> showing a low-speed, closed-course theme park ride in order to build confidence around Tesla’s progress toward actual real-world driverless capability is almost too childish to call a fraud
This guy has lost confidence in Musk's promises.
I never really understood how Musk can claim to be so close to full autonomy when these cars clearly struggle with basic things using the exact same system that supposedly drives by itself. Automatic high/low beam or detecting rain does not work very well. You wold think that a system that cannot detect oncoming traffic (high/low beam), or actual rain has some other problems when it comes to self-driving.
There's a Munro Live video from the event on YouTube where the host is talking enthusiastically about how the Optimus robots are going to take over all dirty and dangerous jobs and the whole time he's talking the video is showing a robot at the event failing miserably at picking up a small bag of chocolates.
The bag has clearly been designed to be easy to pick up, it's positioned on a custom made table top to perfectly space them out and make it easy and still the robot manages to lift two of them by mistake and topple a whole line of them with one ending up on the floor.
Optimus robots were remotely controlled by humans at the "We, Robot" event during Cybercab. They are not ready even for demo purposes.
It was clever trick because many attendees did not even think the possibility hose were teleoperated interactions because Musk did not bring it up. Some investment firms were clearly not amused by it.
It seemed pretty obvious to me. Yet, I was pretty impressed nonetheless, I mean, the sheer dexterity of movement of these things with a human-hand anatomy surprised me, I don't think I've seen it before, and didn't know carnival toys are at this level yet. Is it a commonplace technology? It seems like it must have some practical uses, to be able to operate an almost-human body remotely like that. But, as I've just said, I don't think I ever saw this, only clunky specialized robots with poor movement control. And Boston dynamics pre-programmed toys.
On the cars with matrix-led's the high-low has been solved very well. It used to be bad and now it's almost perfect.
Wipers, dunno, they used to be really bad but I have not noticed the issue in the past few months, maybe they have done something.
That said, I am still agreeing with you on "struggle with basic things", there are phantom breakings in the exact same spots this and other teslas have driven thousands of times. There are occasional cases where it would follow the wrong road-markings into a crash would I not intervene and so forth.
> with matrix-led's the high-low has been solved very well. It used to be bad and now it's almost perfect.
Yes, it works much better on my new Model Y (2024) than my old Model S (2014). But they didn't really solve the original problem, i.e. detecting oncoming traffic vs. street (or other lights) up ahead. The matrix led's darken these areas in any case, and neither street lights or oncoming cars complain of course. But it does not fill me with much confidence.
> Wipers, dunno, they used to be really bad but I have not noticed the issue in the past few months, maybe they have done something.
Then that software update has not come to Norway at least, here they still start wiping in dry conditions, and fail to detect actual rain; so they need to be manually started all the time.
> there are phantom breakings in the exact same spots this and other teslas have driven thousands of times.
Isn't this the problem with AI though? We've taken something we don't fully understand (neural networks) and applied them. We can give them a lot of training data to the point where they seem to be able to do stuff on their own. But when they fail, we really cannot do anything besides adding new nodes to our network and add more training data, and hope for the best. We cannot really say what went wrong, and fix it.
That's for safety if a flake of dirt or dust lands on the front window right over where one of the cameras are, it will do a wipe and try to get rid of the thing that is distorting the image. Or it will do it if you are driving straight into the sun sometimes, presumably because it's doing a best efforts action in case the visibility problem is further exacerbated by dust which can be cleared.
It's strange to me that actual helpful features are misinterpreted as flaws, but that's how it is with new technology I guess.
Failing to detect actual rain, yes I see that too a little sometimes. I think they should augment with audio detection. There is a microphone, obviously, for the speech button, and depending on how things are wired sometimes speaker cones can also be repurposed as microphones.
> That's for safety if a flake of dirt or dust lands on the front window right over where one of the cameras are, it will do a wipe and try to get rid of the thing
Not that dry conditions. I live in Trondheim, where it rains/snows well over 200 days a year, so it's not wiping off some dry dust of the camera area. It just starts wiping the clean window in perfectly fine weather. At other times it rains quite a bit and it does not come on.
The software update to soft click the left stalk to engage the wiper menu via the left steering wheel roller is a nice quick fix for these situations, so they are doing something.
I was not saying that there absolutely is always dust there. I was saying the system is doing a best efforts action just in case there is dust or snow or other debris there which there sometimes is. I'm guessing they have birds in Trondheim.
I haven't tried the roller for the wiper menu, thanks for the reminder. Less rain around here.
> I was saying the system is doing a best efforts action just in case there is dust or snow or other debris there which there sometimes is
I guess that could be the case; but they do seem quite random.
> I'm guessing they have birds in Trondheim.
We do, mostly seagulls where I park my car :-) But you're not dry wiping away anything they produce, that takes a good scrubbing.
> I haven't tried the roller for the wiper
It makes the experience fully controllable from the steering wheel using one hand, which is very good. One light press on the left stalk, and the menu pops up down on the left corner on the screen where you can roll to your preferred setting. I hate having to go into menus on the touch screen while driving.
Honest question, considering I have never driven or been in a Tesla, when you say “it would follow the wrong road markings into a crash would I not intervene”, are you not concerned for your safety when driving it.
As an owner of a Model Y -- I'm aware that its current implementation of FSD would do this. I don't pay for FSD, so it does not do it. Also, this is a bit like someone turning on cruise control and then just letting the car go off the road at a turn. Would you be too concerned for your safety to own a car with cruise control? No, you'd just either not use cruise control, or understand its limitations when using cruise control and steer as needed.
That only seems cosmetically similar to me. If cruise control would usually do the same thing but occasionally try to take a curve for me, I would consider it unreliable. Even if it always got the curve attempts it made correct, it is unreliable if its actions are not essentially a mechanical response.
Right, this is the difference between most car functions and then something like FSD. Reliability. Cruise Control always works the exact same, where the car will break/accelerate exactly as much as it needs to, to maintain the speed it is set to go. Same thing with ABS, where the brakes will be perfectly pulsated when heavy braking. Or differentials. Or pretty much any car system.
FSD is a little too wild west. I also lump ACC and LKAS in here, too. Phantom brake and bad directional input exist for those way too often.
But it’s a function you can choose not to use. My Model Y doesn’t even have FSD, I didn’t buy it. I could buy and use it, but I don’t. It doesn’t concern me that it exists, the Model Y has been an awesome car. If I did buy and use it, its limitations are widely known and I’d expect to supervise it if I used it.
I never paid for any of the autonomous upsells, so for me this is a lane-keeping functionality and I treat it as such, meaning it helps me but I am still fully responsible for my own safety.
It happens rarely enough, and as I mostly drive on the same roads, I know where it happens as well, usually I disconnect the thing before it gets to the problematic place.
So no, it's so much stressful as annoying. But it does remind me that the claims of self-driving is just horse-shit.
I'm struggling to see how you can be so calm about what you just said. To me, the idea of a system that works 99% of the time and doesn't work 1% of the time is utterly terrifying, far more so than a system that works only 50% of the time. 1% is small enough for a human operator to become complacent, but still many orders of magnitude greater than an acceptable risk factor. It seems like exactly the sweet spot that's going to maximise loss of life.
I know what (and where) the problematic spots are for the lane-keeping, so I use it in situations where the benefit is large and the risk is manageable.
There are many places where I won't switch it on and just drive myself.
Also 99% vs 1% depends on how you look at it. The main road I drive is about 200km long, in one direction it does not have any problematic spots. In the other it has 2 lets say 10 meter spots. By distance it's 0.01% problematic. It's not like Desert Bus where it tries to get you all the time.
Yes. Not only that, my car had the matrix led hardware for 2 years before the matrix feature was enabled in a software update. TBF, I was never sold matrix leds so there was no marketing let down.
Also the auto-wipers on my 2003 Saab were working better than the Tesla ones of the current generation.
That said, it's also funny to think that the only things I could complain about owning a Tesla are: wipers, self-driving marketing and Elon being a nutcase.
Before the last one is solved, I would not buy another one, but other than that it's a very good car.
Even Musk has lost confidence in Musk's predictions. He jokes about this. Tough crowd out there though, you being a case in point.
It's normal as a human to make predictions, and it's normal to come to the realization that some of one's predictions have been incorrect on the timelines. This is all just normal real life in the world of technology.
What actually is remarkable is the unstoppable progress that is being made. You can find counter examples, but look at the newer software versions this month, next month, and so on. Things are being fixed bit by bit.
At some point a reasonable person would realize that their past predictions have not panned out and would make their future predictions more conservative.
I don't understand why Musk/Tesla is so insistent on having cars do self-driving purely on cameras.
Both from the technical and the marketing point of view having a car that can do better than any human ever could, even in the most optimal of conditions would be a great thing.
It's cost - LIDARs aren't cheap. Tesla is set up as a consumer car company first and a robotaxi company second, costs matter even more. They could jack up the price of the final product, but presumably it wasn't worth it (at least in his mind?). From a safety perspective more sensors in general are better, if integrated properly. With LIDAR prices are coming down I wonder if he will someday swallow his pride and reintegrate them.
It's almost certainly because he dug the hole so deep in advertising. By saying "we need LIDAR", means that his promise "your car that you own right now will someday be self driving" was a lie. I can't imagine the lawsuits.
I cannot imagine how a musk-fan (at this point, I would classify anyone, who believes his claims as a fan) thinks, but it seems so weird to me to actually cling on that hope, that it should make the promise easy to break. It has been a while he made this promise, and it's nowhere near to be delivered. What does it even matter? How long are you comfortable owning the same car (the one that has more electronics in it than iPhone)? 10 years? 20 years? If he doesn't deliver FSD in that time, it doesn't really matter if it will be the same old cars that will achieve FSD. And if you can achieve FSD with better technology on a newer car, it's a win-win, it doesn't really break the promise that your 20 y.o. Tesla will achieve FSD. Maybe in 30 more years it will.
Both the Valeo SCALA and Luminar LIDAR sensors are in the $500 ballpark (at volume). Even a few times over, this seems like a petty cost to forgo on a $50k car.
Anyway, Tesla have supposedly recently bought a massive pile of Luminar sensors, so I expect them to be in one of these vehicles soon.
I think the robotaxi might be the only actual viable approach to developing fsd.
You control the hardware and software, you drive in real world conditions, you can deploy regionally and control the hours and conditions you operate in letting you gradually increase when you operate as you judge it safe.
And if the robotaxi makes a mistake, well it's still in testing phase/beta.
If Tesla comes out with FSD then it has to be damned near "perfect" in all conditions, which just is never going to happen out the gate?
You can even design some type of system where the taxis can only physically go on a routed road, and then you won't have to worry about real world conditions. And then you can just integrate some sensors into the routed roads to detect the vehicles' location and then you can guarantee they'll never crash.
Oh! Oh! And then for efficiency maybe you can chain a ton of robotaxis together, so you only need one drivetrain... wait a second...
Why not have at least radar? Plenty of cars have radar for adaptive cruise control and blind spot monitoring, it can't be that expensive. Two sensor kinds would still be an improvement.
I think if you have two sensor systems with one being significantly less capable than the other, the gain from having both quickly diminishes. I believe Tesla dropped the radar because it was simply outclassed by the camera in 99% of circumstances. So the radar really just becomes a hacky failsafe for bery rare circumstances, and Musk would like to avoid such a situation, he wants a "clean" solution.
Sensor fusion could cover that though, cameras aren't good if they are obstructed by dirt, particles (fog, snow, heavy rain), don't capture enough light, and probably many other cases. Fusing camera input with a radar to cover shortcomings of the camera sounds sensible enough.
What I'm saying is that it might be the case that "capability c" (somehow measures as a positive scalar) of your fused sensor system is not the sum c = c_1 + ... + c_n of the capabilties of each sensor c_i, but more like c = (c_1^p + ... + c_n^p)^(1/p) where p is a large power. The case p = 1 is where every sensor contributes to the whole system with perfect efficiency. The case p = infinity, i.e. c = max (c_1, ..., c_n) is where really only the best sensor does all the work and the weaker sensors can not add anything useful. The case of p > 1 somewhat large represents that all sensors contribute but the real heavy work is being done by the best sensor and other sensors only have significant impact if they have roughly equal capability. So I believe that the team at Tesla has the intuition that p is too big for radar to be worth it.
People drive all the time in snowstorms around here in Sweden and Norway, lower speeds but outside of a major blizzard with zero-visibility people still drive and reasonably safely.
I used to drive through torrential storms in Brazil, where the highest level of the wipers was barely enough to keep water out of the windshield, never had an issue when keeping safe speeds either (even though Brazilian traffic is murderous).
Driving in a torrential rain can be terrifying. Literally can't see practically anything in any direction.
Haven't seen an equivalent snowstorm so far. Even in the worst case there's still at least 30 ft / 10 m visible range. Which is not great, but at least something.
In both cases, if at all possible, better to stay put!
But they now charge $5000 per year subscription fee, and they can demand the car be bought in for some onboarding. People are ready to pay for it and do whatever it takes to make their Tesla dance, so cost is not an excuse really.
I've seen this argument brought up a lot, but I've never seen someone attempt to answer the counterfactual. If LIDAR was added to Tesla's now, how much would FSD improve? I've seen people here who have used it say that with the current software FSD is already pretty good, but my guess is that even with LIDAR they'd still be far from L4. So what kind of difference would we expect to see if, say, tomorrow Tesla announced that they're going to start using LIDAR?
The worst thing about my Tesla are the phantom brakes and "emergency lane assist" function (which should be called "steer into opposing traffic for no reason". If I forget to turn those off before starting, my ride will be pure horror.
LIDAR probably would help to make Phantom Brakes happen less often, simply because there would be another info source for "is there REALLY an object that is dangerous to me?".
Those who say that FSD is "pretty good" are living in a fantasy world. There is hard data on miles between critical disengagements (which really should be called "if the driver doesn't respond within a fraction of a second, people will die"), and depending on region, model, weather etc it's between 13 and 115 miles right now.
Over here in Germany there are statistics that a human driver will have the equivalent every 155,000 miles.
"Pretty good" just doesn't cut it when it's about the risk of killing people.
My Model 3 right now detects about 60-70% of school children crossing the road (keep in mind roads in Germany are narrow, and humans including kids are using the roads, too). 30%-40% of those I would kill every morning on my way to the office.
And the thing is: 70% isn't enough for this, 80% isn't, 90% isn't, 99% isn't, and 99,999% isn't.
Side note: People constantly claim that Waymo is autonomous. It's as autonomous as a tram. They only work because it the cities they operate every single road they use have been mapped by hand and is constantly updated. Send a Waymo into my city over here, and will also kill a couple of kids per day. Years ahead of Tesla? Yes. Good enough? Hell no.
Looking around at the percentage of drivers around me on the road with their face looking at a phone, while moving, is probably in excess of 5%. These people won't see the kids crossing the road, either.
Adding an extra sensor onto something not designed for it may not do much. So if that specific question is important to answer, the answer is probably 'not a whole lot', especially since the AI has been training on images only, AFAIK.
The problem isn't that the sensors need to be added now, the problem is that without the sensors the cars are half blind and are more likely to make bad decisions and refusing to add them at all shows bad judgement.
Rain, snow, mist, fog, darkness -- these are things cameras are very bad at seeing through. Is that thing in front of us solid? Without extensive training on that particular type of object, the AI with the camera has to guess. I'm sure these aren't the best examples -- I am not an expert in LiDAR. But I do have experience with computer vision with camera systems, and they are woefully insufficient for life and death decision making systems when they haven't been trained on the exact specific scenarios they will encounter.
My guess is that it's not the lack of sensors that causes the problem with autonomous driving, rather it is the processing of the data. What should the car do if there is a flood covering a dip in the road?
> Both from the technical and the marketing point of view having a car that can do better than any human ever could, even in the most optimal of conditions would be a great thing.
Given that the leading causes for road accidents are speeding, DUI and distractions [1], a car doesn't need to be better than any human reasonably can. It just needs to be on the same level as a well-rested, sober human without a phone in their hand.
The decision LIDAR vs cameras is the pareto principle in action: what matters is the 80% - the mass that causes the most accidents.
The history of technology is pretty much directly opposed to your wishes.
The prosperity of technology basically solves the problem with the most terrible hardware possible.
For example, if you have a really good microprocessor that works, what you should do is shrink it until it barely works or speed it up until it barely works to get a similar version for cheaper or faster version for the same price.
Same goes with cars. If you have a car that works with ultrasonic and lidar and cameras, eliminate the expensive sensors and reduce the cost.
Reducing the cost will let more people afford the cars, and sell more cars.
To Musk that is basically his plan all along - to replace ICE cars with EVs. Just wish we could keep turn signal/wiper/headlight stalks.
It does in certain strictly determined conditions + human assistance.
But the point Tesla didn't go through "FSD with ultrasonic + lidar + cameras", and then start to reduce the sensors. They went from "no FSD" to "broken non-working FSD, but now only with cameras". That is not how you strip down a working product to reduce costs.
Waymo has sensors but they don't sense purely in real-time. Meaning the roads are mapped out in detail ahead of time. This is fine and effective, but it makes the comparison not really fair. A human driver will do splendidly in an area they've never been before, whereas a Waymo taxi won't.
I have an idea, replace the existing cameras with early 2000s era webcams. Since worse technology solves problems, it would work better.
If Musk had gone all-in on every sensor possible and put them all over the cars like strings of Christmas lights, I assume you would be defending that decision with just as much fervor.
Perhaps the worst hardware that can solve FSD is hundreds of sensors of all types plus a fleet of attached drones for other angles. The worst thing that could work is different from the worst thing.
They shrink their chips, they up their clock speeds and keep the costs the same.
They also remove hardware - they've relentlessly removed ports, intermediate interfaces (like removable batteries, memory and processor sockets) and changed screws into glue.
> you have a really good microprocessor that works,
and applying this analogy to the current situation: do we?
normally you have to sell a product that works first, take the profits and then invest that into improving the components to make more profits. I think Elon has cracked the code: sell the product you don't have to people who would love to have it, and you don't even need to produce a version 1 that works - just move on to solving the problems that you want to. the market will pay you either way.
Musk gets obsessed with big bets when people tell him something isn't possible. That, plus the hardware costs are much saner for Tesla if they don't need a bunch of lidars.
That said, this particular big bet looks pretty bad for Musk. He's been confidently promising self-driving either "this year" or "next year" for the last several years, and so far Tesla appears nowhere close to getting L4/L5 driving down.
I mean, Volvo is already delivering the EX90 with a LIDAR integrated from factory for similar money to the Model X - and I can't imagine Volvo is losing money on them, so it can be done. Tesla just chose not to.
And yes, I know the EX90 is currently an absolute mess in terms of software and their LIDAR isn't even enabled for driving assists yet, but the hardware is there so it can be built on scale.
I presume there is a bigger game here. Androids, eg humanform robots, will require the capacity to navigate on camera alone.
Specifically, loads of object detection with smaller objects, imagine an android cleaning off a table after dinner. Visual spectrum, camera identification is vital here.
So it's all one ball of wax. Tesla is not juet a car company, likely in 5 years the number of androids they sell will dwarf car sales.
> Androids, eg humanform robots, will require the capacity to navigate on camera alone.
But those cameras don't have to be simple visual spectrum cameras, do they? They could have, for example, in each eye, LIDAR, infrared, several focal lengths of visual, etc. Not to mention being able to augment those sensors at places humans just don't get sensor data (e.g: back of head; outward facing from shoulders, hips, etc.; constant presence monitoring via ultrasound / IR / 2.4GHz, etc.)
But those cameras don't have to be simple visual spectrum cameras, do they?
They do, if they want to do comparative analysis on the trillions upon trillions of existing photographs, and even motion picture frames.
Not to mention the entire AI industry is working full out for perfect image recognition, again in our visual spectrum. Something apparent to a person working with openai 5 years ago.
Tesla will likely never accomplish anything with robots because they can’t even build a car to typical market standards of quality and robotics is even more greenfield and unrefined in its mass production processes. I guess maybe it’s possible if someone like Shotwell takes over and Musk is relegated to a similar figurehead position.
Because he needs to maintain market value. Musk is part inventor and part salesman. He needs high valuations to keep risky endeavors afloat. What's better than sell the future.
This is a good practice to reduce at least some amount of dependency to outside US. Also this is the definition of innovation; making high tech out of low tech.
one answer is due to the fact that humans also do this with just 2 pretty bad cameras and a lot of offloading to the cortex.
It also simplifies the stack a lot to have a single set of sensors, so the software becomes mostly: getting good training data (iterative loops from failing production cases) and an efficient training algorithm.
This scales to more than just AD and also can leverage new breakthroughs from academia
> one answer is due to the fact that humans also do this with just 2 pretty bad cameras and a lot of offloading to the cortex.
No, humans do significant sensor fusion.
There's binaural audio: useful to detect and have a rough position of emergency services and/or high speeds cars(it's true that its a relatively unencumbered channel, but that makes it all the more valuables for emergencies)
And there's a working if imperfect IMU (performance can be altered if the power supply is set to an alternate mode) to sense all kinds of acceleration: for fine course correction on acceleration and bearing, for getting the road condition and adjust the driving profile accordingly, etc..
Humans don't just have two eyes though (and our eyes are pretty damn good as far as organic light sensors go), we have 3 mirrors giving all-around sight, a whole body full of nerves providing feedback, and excellent 3D reckoning to keep track of other vehicles.
Our eyes (+ brain visual processing) are way better than the vast majority of cameras you can buy, judged on things we care about in driving. Typically only high-end film cameras approach the dynamic range of our eyes, for example. The only major thing that even cheap cameras easily beat us at is zoom, but that is mostly irrelevant for this use case.
Humans have millions of years of evolution in vision processing, and yet we still regularly have 100+ deaths a day in the US, plus many more injuries and fender benders.
I don't doubt that it's possible with machine vision alone, but it makes the challenge substantially harder.
Another way of saying this is that we do this with a system comprised of human eyes and a human brain. We’re very good at making machine eyes, but the brain part is proving extremely difficult for us to reproduce with machines.
Yup - if there was a camera today that could consistently (and without any fiddling) reproduce things the way my eyes see them (especially in high dynamic range or low light situations), I would buy it immediately!
Humans drive with only 2 cameras pointed in a single direction. A Tesla has 8 cameras pointing in every direction. So by default it would already have superhuman awareness. But the difficulty is not perception, it is planning. Turns out to drive among other humans on public roads, you need to correctly understand and predict human behaviour first, which is like half the way to AGI.
Yes - but humans possess other senses, and crucially, our eyes have a much greater dynamic range than even the best cameras (and they are not fitting the absolute best cameras, because that would be even more expensive than LiDAR).
It also does away with one of the crucial aspects of "better than human safety" - driving in conditions where a camera is uselss - think moderate to dense fog, heavy rain etc. etc.
We don't have any senses that aren't trivially replicable in a car. Microphones and accelerometers were already plentiful before. And HDR is not really an issue anymore since all carmakers dynamically vary exposure times in their cameras. Other carmakers have tried to put LiDAR and stuff into cars for many years, but it didn't magically give them self-driving either. The key issue to solve is planning - not perception.
In case you really have two eyes pointing in a single direction, you should see an oculist urgently. Mine are doing a perfect 3D picture. Also, unlike the eyes of my Tesla which becomes blind when it rains, or if there is any light reflection, I can still see things.
Also, should your eye cameras be in fixed position, you even more urgently should see a doctor. Mine can be turned 180° on the X-axis, and 120° on the Y-axis.
And finally: My eye cameras have working high-speed auto-focus, HDR, are protected against rain or snow, can operate at very low ISO, and have retina resolution.
Human eyes have much higher resolution than Tesla's cameras, better dynamic range, no issues with rain or dirt on the lens, can take measures to avoid most glare from the sun, and are connected to a brain that is vastly better at driving than AI.
It makes no sense to say "humans can do it with only X, so machines should have to as well". Do cars run? Do planes flap?
I do agree that planning should be the hardest part of self-driving, but go and look at any Tesla FSD video on YouTube and it's clear they haven't solved the perception part yet either. Cars wobbling all over the place, morphing into vans, etc.
I think this article does a good job in describing why I always found Tesla's tech demos so jarring.
There's a certain amount of "movie magic" that goes into any tech demo to make it work well on stage but Tesla's demos never gave me a sense that what I was seeing would actually line up with how their cars will perform in the real world.
You have to be ambitious and aim high in the tech industry, but at some point you have to level out and stop getting high on hype.
So does Tesla FSD use cameras to create point clouds in real time with photogrammetry that they then segment and classify via AI and act accordingly, or is it something else entirely?
Something else entirely. Similar to an LLM, V12 of FSD is trained on on tens of millions of clips of people driving. The model takes in the history of frames over some period of time and spits out a path for the car to take, conditioned on the route from the regular route planner. As with LLMs, this approach has the benefit that it scales well according to a set scaling law: add more data, add more training and inference compute and your model will get predictably better. Tesla has the ability to sample driving behavior from almost any of its 6+ million cars world wide, based on certain triggers. The video (usually 30 seconds long) is uploaded to Tesla along with the actions that the human driver took during that time. Training a model like this is called Imitation Learning, or Behavior Cloning, and it's all the rage in robotics right now. But Tesla also has direct intervention data from its 0.5m users running FSD. Whenever a user intervenes (by disengaging the system through the steering wheel or brake, or by pressing the gas), the clip gets sent back. Tesla can then train on this clip directly to correct bad behavior, this is called Expert Intervention Learning, and it's mathematically proved to fix certain problems with raw Behavior Cloning.
"Ploy for credibility, retreating even deeper into fantasy, pretense of reality, pure fantasy, contained within a world of fantasy, retreating into misinformation-fuelled fantasy world, runaway ego, less palatable, self-soothing fantasy bubble, fake statistics, blown deadlines, litany of bullshit, chronic inability, unprecedented publish delusion or deception, having drawn a poor hand, financial-cognitive nightmare, it’s hard to imagine anyone seriously believing that a night of delusional Disney Adult cringe, too childish to call a fraud, Musk’s increasingly-degenerate gambling run is slouching toward one last big coinflip"
Some people have a rather unhealthy obsession with Elon Musk, and ought to take a few hours off the Internet and get some sunlight.
Do you have any actual evidence of that, or is it something you like the idea of because it's difficult to accept that Musk is a real arsehole and total idiot about some things (e.g. steering wheels, divers), yet also a very successful engineer (yes) at other things (rockets, electric car platforms).
Why would it matter whether he technically founded Tesla? He basically built it from scratch regardless.
Also What does that have to do with him (supposedly) being a talented engineer? He’s a brilliant strategist/visionary/CEO obviously but Cybertruck was a product of his engineering decisions then.. yeah.. probably he isn’t.
Right, consumers are often stupid beyond belief and will buy an objectively worse product purely for prestige and signaling to other humans. It's that social part of the brain going off.
That doesn't change the fact the Cyber Truck is an inferior truck to just about any other truck in just about any metric. It's a testament to the sheer arrogance of Man and the failure of modern American education.
Looks like it’s designed by a not particularly talented three year old?
of course not everyone is bothered by that. All the water/environment related issues (I guess you can get the “optional” coating) and the absurd repair costs seem like a bigger issue.
> Tighter turning than everyone
Doesn’t the Hummer EV have a 25% smaller turning radius? Also a removable roof.
Do you have any evidence of him ever being one? I mean he’s an extremely successful CEO and decision maker which is almost infinitely more important than just being an engineer.
I mean was Jobs ever considered an “engineer”? And I’d bet he was significantly more involved in the design of Apple’s most successful products than Musk was in SpaceX (and possibly even Tesla if Cybertruck was the result of him going of the reins).
Go and watch the videos of him touring SpaceX. He's clearly involved in the engineering.
You just don't want to believe that because it complicates your view of him as a purely worthless arsehole. You can't accept that he's an arsehole and a skilled engineer (or project manager if you want to gatekeep "engineering").
Imagine him coming up with the spaceship equivalent of the CyberTruck at SpaceX. Even Boeing would seem like being great at what they do at that point…
You read any biography of him any testimonials from people who've been with him. It's well known and obvious he gets into the weeds. THATS why his companies are successful.
> It's well known and obvious he gets into the weeds. THATS why his companies are successful.
What about the PayPal stories? Or those are all supposedly “not fair and balanced”?
> It's well known and obvious he gets into the weeds
That’s why the Cybetruck looks like it was draw by a 3 year old?
> THATS why his companies are successful.
Why? You don’t have to pretend you are an engineer to build an extremely successful company as CEO while having significant impact on the products it designs. Just look at Apple.
>What about the PayPal stories? Or those are all supposedly “not fair and balanced”?
What stories?
>That’s why the Cybetruck looks like it was draw by a 3 year old?
So? This doesn't serve to disprove my point. Elon gets into the weeds. Your opinion of it doesn't change the success of the company and the technology.
>Why? You don’t have to pretend you are an engineer to build an extremely successful company as CEO while having significant impact on the products it designs. Just look at Apple.
I never said anything about apple. I said that's why HIS companies are successful. There's plenty of reasons why other companies are successful.
It's a well known fact Elon Gets into the weeds of engineering and tech. Just watch the video.
Yeah, that's true for the incredible innovation in Neuralink and Paypal and OpenAI and Tesla, too. It's a total coincidence that Jeff Bezos and other billionaires fail.
On the other hand, if we shortly discover an asteroid on course to meet Earth within a concerning time frame or some other predictable disaster coming down the pike, they'll be a mass effort to take giant steps to ensure he and others are not wrong.
You're out of touch with reality. I'm not a fan of everything musk says but you're not clear about how ludicrous it is to catch a rocket with chopsticks. You're only going off your gut feeling from social media and popular opinion.
This is the reality. Musk made bets that have a lower chance of success then winning the lottery. If even one of the bets succeeds it means he's smarter than normal. That being said, Musk has succeeded in way more than one bet.
> You're out of touch with reality. I'm not a fan of everything musk says but you're not clear about how ludicrous it is to catch a rocket with chopsticks. You're only going off your gut feeling from social media and popular opinion.
Not sure what you're getting at here. I just praised SpaceX, but you're acting like I downplayed its success.
and then he launched a bunch of dumbass initiatives, and Peter Thiel power-played him out, and ditched "X".
his real victory was surviving long enough to make money from the merger and not get totally fleeced on the way out.
and given the number of Russians using Starlink, I suspect it's a matter of time before his clearance is revoked and the US FedGov backs away further from SpaceX. wouldn't be surprised if there are ITAR violations there, too...
Nothing that happens in the future can negate the fact that he promised autonomous driving would available in Telsas by now (actually long before now), and it isn't.
Tesla can never be the first to large-scale autonomous driving since Waymo is already delivering it. Sure, Tesla will probably deliver it eventually. Whether they can overtake Waymo's rollout is the open question.
> Nothing that happens in the future can negate the fact that he promised autonomous driving would available in Telsas by now (actually long before now), and it isn't.
Absolutely, he is well known for setting extremely optimistic stretch goal timelines, and is almost always late. In virtually every case he is still years or decades before the competition, with autonomous driving being the biggest miss to date.
He certainly does deliver though.
Given that Tesla are currently making 2 million cars per year and have 5 million on the road, I think it extremely likely they will have by far the biggest autonomous fleet.
Musk inspired and hired a lot of great engineers that made this all possible. His vision, passion and capital are crucial in this, but he dis not catch these himself, nor did he engineer any of the control systems to the detail needed to make this work, I'm sure.
I wonder how the mood is at SpaceX. I mean, those engineers see how he unsafe the production workers at Telsa have to work, they see how shit he's treating everyone at Twitter and basically sinking it, how long before someone the mood turns sour and we see an outflow to the competition?
The problem I see here is: what competition? There doesn't exactly seem to be any in the rocketry space. There's Boeing, but that place is a disaster right now. I guess ULA is still around, but they're not doing anything groundbreaking at all AFAICT, just using old rocket designs to launch some satellites. Where else can these people go?
Twitter people can go lots of places, both competing social networks, and also various other internet-related or software places. Tesla people can go to lots of different competing automakers. But I don't see anyplace nice for SpaceX people to go.
Plenty of of want-to-be competitors, but not so far with the amount of capital behind Musk. And space remains capital intensive, if not for hardware, then for people.
You say that, but at least half a dozen of these companies in the linked wikipedia list, all western, launched rockets last year. And several more are on track to launch within the next. Rocketlab is a serious competitor with multiple spaceports, a rapid launch cadence, and reusability and next-generation launcher projects in the works. Blue Origin is supplying engines to ULA and seems almost ready to test New Glen. Firefly had a successful launch. Stoke is working on a resuability concept that's even more exciting than starship, and advancing steadily through testing flight capable hardware. The list goes on.
Rocketlab and others even have products which standardize the bus part of the sattellite or probe, so you just attach your science experiment or sensor.
There's some new competitors, like Stoke, Relativity, Blue Origin, bunch more I'm forgetting.
This sort of diaspora of talent with experience is good to have, if there is capital
It's right around the corner. The evidence that Musk is actually really dumb is riiiiight around the corner. And will continue to be, for a very long time.
People at twitter seem to be very happy, major features like community notes are being shipped and having a huge impact on the platform and the company is much healthier after he removed the employees that weren’t doing anything.
His track record for being behind all these mindblowing innovations is off the record, but we don't like his face, so he doesn't deserve any credit for it!
The guy behind the article is bashing on Musk on constant basis. What a biased, one sided, bullshit article. You can also check his also rantings. I guess.. he wanted to work for Musk, but was never able to? Why the hate?
He even deleted my comments when trying to argue on some of his statements. Lol.
Yeah, it seemed pretty balanced and well thought out:
"Ploy for credibility, retreating even deeper into fantasy, pretense of reality, pure fantasy, contained within a world of fantasy, retreating into misinformation-fuelled fantasy world, runaway ego, less palatable, self-soothing fantasy bubble, fake statistics, blown deadlines, litany of bullshit, chronic inability, unprecedented publish delusion or deception, having drawn a poor hand, financial-cognitive nightmare, it’s hard to imagine anyone seriously believing that a night of delusional Disney Adult cringe, too childish to call a fraud, Musk’s increasingly-degenerate gambling run is slouching toward one last big coinflip"
Facts sprinkled with personal interpretation, a normal, albeit not an extraordinary, blog post?
I admire some of the things Elon Musk has made happen, but the guy has obvious ego issues and acts often like a villain from a child superhero movie. If a blog post is "biased, one sided, bullshit" for calling this behavior and all the accompanying unkept promises, what they are, together with some creative words and interpretation, why do we have blog posts in the first place?
It's not that bad. He's one of the rare people criticizing Musk while not exhibiting the delusion that Musk's predictions are "promises." So he avoids some of the more clueless mistakes although he still does manage to miss that the point of the event was just the unveiling of a product (the robo taxi), not its final delivery.
whoever wins the election, Musk is supplying NASA with spaceships for now and the immediate future and next probably will take over Boeing to make self flying personal airplanes
The ending spoilt the story for me.
Low quality election propaganda.
I am surprised its on the front page,
since thankfully election posts are
extremely rare. (Thank you for that guys)
Arguing the failure to deliver self driving,
even though it has been promised repeated
and greatly failed is fine.
The political rage rant at the end is not.
(This of course goes for any side in this election)
>"Ultimately, Musk’s increasingly-degenerate gambling run is slouching toward one >last big coinflip: the 2024 presidential election. With Musk going “all-in” on >Donald Trump, and musing that he will end up in a prison cell if Kamala Harris >is elected, it’s clear that his main political issue is his freedom to keep >rolling over his endless confidence game without legal consequences. If Trump >wins and delivers Musk the impunity he craves, the line between amusement park >fantasy and $700 billion self-driving juggernaut will all but disappear, and we >will all find ourselves living in Muskworld’s house of mirrors.
Doesn't read like a rage rant to me. Reads like a sober assessment of Musks behavior in regards to politics recently. He has gone full Trump fan and Kamala hater. He's become what he says he hates. He's become a demagogue.
If anyone is going to make self driving work, the odds are stacked in Tesla's favor. They have more data and distribution than any of their competitors. If/when that day comes, a decade—or two—of relentlessly pursuing that goal while making hyper optimistic predictions will be entirely irrelevant.
Root for the ambitious… The alternative is to ignore or tear them down. And it doesn't appear to be very fun to be in either of those camps.
I literally jumped for joy yesterday when SpaceX achieved a major milestone along their ambitious path. It's been fun following along and cheering for them.
I feel similarly excited for Tesla when they demonstrate robots with incredible dexterity. Even if they're not fully autonomous. They will still be a game changer for workplace safety. Imagine sending one up an electrical tower ladder in Alaska to replace a tranformer while an operator uses FPV goggles and gloves to control the robot from an office building in Fairbanks over a Starlink connection. Or a robot enabling a scientist to take ice cores in Antarctica from the warmth of McMurdo Station. This scenario would provide enormous quantities of training data as well.
The baseline AI-powered mobility and object avoidance subroutines are enough to make that viable while falling back to remote, human control for 'high resolution', less routine tasks. This same principle applies to robotaxis… And perhaps many Teslas already on the road if they can be retrofitted with the appropriate networking equipment.
> If anyone is going to make self driving work, the odds are stacked in Tesla's favor.
How disconnected from reality are you to make this statement? https://waymo.com/blog/2024/06/waymo-one-is-now-open-to-ever... Waymo self-driving has been open to general public since June 2024 and Telsa still doesn't have full self driving technology and yet here you are claiming odds are stacked in Tesla's favor.
> I feel similarly excited for Tesla when they demonstrate robots with incredible dexterity.
Did you never hear about Boston dynamics? https://www.youtube.com/watch?v=tF4DML7FIWk This is 3 years old video and Tesla's robots don't have as much dexterity as the robot from the video. And unlike Tesla's robot, Boston Dynamics robot is fully automated.
>> If anyone is going to make self driving work, the odds are stacked in Tesla's favor.
>
> How disconnected from reality are you to make this statement?
I suppose we all live in our own bubbles. I've not been to SF in many years. I've never seen a Waymo on the street. However, I've sat in a Tesla myself and watched it navigate the roads in my hometown. As have many others across the US. And so I feel rather connected with reality in that small, tangible way.
I should have been more precise with my words. By 'make self-driving work', I meant deploy that technology cheaply and widely. But retroactively I understand my clarification appears like I'm moving the goal posts. I alluded to that though: 'They have more data and distribution than any of their competitors.'
>> I feel similarly excited for Tesla when they demonstrate robots with incredible dexterity.
>
> Did you never hear about Boston dynamics?
I have, and I am also excited for them. I believe you missed my point about being happy for people trying to do difficult things.
Maybe. I guess another possibility is some people deliberately ignore evidence that doesn't fit their worldview for reasons that are yet unknown to me.
> However, I've sat in a Tesla myself and watched it navigate the roads in my hometown.
https://www.tesla.com/en_gb/support/autopilot
"Before using Autopilot, read your Owner's Manual for instructions and more safety information. While using Autopilot, it is your responsibility to stay alert, keep your hands on the steering wheel at all times and maintain control of your vehicle. "
This is not self driving.
> By 'make self-driving work', I meant deploy that technology cheaply and widely.
I see only one company deploying self-driving technology and that's not tesla.
I appreciate you sharing articles. Do you feel that we're trying to establish how well I see 'reality' or about Tesla's odds of deploying self-driving cheaply and widely. It feels like the former, but maybe I'm not being charitable enough in my reading of your posts.
Simply sharing the evidence that helped shape your own worldview does not necessarily invalidate mine. The first-hand evidence I shared supported my claim that Tesla has wide distribution, lots of data, and falls somewhere on the spectrum of self-driving (admittedly not fully autonomous).
The evidence you shared shows that Waymo has limited (but increasing) deployment, limited data (only the 3 metro areas), relatively little revenue, and—to its credit—fully autonomous driving capabilities under city-driving circumstances.
Waymo is taking a different development path and I'm rooting for them too.
But I'm still not convinced that the odds are in their favor. Can they build a relatively affordable car and especially their LIDAR system at scale? Will they be able to scale production before Tesla cracks the fully autonomous nut? I personally still think the odds are still in Tesla's favor. Note that I did not ever say: 'Tesla is the only one who can or will make self-driving work'.
> Root for the ambitious… The alternative is to ignore or tear them down.
Some resources are mostly zero-sum, support being one. That means that buying their product or even buying their stock means not supporting their competitors.
At this point, after so many hollow promises from Tesla, why shouldn't the money support plenty of other competitors as well?
I love my Model Y, and the supercharger network. The charger network in the US is so much better than any of the competitors in terms of reliability, why would I support the others? The other car manufacturers constantly roll back their hollow promises to electrification, while Tesla has been all-in since day 1. Why would I support Tesla's competitors? VW lied and lied about the emissions their cars put out, why would I ever buy a VW car? On and on... I'm happy with my choice of a Model Y, and I'm not compelled at this point to choose a competitor for our second or next car.
It was a difficult conversation when I had to explain to my children that my love and support for them was zero-sum.
I.e., I think you missed my point about being happy for people attempting to do something difficult and celebrating their successes. I can root for them as much as I please and it doesn't hurt a soul.
I'm afraid you're arguing against something I didn't say.
They certainly do. Those Mercedes don’t have all the cameras and sensors and are not recording it all and sending it back to HQ.
For the newer cars that do there was an awesome post on HN from someone who consulted with them and they had a single server at HQ with a single thread to capture all this data.
They had no idea why it wasn’t working.
I don't understand what's the message here, asides from that the author is bitter. Musk and his team have made incredible progress in many important directions, to the point that it feels the next industrial revolution is just around the corner. Let him go all in.
You really don't? 8 years ago Tesla did a (fake) video of their car "self driving" on real streets. 8 years of promises and in effort to convice their investors that they are actually making progress towards that goal is.....a movie set with fake streets to drive on? While Waymo is delivering real self driving taxi rides in the real world? What isn't clear about this message to you, exactly?
This, and that when you get involved in politics, a big percentage of people is going to scrutinize you more thoroughly. Before, it was fun to believe everything he said, but now it is less so.
I agree with the tesla stuff but last I checked waymo had self driving only for pre-mapped roads and a lot of human assistance in the background. There was a lot of fuss about it when veritasium released a sponsored video about them.
Do you think the data from the one doesn't apply toward the other? Is politics really enough to get people to stop believing in incremental problem solving? Can we communicate entirely in question form?
If the data from supervised driving can he used for unsupervised driving, that increases the depth of Tesla's failure to keep up with competitors like Waymo, who are doing a better job with less data.
I don't really know what you're trying to say with the politics distraction, so I guess the answer to your last question is "apparently you can't."
> If the data from supervised driving can he used for unsupervised driving
So you're unaware of the AI day presentations about their driving training pipeline. Learning about their engineering in that regard really changed my perception of the likelihood that they could solve the problem. Give it a watch: https://www.youtube.com/watch?v=ODSJsviD_SU
> failure to keep up with competitors like Waymo
Looks like each of them lead in different ways to me. Waymo leads within certain specific city limits, Tesla leads on the open road. These seem to be direct consequences of their engineering decisions. Question is which will converge on working everywhere all the time.
I think the issue here is that elon is one of many people who make their money not from building the next technology to underly an industrial revolution, but from making you feel like one is just around the corner. It’s an aesthetics vs functionality divide. Aside from spacex, which I’ll grant is really impressive, what novel technologies has musk helped develop that you feel will (or currently are) transforming the world in such a dramatic fashion?
There was already plenty of medical bci research, neuralink is just continuing that work with a talented group and lots of money from Musk. Neural interfaces are definitely a powerful idea and they would be a huge deal if made consumer-grade, but neuralink is far (decades probably) from making it sufficiently non-invasive to be viable for anything other than medical uses, and that prevents it from changing things dramatically. If Musk funded fusion research, I wouldn’t say he’s ushering in the next industrial revolution, I’d say he’s contributing money to a noble research effort.
Popularizing electric cars, while similarly noble, doesn’t qualify as revolutionary in my book. The move to electric is fundamentally just an increase in the energy efficiency of cars, which is great, but not a transformative leap. Furthermore, EV’s have been doing kinda bad here in the US as of late and automakers are having to scale back their rollout because of lacking demand. Maybe eventually we’ll be able to credit Musk with a big cut in car emissions, but in the best case scenario where we magically made the switch and powered every electric car with solar, we’d be cutting US emissions by a fifth.
Again, the parent comment compared Musk’s contributions to the industrial revolution, which saw a complete transformation in how humans do labor. It wasn’t a percentage gain in efficiency, it was a force-multiplier that spawned whole industries.
I disagree. I think that I’m measuring progress partially in terms of change, and that in order to feel like Musk was heading a revolution, I’d need to see the possibility for great change as a consequence of his actions. That change could be social/political, but I’m not seeing any. Right now the world I imagine coming about with Musk looks mostly identical to the world I imagine would have happened without Musk.
This guy has lost confidence in Musk's promises.
I never really understood how Musk can claim to be so close to full autonomy when these cars clearly struggle with basic things using the exact same system that supposedly drives by itself. Automatic high/low beam or detecting rain does not work very well. You wold think that a system that cannot detect oncoming traffic (high/low beam), or actual rain has some other problems when it comes to self-driving.
reply