1]: https://www.ndtv.com/world-news/elon-musks-pedophile-dig-aft...
Now it's robots, $10 trillion market.
Cruise was/is fucking awful. But these tesla taxis somehow 10x worse than Cruise. Going to end up live the governor of this state and without the ability to recover funds due to awful tort reform.
Some of those things that went wrong were dropping people off in the middle of an intersection, not recognizing railroad track guard rails… and simply not being able to get them to their destination at all.