Tesla Engineer Says Musk Overstated Tesla Autopilot Reality(bloombergquint.com) |
Tesla Engineer Says Musk Overstated Tesla Autopilot Reality(bloombergquint.com) |
I think when domain experts speak up like this we should listen. The autonomous vehicle technology seems like it has a long way to go.
Here’s how it looks to be in the front car of the train when there’s a supervisor: https://imgur.com/geFIidK
Outside of the rush hour, that lid is closed and it feels like the Half-Life train at beginning of the game.
I know it’s not possible to build trains to everywhere and USA is huge but I suspect that the answer for self driving would be somewhere in between of human+ level intelligence and dedicated infrastructure.
Unfortunately, most new transportation methods that are not trains are just attempts to build trains, while claiming they are not trains and will cost less than trains. But inevitably end up costing as much as trains.
Trains.
On opening, the line was equipped with a fixed-block Automatic Train Operation system (ATO). The train operator closed the train doors and pressed a pair of "start" buttons and, if the way ahead was clear, the ATO drives the train at a safe speed to the next station. At any point, the driver could switch to manual control if the ATO failed.[27] The system, which operated until 2012, made the Victoria line the world's first full-scale automatic railway.
- from https://en.wikipedia.org/wiki/Victoria_line#Service_and_roll...
https://migogkbh.dk/wp-content/uploads/2019/09/Metro-1200x72...
We are generally less optimistic about timing, but that was more true a couple years ago (before all the big cos' 2019 deadlines got pushed back) than it is now. Current timelines seem like an achievable challenge to me, and I don't think I'm unusually optimistic.
Musk's timeline pronouncements, as usual, just aren't taken seriously (and imo hurt the industry's credibility). I like the guy and respect what he's done, but for whatever reason he says really out-there things that are best ignored until he shows the receipts.
1. Human-driven cars are dangerous.
2. Autonomous driving (currently) is dangerous.
AD solves some human driving problems (like inattention) but introduces new ones (like driving under trucks).
Some of my more reserved comments get downvoted, and then flooded with replies about 360 vision and cars never getting sleepy.
IMO: Most of the autonomous driving tech will evolve to be driver assistance and safety tech in the upcoming decade. I still believe we'll get autonomous driving someday, but not until we have a few more AI breakthroughs.
Until then, "Autonomous driving" will be like "Duke Nukem Forever."
Skepticism of self-driving tech and "AI" in general is a very popular position here.
Here, anyone even just tangentially involved or know the some details of AI/ML are highly skeptical that Elon's vision of self driving cars will occur in our lifetime. Yet, Elon keeps pitching it, the press runs with it, and the public grabs it up, patiently waiting.
I'm less worried about the autonomous vehicle I might be driving and more about the ones others are driving. Here's an idea, how about being able to geofence a neighborhood as "<= Level 3 autonomy only". I'm only somewhat joking. I know it smells of NIMBY but it might be useful around schools, daycares, etc.
The Boring Company might be onto something. These tunnels are a very controlled environment, highly predictable. If they can be be made cheaper it might just be the initial "killer app" of self-driving cars.
The Boring company is essentially redesigning trains. With single-family train cars that are also very overpriced normal cars. That need to be substantially modified into very inefficient train cars to go down the tunnel.
I'm not trying to contrarian or anything but most of my life those have been my primary mode of transportation and it's worked out great especially if you throw a bike or now I guess a e-scooter into the mix
Yet, this fake it til you make model has enabled reflexivity in the growth and stock price allowing Tesla and other Musk ventures to grow and accomplish a lot with amazing speed.
An interesting study in Soros' theories. Turns out that lying about timelines, product capability (even existence), liquidity, buyouts, crypto pump and dumps, and defamation is heavily rewarded by the market. Particularly if your fame, past accomplishments, and world savior persona makes you immune to prosecution.
The problem with any non-vision based location system to compensate for covered roads is that it isn't what humans use. In winter in cold climates sometimes the roads are covered with snow for days to weeks and road margins drift for the entire season. Humans just form new emergent lanes. These lanes often aren't the "correct" lanes that non-vision based absolute position would find. And they're very difficult to detect with computer, or even human, vision.
Having one set of laws of autonomous cars in arizona and another for minnesota will lead to a lot of problems.
People are so poor at driving on ice/snow that it seems like a much lower bar to clear.
For self driving cars its very hard, roads might be a lot narrower and visibility is limited because of snow banks, also you need to know what part of the road you should drive, sometimes there can be deep grooves and you might get stuck. If you stop at traffic lights etc. it might be hard to get going and how does the computer know that this crossing is very slippery and it takes 3x more time to take off?
I'm thinking about those roads where it's nominally a two lane in each direction, but the snow banks make it a 1.5 lane in each direction (if that).
Where one doesn't start into the intersection when the light turns green - to give any cars that have the need to slide through. And likewise the ability for the robotaxi to realize "well, I've got no traction - guess I'm going to slide through this intersection and the left turn is impossible - I'm going straight through."
A robotaxi in LA, SF... ok. Pheonix - sure. But I've yet to see any examples of a self driving vehicle able to drive in winter condition roads in the midwest.
Elon is a marvelous marketer and salesman.
He's frequently MISIDENTIFIED as the new Edison or Tesla. He's NOT that - both of those men were insanely technical and enough so the jump into technical problems and add value to work being done by their employees. Musk is not that. He's had very little to do with any technical aspects of any of his companies. But there's been plenty of PR spin to imply the opposite. That spin has always been a lie.
I've had Silicon Valley bosses just like him. Our recent president is someone with the same personality just with fewer IQ points. But the same otherwise.
Disclaimer: I don't like the dude, he's an edgelord jerk.
Just a minor nit, "ADAS systems" is redundant.
Isn’t this the norm in any company? It’s always a struggle to keep a lid on sales promises!
>Full Self-Driving Hardware on All Cars
>All Tesla vehicles produced in our factory, including Model 3, have the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver.
How was it possible to know in advance what hardware will be required?
I suspect that most of the hardware is the ability of a computer to drive the car. Steering/throttle/braking inputs for example.
It could be that what you mean by 'hardware' is some combination of intelligence and sensor system.
For example, on the highway, truck infront of you drops some debris.
What do you do? It can be dangerous to suddenly swerve or brake, but also dangerous to hit something.
What you do In real life is quickly evaluate. You are looking at the object that fell and estimating how deformable it is. For example, if it looks like a wad of paper, you would continue on. Or you estimate if you have clearance based o how big it looks and your mental model of your car’s clearance.
You also look at the traffic beside you and behind you to evaluate risk of braking or swerving.
You are also doing object recognition. If it is an empty box, you may be ok hitting it, but if it is an infant car seat that fell, you will brake hard/ swerve even if it is dangerous.
Also, for people who bring up the safety of human drivers, one thing to consider is the safety of a hybrid approach where the human is still in charge but you have safety features like lane keeping, blind spot monitoring, automatic emergency braking, etc. That hybrid approach may be actually safer than either only human, or only computer driving.
Book suitable even for lay people, written by someone who has been working for years in this field:
[1] Artificial Intelligence: A Guide for Thinking Humans by Melanie Mitchell
People to follow:
[2] https://twitter.com/missy_cummings/
[3] https://twitter.com/rodneyabrooks/
[4] https://www.twitter.com/MelMitchell1/
Edit: For criticism of Elon Musk and/or Tesla:
[5] https://www.elonmusk.today/
[6] https://en.wikipedia.org/wiki/Criticism_of_Tesla
[7] https://en.wikipedia.org/wiki/List_of_lawsuits_involving_Tes....
[8] https://www.currentaffairs.org/2021/04/surely-we-can-do-bett...
[9] https://troll.tv/videos/watch/54bc7bd0-8691-4359-aa7d-dc5148...
I don't know if 10 is the right multiplier here, but the point is the difficulty is exponential.
Musk's engineers know this but for whatever reason he can't or won't understand it.
I think that joke should have been reserved for Saturday night live.
"“Elon’s tweet does not match engineering reality per CJ,” Miguel Acosta, chief of the California DMV’s autonomous vehicles branch, wrote in the memo"
So someone in the CA BMV claims a Tesla engineer (CJ) said this, but doesn't quote them, and the article doesn't elaborate on it.
Maybe it's true but they're clearly reaching a bit with the title.
However in today's world there isn't a good option. Even if we would put in arbitrary subways, there are still jobs [like plumbers[ that cannot reasonably use transit. Thus self driving cars must be a part of my better future.
1) sleeping/reading/etc whilst driving long-distance 2) getting home from the city after drinks 3) pick up kids from school/practice (this would be a major time saver)
People are looking for something to believe in, which will make them feel like the future will be brighter.
Musk has this ability . Unless he goes crazy and becomes overleveraged with debt he'll be fine.
He'll constantly move the goalposts and keep giving people hope, if Autopilot is proven not feasible, he'll just move onto the next big thing which will fill people with enthusiasm .
He's essentially a politician, a techno-utopian cult leader who is also a CEO and major owner of the stock.
His best intuition was understanding that telling people that the future will be brighter is itself a product which people are willing to buy. Much like people pay for insurance to have peace of mind, people gladly pay to be reassured that the world will be better.
His second best intuition is to never bankrupt the company trying to deliver, just move the goalposts. As long as you don't bankrupt the company you can always move the goalposts
Apple -- best iPhone ever so what happens to phone I purchased last year (reality - there a decent upgrades)
IBM Watson -- commercials looks like you can layoff 20% of employees as this things can work like wonders (reality - its not that smart)
Tesla -- Autopilot is not 100% auto pilot human supervision is needed at times. We are far away from 100% autonomous vehicles. I wish they selected a better name for this feature.
I joke, but seriously, how long can this Musk worship go on? Musk invested in Tesla AFTER Martin Eberhard and Marc Tarpenning built the Roadster. He later sued them for founder status and ousted the real founders from their own company. It's now almost 20 years later, and Tesla is still a tiny little car company that has not had one profitable year (a fact which actually benefits Musk right now as the share price can not be anchored to anything real... the possibilities are infinite ;) The one company Musk actually founded; SolarCity, was an epic failure which needed to be saved from bankruptcy by Tesla and SpaceX. I honestly suspect that Tesla might have been a bigger success, had it remained with it's original founders.
By the way, there are at least 11 lawsuits and investigations of SolarCity listed on it's wikipedia page. That's quite a rap sheet for such a young company with such noble goals. https://en.wikipedia.org/wiki/SolarCity#Lawsuits_and_investi...
It's as if Apple made the best computers--which some people think they still do--but Tim Cook was tweeting every day about how the MacBook Pro can solve the halting problem. No it can't; of course it can't. Why would you continue to drag down your company's good name by lying like this?
What is more interesting is why he hasn't learned that his worldview is so fundamentally wrong. And for that, I would say that it's hard for very rich and successful people to acknowledge they are so completely wrong about important things.
[1] https://www.thedrive.com/news/34144/the-tesla-model-y-is-alr...
[2] https://www.thedrive.com/tech/29200/customers-revolt-as-fix-...
But once you are a company that size the goal is to grow and accomplish amazing stuff but with great margins.
Tesla has essentially arrived at destination with regards to MarketCap, it won't become a 5 trillion dollar company. Nobody will. There is such a thing as the Sherman Act and Antitrust.
Tesla still lags behind in every other metric (especially margins) compared to similar companies of that size.
Seems to me Musk will have to work a whole lot just deliver and thus mantain the current valuation and his level of net worth right now.
Alternatively he could get to 5 trillions by promising new products at an increasingly higher rate and thus avoiding the antitrust laws. Those laws were written for a company which actually has a market dominance, not a company which has a stock market dominance based on the promises of the CEO and people believing in it and buying the stock.
I think from what we've seen, he's going in this direction exactly, he'll keep inflating the bubble that's for sure. He only has one gear
Tesla has shown that it is taking increasing steps to ensure people are using fsd responsibly. while autopilot used used simplistic sensors to gauge operator attentiveness, fsd uses cameras to watch to driver. as elon has said, with autopilot, the company has data that proves that it reduces the risk of accident. he has said there is a moral argument that it is immoral to wait until it is "complete" before making it widely available. this evidence is regularly ignored by people talking about the dangers of Tesla/elon automation claims.
Object recognition has been a part of autonomous driving systems for a very long time. The difference between a stroller and a piece of paper is pretty well understood by these systems.
> For example, on the highway, truck infront of you drops some debris. > What do you do?
The answer is to always keep a large safety margin between you and the vehicle in front of you when traveling at high speeds, so you have time to react and take the safest course of action. Seconds make a huge difference in this scenario and it's easy to get them. No autonomous system is going to be hugging a truck bumper, but rather is going to always maintain a significantly greater safety distance than an aggressive human driver would.
> If it is an empty box, you may be ok hitting it,
You can never be sure that the box is empty, so I'd recommend always swerving to avoid.
1) Hit the object 2) Swerve 3) Brake
Hitting the object is almost always a bad idea. On a highway the best option is almost always 2. On an empty road, the best option is almost always 3.
Saying things like "You also look at the traffic beside you and behind you to evaluate risk of braking or swerving" implies that humans are good at this and computers are bad at this. I'd wager humans are bad at this and computers are good at this. And in either case, situations like these represent a small fraction of accidents on the road. If the computer arbitrarily chose to recklessly swerve to the nearest lane every time it would still likely outperform human drivers in terms of safety if it solves the low hanging fruit elsewhere.
Somehow people manage without this.
Without any real knowledge of the problems the engineers are hitting, my current take is that to produce a reasonably good autonomous driving system requires killing some people. Slowly adding features will send more information to the mothership and they'll have to push the envelope to learn enough.
> Musk took an active role within the company and oversaw Roadster product design at a detailed level, but was not deeply involved in day-to-day business operations.[12] Eberhard acknowledged that Musk was the person who insisted from the beginning on a carbon-fiber-reinforced polymer body and that Musk led design of components ranging from the power electronics module to the headlamps and other styling.
https://en.wikipedia.org/wiki/History_of_Tesla,_Inc.
Do you have a source on your company founding point?
https://en.wikipedia.org/wiki/Zip2 https://en.wikipedia.org/wiki/X.com
I don’t have time to keep going through this, but even the tone of your whole post is extremely smarmy/troll-like. Usually discourse here is a little more polite.
https://meaww.com/elon-musk-tesla-cofounder-lawsuit-settleme...
Isn't this true of most growing companies? Like Amazon, Uber, YouTube (now Google), etc.
I believe Amazon finally has. I don't know about YouTube, but I suspect they are. Uber has a bad business model and so won't (unless their raise prices)
If you mean driving like a human does with no priors over the area and in nearly any circumstance, I would agree.
> why not just build more advanced subways/underground trains?
The train business is horrible to be in. Just look at Bombardier Transportation/Alstom. City A in State B and country C decides it needs a new train. The existing infrastructure was all custom-made in the 60's. Bombardier/Alstom has to design a mostly custom train for it. Everyone bike shed on the design. Then, because it's a government project, and State B financed some of it there's a clause in the contract that is must be manufactured in State B of Country C. Not only is a new production line (manufacturing hell) for the almost completely custom design required, they must also set shop in this new jurisdiction and work their supply chain around that. Ribbon cutting ceremony, positive polls for reelection.
Few years later, City D in State E of Country Z wants a new train system. The whole process starts all over again. Can't just retool the existing plant in State B of Country C, it has to be a new one too in State E of Country Z, or a local manufacturing partner (ribbon cutting ceremony obliges). Meanwhile the press in State B of Country C gets angry as workers are laid off since there's no more work to be done.
By using standard road vehicles, Musk eliminates one of the biggest cost. It's all off the shelf stuff that's user-supplied.
I’m sure millions of people had the idea of building an electric car before Eberhardt and Tarpenning, but as the quote goes “ideas without execution don't matter; execution matters.”
Almost literally every decision trades off one danger for another: to pick an adjacent example, airbags introduce dangers that are different from the ones they solve. And yet, "airbags have dangers! Take them out of your car" is a woefully incomplete understanding of the risk landscape, driven more by our Neanderthal brain's superstition than our Homo Sapiens reasoning about risk.
That sort of magical thinking is a hurdle that every shift in the risk landscape needs to overcome ("what if vaccines have microchips in them??").
This is, of course, dependent on actual comparative safety stats, but I've yet to hear a safety argument against taking a Waymo ride over an Uber in Chandler[1]. If the same level of transparency and rigor is used in more complex environments, I'd take a cruise/waymo car in sf too (once they launch).
[1] https://www.forbes.com/sites/bradtempleton/2020/10/30/waymo-...
As engineers we can pretty much completely characterize the airbag space and make predictions about where airbags are dangerous, and then mitigate those. What makes AD a bad idea is that such characterization is impossible with the AD space because AD is currently done almost exclusively by feeding massive amounts of data into a neural net pattern matcher. In other words we have no idea why AD works, and thus we have no idea when AD won't work until somebody dies. This is no way to do engineering, especially when lives are at stake.
[0]: I don't think this is good, it's just a statement of fact.
There's always "something" that technologist think is right around the corner, until they realize that the hard problems are harder than they anticipated.
Deep learning is great for a few narrow cases but there is no path to general human-like intelligence required for a complex human-like activity like self-driving which has an infinitely long tail of edge cases you can't train for with these large and "dumb" ML models equivalent to curve fitting.
The mistake of Tesla is betting on covering more and more of these edge cases incrementally but what's needed is qualitative change rather than incremental improvement.
By qualitative change I mean actual model of human-like intelligence, including reasoning and "common sense", which is a prerequisite for self-driving. The solution to this problem requires more than the sum of its parts.
It's kind of like planning to build an airplane before figuring out Newton' basic laws of physics.
I'm not sure that is possible on roads.
The fact that it tends to be in businesses that are among the most mature and hardest to get into is really amazing.
All of Elon's companies are heavily subsidized by US tax payers (energy credits for Tesla and solar, rockets are paid for by NASA, military, etc...) So to use tax payer money you need to have the goodwill of the tax payers behind you. Its worth every penny to lie and bolster your image of doing cool shit so people don't mind their tax dollars continuing to go towards your cause. It's 100% PR. Without US tax payers none of these companies would still exist.
Before launching in any city, you have to validate that you can handle the idiosyncrasies of its driving environment. Perhaps the business side decides that Detroit isn't worth launching in, but <other snowy city> is drivable 200 days/yr and that's enough to warrant launching an intermittently-available service.
My ultimate point is that AVs can cover a lot of ground from a business perspective without getting to full universal-availability. This is helped along by a combination of validating each city independently[1], limiting operating conditions, and safe-fallback teleoperation for occasional use.
Validating a 2021 AV in a snowy environment isn't an inherently harder problem than validating a 2016 AV in a normal environment (with a safety driver), and a similar approach will be applicable.
[1] Presumably less and less narrowly each time
> I'm not sure a want a human driver who only drives in winter - they will be out of practice when we need them
Do you feel similarly about making cities bike-friendly?
If not, the robotaxi could drive itself to Florida and leave any customers who relied on it up north in a lurch, and also mess up the traffic and transportation economy of Florida in the winter. Screwing two economies with one car = bonus.
I’ll say that’s a rather naive view. It sounds like saying Mark Zuckerberg is not lying about FB and privacy, it is just his worldview that nobody should have privacy, except of course him and his family.
Why isn't he working on stuff in a full open source manner like Linus of Linux did back in the day?
Also , let's assume you are in it to solve the problem and believe that the for-profit model is the better than opensource ...then why not live modestly like Warren Buffett?
Guy knows what he's doing. He wants to be famous and rich and surronded by yes-men and yes-girls.
Governments already claim a monopoly on roads in most places. Imagine if those roads had some sort of cheap embedded track that a number of vehicles could latch onto. Then you just need some sort of computer to coordinate the vehicles using the track, and understanding where the vehicles are on the track.
I'd like to see the numbers on that. Given the cost for the CHSR is around $100B, I don't think that statement is true. The bigger issue here is the US's inability to build fast trains, affordably, and on a reasonable timeline.
I think a very conservative guess is that there's been $40B invested in this space. That would, for example, cover the south bay BART expansion 4-5 times over.
I've become more cynical about CA (and the US in general) lately. By reasonable costs it would, but I think the various corruption interests would find their way to increase the price tag until it is only half done, while they enjoy their $40 billion.
Also there are considerations other than time, it's much more convenient that driving or flying. I know in USA there's last mile problem going out of railway station but that can be solved by car trains.
