Slashdot Mirror
Tesla Will Open Its Security Code To Other Car Manufacturers (engadget.com)
Tesla CEO Elon Musk announced he would share the source code for Tesla's car security software with other manufacturers, adding that it would be "extremely important" to ensure the safety of future self-driving cars. Engadget reports: Musk didn't provide a timeline for availability, and you might not want to get your hopes up when it took years for Tesla just to post any source code. And this isn't strictly a selfless gesture. If rival brands adopt Tesla's approach, it could set an unofficial standard for connected car security that would look good from a marketing standpoint. The code could provide a boost to connected car security if and when it arrives. There are few common frameworks (technical or legal) for safeguarding networked vehicles, and security might not always be a top priority. This could give companies a baseline level of security that would save brands the trouble of developing an effective defense from scratch.
I can't see many if any other automakers taking in code Tesla has produced, pretty sure most would prefer to go their own way... I am also pretty dubious if they did, that it would provide much marketing benefit to Tesla.
Even from a pure technical standpoint I wonder how much use you could get from what Tesla is offering when each company would have pretty different approaches to self driving cars.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
"give companies a baseline level of security that would save brands the trouble of developing an effective defense from scratch." - Or make a monoculture exploitable by kit that most mfctr's will never fix in existing models?
Toyota has hundreds of security engineers and produces about 9 million cars per year. That's about 11% of total auto production.
Tesla produces about 100,000 vehicles, or 1% of what Toyota produces. It's going to be tough for a tiny boutique operation to even *influence* the standards uses by the auto industry, much less *set* the standards. It's like me and my company trying to set the standard for HTTP authentication. Our comments were read by some of the significant players who have some actual say in what standards are adopted, but organizations such as Netscape, Microsoft, and Apache designed and selected the standards.
A more realistic goal for Tesla might be to have a voice so that standards adopted by the industry don't completely screw them. Since Tesla is 0.11% of the auto industry, statistical noise, the industry as a whole doesn't care what they do. VW, which produces a hundred times as many vehicles, actually affects the industry.
has something of value to contribute to car "security", given the number of exploits that even allow some models to be driven around remotely.
Fuck off, this is legitimate technology news.
This is all about gaining a monopoly.
And, having seen SpaceX software design "processes", I don't trust any Musk enterprise software for life safety critical applications. There's a reason he had to buy his space rating from congress.
No, it isn't. A legitimate technology news is "Tesla has opensourced its security code and it is available at this URL, along with the license and documentation". "Tesla may or may not share something with someone" is just paid shilling and musk ballsucking.
Unlike the funding
It's an Elon Musk tweet. He promises all kinds of shit that never happens. 100% solar powered superchargers, self driving cars in 2017, built in dashcams... The list is quite extensive.
It's not news.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
Toyota has hundreds of security engineers and produces about 9 million cars per year.
Security concerns around self driving cars are many orders of magnitude more critical than anything traditional auto makers have to deal with.
What is the most complex thing todays auto makers have to worry about? Pretty much it's wired input for gas and brake, that's about it... oh maybe something as important as On-Star's ability to stop a car running.
All of that stuff is nothing compared to a self-driving car that is not just directing speed but also direction.
I wrote elsewhere I'm not sure what Tesla has to offer will be taken in by other auto makers and probably wouldn't be of use to them. However I do think Tesla is vastly ahead of other car makers in terms of practical self-driving car software, both because the not-quite-self-driving car Autopilot feature has been in production for some time, and because of the vast quantity of sensor data Tesla has around to process and try new systems against.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
You mean the code that was hacked each year so that brakes could be activated remotely? That code base? Yeah - I don't see other manufacturers really giving $0.02 about it...
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
To be fair, a few slashdotters with a month of time and existing FOSS tech could probably create a more secure car than the industry as a whole.
This is my signature. There are many like it, but this one is mine.
Well, they have to do SOMETHING to deflect from Musk's blantant SEC manipulation and upcoming fraud charges...
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
Don't forget a $35,000 model 3!
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
Yes OnStar cars have been similarly hacked, despite GM producing many millions of cars....
I'm not saying Tesla's security is perfect either, I am saying:
A) How much more important security is with a self-driving car where a hacker could literally drive a car, and
B) Tesla has had a much longer time with cars in production to think about this.
It's worth noting that the hack YOU linked to involved going in through the web browser that was actually IN THE CAR ITSELF. So you'd have to direct someone to visit a web site using the in-car browser.
The OnStar hacks have been purely remote, initialed from afar. And the hack you linked to did not affect the self-driving features, just the brakes.
Tesla is clearly being very careful and doing something right when they deliver car updates over the air and yet have not had any serious real-world exploits.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
The trolls are at it hardcore lately, huh?
Good job to Tesla for announcing that they're planning on open sourcing their code! This can't be anything but good news for the auto industry. There's a lot of people that are worried about their autonomous cars getting hacked, this will provide a good baseline of security.
It is pitch black. You are likely to be eaten by a grue.
It's going to be tough for a tiny boutique operation to even *influence* the standards uses by the auto industry...
"Going to be??" That's funny.
Well, they have logged over 5 million miles of fully autonomous driving on public roads. Oh wait, that's Waymo.
> the vast quantity of sensor data Tesla has around to process and try new systems against.
You're kidding about that last part, right? BMW, and several other auto companies, get more sensor data in a week than Tesla does in a year, simply because they've produced and sold millions of vehicles with driver assist. Tesla does have a few advantages, such as a very charismatic CEO. There huge weakness is when you talk. About a "vast quantity" of - anything. Just using BMW as one example, millions of BMW cars are on the road with systems including Active Lane Keeping Assistant, Frontal Collision Warning with City Collision Mitigation, and Active Cruise Control with Stop & Go. Being a small, boutique player has its advantages, but the only thing vast about Tesla is its CEOs imagination and ego.
Well, they have logged over 5 million miles of fully autonomous driving on public roads. Oh wait, that's Waymo.
Is Waymo Toyota? Or any other traditional car maker?
OH! Thanks for re-enforcing my point. I didn't say Tesla was the ONLY one ahead of the large car makers... I just said the car makers producing millions of cars now are way behind.
BMW, and several other auto companies, get more sensor data in a week
That is utter nonsense, you appear to be equating extremely primitive sensors like sonic devices and lane recognizers to a full 360 multi-camera camera rig or LIDAR (Tesla doesn't use LIDAR but Waymo does). Come on man. Sensor data from sensors that are not going to be the ones eventually used for self driving tech is vastly less valuable than data that is from the actual hardware that will be running the system.
There is a world of difference between lane assist (which incidentally can often go wrong, have you been in a car that has this???) and a car truly steering itself.
This also brings up another point, it doesn't matter how many millions of BMW cars have Lane Assist on when they have no way to get data back to BWM. All of that data is lost, of no use to BMW in development. Tesla has the ability to retrieve a lot of data off car hard drives during service or even to stream key factors back on the fly in a way almost all of the many millions of cars other car makers produce, cannot.
"There is more worth loving than we have strength to love." - Brian Jay Stanley
Which idiot thinks it's a good idea that cars are always online and send all kinds of data back to their manufacturer and ready to query for all TLA's, reducing privacy to zero.
> Tesla has had a much longer time with cars in production to think about this
What?!? Tesla is just now this year and last starting to mass produce cars, at the rate of about 100,000/year. That's the same rate of production that Fors had in 1913, when. They produced 100,000 of the model A.
Mercedes and Freightliner have had various types of "autopilot" (driver assist) since 2000-2002, about the time Elon Musk was hired (and quickly fired) by PayPal.
Tesla is a new company, and a small company, led by someone with big dreams. Who has "a longer time to think about this" is very much to Tesla's disadvantage, since other companies had working systems on the road in production vehicles before Tesla even starting thinking about trying to develop something.
The problem with the CEOs big dreams is he can't pick one - Tesla, SpaceX, the Hyperloop, Solar City, the Boring company - what does Elon Musk want to do when he grows up? He can't decide.
Perhaps so. There are a few of us who understand security. There are many on Slashdot who think they do, and steadfastly refuse to pay any attention to those of us who have been doing security as a lifelong career.
Not knowing is okay - smart people can learn. Refusing to learn because you think you know it all ensures failure.
Mod parent up. Enough is enough of Enron Musk's constant and meaningless attention whoring and his childish attempts to divert the spotlight from his incompetent and fraudulent "business" tricks.
GTFO, empty homophobic racist suit and tie.
We really should make one of those Hitler clips with Slashdot's Tesla short seller's tears.
How’s that short treating ya lyn?
Other manufacturers are not behind, they are just coming at it from a different direction.
Nissan has an advanced self diving programme, for example. It works very well, but they are waiting for the necessary sensors to become small and cheap enough for consumer use.
Tesla decided that they could do self diving with minimal sensors. They mostly rely on cameras, cameras that don't even have self cleaning capability or HDR and which get blinded by sunlight and tunnels. Their hope is that they can train a neural network to drive with just those. So they need to collect a huge amount of training data and do a massive account of testing and evolution.
And it's not working; they just announced a new hardware revision that has to be retrofitted to existing cars.
Just because other manufacturers aren't doing a public beta with their customer's cars doesn't mean they aren't keeping up with or exceeding Tesla.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
If they build "secure" code like they build models 3s, there will be huge panel gaps everywhere and the bumper will fall off when it rains.
Yes Tedla has flaws, but I think open sourcing something this important helps the industry as a whole - from creating standards, to government regulation, to potentially even having the development community at large participate - although a consortium of manufacturers agreeing on a code base that is actually secure would be more reassuring
Toyota has hundreds of security engineers and produces about 9 million cars per year. That's about 11% of total auto production.
The most common cars are the most commonly stolen cars, because sales of their parts once scrapped are easily disguised among all the other parts sales.
But if you want to tell us that Toyota is serious about software, you're going to get your face laughed in, because what came out when they were accused of sudden acceleration is that they not only don't follow industry best practices, they don't even follow their own [inferior] internal practices. Multiple code paths which could cause the sudden acceleration problem were identified within their code as a result.
A more realistic goal for Tesla might be to have a voice so that standards adopted by the industry don't completely screw them. Since Tesla is 0.11% of the auto industry, statistical noise, the industry as a whole doesn't care what they do.
Bollocks and garbage. The whole automotive industry is watching Tesla very closely, and emulating them in any particulars they can manage. It's now becoming popular to put large portrait-orientation LCDs into higher-end models, and all major automakers with two cents to rub together (i.e. everyone but Mazda and FCA) is bringing out at least one better-than-compliance EV.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
In a few years, TSLA will have a mult-trillion dollar market cap, bigger than all the other car and energy companies combined. Do you even land a rocket on a boat, bro?
Copyright (c) 1990 - 2014 Dice. All rights reserved. Use of this comment is subject to certain Terms and Conditions.
> all major automakers with two cents to rub together (i.e. everyone but Mazda and FCA) is bringing out at least one better-than-compliance EV.
So they're emulating BYD, the largest maker of electric cars? If they promised super-expensive cars, took deposits from customers, and didn't deliver cars for years, that would be following Tesla. BYD actually mass produces electric cars, in higher quanities than Ford produced the Model A in 1913.
Tesla has the production of 1913 Ford, before the model T, with the hype of PT Barnum. Like Barnum's endeavors, Tesla is a show, which may eventually become a car company in 50 or 80 years.
That sounds like another great reason to never get into a self-driving car. Even if someone you trust is at the wheel the software can easily remove them from gaining/retaining control of the car.
"Opening security code to other car manufacturers" (as per the language used in the /. headline) simultaneously exposes a problem with the open source development methodology (it's not about software freedom, users included) and points to a conflict self-driving car manufacturers and distributors aren't terribly interested to get you to consider—various authorities (police, FBI, etc.) want you in a vehicle they can more easily lock you into and commandeer.
Software freedom (the freedom to run, inspect, modify, and share published computer software) would work against outside controlling interests by giving you the keys (pun intended) to control your car to the limits of your expertise and willingness to learn or hire someone to work on your behalf. You could choose to let some authority control your car or you could edit out the code which gives anyone else that power and retain control of your car. After all, any free software activist will tell you it's your car therefore it's your computer in that car. We've seen what proprietary software interests think of your health and safety. I don't think we need to wait for more evidence that they don't have our interests in mind. We all deserve software freedom for all of our computers.
Digital Citizen
No short here! I'm smart enough to not short such a volatile stock with a crazed CEO at the helm who engages in stock manipulation. Curious about when a $35K Tesla will finally ship, though...
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
BYD actually has a headquarters in Los Angeles and a factory in Lancaster, California. Rather than selling one car at a time (and maybe deliver it five years later), BYD see vehicles 100 at a time, providing fleets of taxis and buses. If you've ridden the bus in New York or Chicago you were probably riding in a BYD electric vehicle.
How can any vehicle be considered 'secure' when there'll always be a backdoor for the government to seize control of your vehicle remotely, not allowing you any say in the matter? Any so-called 'security software' will be a joke, because you know that criminal hackers will be able to break into that backdoor and do whatever they want. This is why there should be a hardware switch that allows the owner of the vehicle to disable any radio transceivers that could be used to control the vehicle remotely or otherwise influence it's operation, including any 'updates' to it's firmware/software.
How’s that short treating ya lyn?
Hard to short something you weren't dumb enough to buy in the first place. Only one of us in a position to short is you.
I just said the car makers producing millions of cars now are way behind.
Maybe those other manufacturers are doing it right - focusing on actually building cars at a profit, and leaving specialty things like auto-pilots to specialized companies. It doesn't make sense to design your own fuel injectors, tires, or starters - let others do that. Likewise, let someone who is specialized in auto-pilot technology spearhead it and add it to your cars when it's ready. In the mean time, keep pumping out millions of cars that actually turn a profit unlike anything shipped by Tesla.
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
And Linux on the desktop will completely take over. By 2002 there will be almost no people left running Windows 98. Microsoft will go out of business.
Yeah,
In a teardown if the Model3 it was noted how well integrated the systems in the car were, well ahead of what the big automakers were doing. It was started that this level of integration would allow for a level of profitability you just can't get from off the shelf systems. Tesla is clearly looking at the long term as opposed to what happens tomorrow. The teardown also pointed out how they have a way to go with there assembly techniques and structures, things the big automakers have mattered. But despite this they have the most profitable EV noon the market. Even GM can't say that.
Nissan has an advanced self diving programme, for example.
Tell me more about this amphibious, submersible Nissan. :)
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I just said the car makers producing millions of cars now are way behind.
Maybe those other manufacturers are doing it right - focusing on actually building cars at a profit, and leaving specialty things like auto-pilots to specialized companies.
Nope. They are all buying companies working on self-driving systems so that they can control the whole. They disagree with you on what "right" means. Except, of course, for the smallest automakers, or the ones rapidly circling the bowl like FCA.
It doesn't make sense to design your own fuel injectors, tires, or starters - let others do that.
GM used to build their own fuel injectors for heavy vehicles, back when they owned Detroit Diesel. The old DD injectors were pretty cool, too... injection pressure was produced in the injector itself. That meant less pressure in the fuel pump, and the ability to use a common rail fuel system. This went back to the 1940s. Even then, though, everyone else was buying them from someone else, so yeah.
On the other hand, Tesla has the best EV motor in the industry, literally. And they have it because they designed it... They also have the best battery packs in the industry... ditto.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Curious about when a $35K Tesla will finally ship, though...
When Tesla stops selling more expensive models as rapidly as they can produce them, of course. It's a business, not just a game. And they're collecting $10k in profit from each sale right now, they aren't going to give that up any sooner than necessary.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Rei, do you think everyone in the world is shorting Tesla?
There isn't enough stock to for that.
Now wipe your chin and go back to the box, Gimp.
And they're collecting $10k in profit from each sale right now,
So much profiting...
How come their losses increase, then, year after year?
There are no plans to "opensource" anything. What the attention whore tweeted about is and idea for a plan to license the code to some automakers. You know, kind of like Oracle and Microsoft license parts of their products, including source, to third parties.
It ain't going to github for you to inspect.
Yelon Mask and his friends Larry Page and Marc Suckerberg.
You won't believe it, but it is available for sale and is delivered on schedule and without "paint chips". Also, it is also very polite, doesn't call heroes who save children "pedos".
No, in a video it was noted how integrated its computer motherboard was. By a guy who obviously knows a lot about body work, but very little about electronics.
would allow for a level of profitability you just can't get from off the shelf systems
Yeah, a negative level of profitability. You can't get this off the shelf, really :)
The "sudden acceleration problem" was a made-up story by the US to help domestic producers, just like the Audi "pedal" issue. The "issue" has not manifested itself anywhere else in the world.
By coincidence, the US is the only country of the 50 or so I've been to, where I've seen people driving with the right foot on the gas pedal and the left on the brake.
I think most of us were pretty certain a long time ago that it would require a faster computer than what they were using. That said, if they're really able to get their machine learning model to process 2,000 frames per second with the new hardware, up from 200, they went from limping along at 25 fps per camera to ostensibly processing 250 fps per camera, which finally exceeds the actual frame rate of the camera hardware (60 fps).
I wouldn't call that "not working". I would call that grossly underestimating the compute power required for self-driving. Chances are, they'd have encountered the same problem if they had used LIDAR like Waymo. It isn't a panacea; it's just another data source.
Check out my sci-fi/humor trilogy at PatriotsBooks.
While I agree that the rest of the industry is unlikely to use their code I'd like to point out that your big numbers does not translate to more secure systems. Modern cars have more computers than ever and it's a pretty common opinion by experts that the industry have integrated all these systems with little to no concern for security, just like many airplanes, sadly.
But hey maybe these hundreds of "security engineers" at Toyota are actually creating a new secure by design car computing system. It's about time!
Funny, Iâ(TM)ve charged at several 100% solar powered Supercharger stations. I guess you donâ(TM)t know shit.
I don't think they will be able to do it with cameras and neural nets at all.
Maybe they will be able to get it to drive along a highway with that system. But they promised it would park itself too. They are going to have to teach that neural network to handle every different kind of car park, navigate around them looking for spaces. Different road surfaces, different lighting, different layout of spaces, and no map data or GPS for assistance.
The basic problem is that they are hoping they can teach the neural net to drive just based on what it can see, where as every successful self driving project has worked on the basis of building up a 3D model of the world around the car and navigating through it. Lidar is immensely helpful for doing that. In theory a camera might work, but getting a neural network to recognize geometry and spit it out in a format that can be combined into a 3D model is science fiction level strong AI.
The best part is that they activate the autopilot computer in your car even if you don't use autopilot or didn't even pay for it. It wastes your electricity, your energy to help train their neural nets.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
Well Toyota make a "Tank", although I'm not sure if it is actually watertight :)
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
I think they might drop that model, or at least make it so unattractive that no-one buys it.
Hyundai and Kia beat them to the affordable long range EV with models under the $35k mark and similar range to the much more expensive Model 3 Long Range.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
When Tesla stops selling more expensive models as rapidly as they can produce them, of course.
The thing is, the work into which Tesla is actually, is "building manufacturing capacity as rapidly as they can".
The fact that they can already ship cars (relatively) rapidly out of those currently built manufacturing plants is just added bonus.
And thus yeah, you're entirely right :
And they're collecting $10k in profit from each sale right now, they aren't going to give that up any sooner than necessary.
They indeed need all that sweet money to pour into their "manufacturing capacity" building. For now they are not pushing that many cars out thus :
- they need more manufacturing plants before being able to push them out faster
- these manufacturing plants are expensive to build
- thus while they are still pushing small numbers of cars, they need to push the more expensive ones with bigger margins.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
And they're collecting $10k in profit from each sale right now,
How come their losses increase, then, year after year?
Yes, you know, there's a big lie in the car industry (all manufactrurers including Tesla) :
In reality, cars actually grow on trees~ For real~~~
Thus there's no way that Tesla needs to invest any money they make plus any investors' into building more plants to increase their manufacturing capability.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Even from a pure technical standpoint I wonder how much use you could get from what Tesla is offering when each company would have pretty different approaches to self driving cars.
You could at least get a lot of idea sharing and at a lot of brainstorming by comparing these different approaches.
Thus eventually faster development of these different approaches, even if not all end up converging on the same code base.
Kind of how large scale opensource project have managed to pull out incredible features.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Yeah? What did you charge, your smartphone?
What is the most complex thing todays auto makers have to worry about? Pretty much it's wired input for gas and brake, that's about it... oh maybe something as important as On-Star's ability to stop a car running.
Maybe that the case in the US where the manufacturer seem to be only obsessed with constantly bigger cup holders as the selling feature.
That's not necessarily the case everywhere (Certainly not here around in Europe).
However I do think Tesla is vastly ahead of other car makers in terms of practical self-driving car software, both because the not-quite-self-driving car Autopilot feature has been in production for some time,
Again, numbers matter.
FCAS (Foward Collision Avoidance System), dubbed "City Safety" and using a forward laser grid has been available as a standard on every single one of the cars that VW currently produces, including the lowest cheapest one (like the Up). Not an option, a standard feature coming built-in. More exensive VW event get cameras (including LDAS - Lane departure alert systems). That makes it an extreme large number of cars currently on the streets.
Other brands like Volvo has been putting FCAS and LDAS and more recently lane following built-in in all their higher range of cars.
German manufacturer (such as Mercedes) and asian manufacturer (I think mostly brands) have been using stereo cameras for better stereoscopy depth perception, whereas Tesla only recently started having multiple forward facing camera (and might only using them for different ranges and filed of views, though they could eventually be used for stereoscopy on their overlapping part of FOV).
The thing that currently Tesla is doing with their Autopilot driver assisting technology has been done for quite some time by multiple European and Asian manufacturer, often with more and better sensors.
Tesla's only main advantages are their over-the-air car communication, leading to better "in the wild" data gathering, and faster push of updates.
That has enabled them to make progress much faster, enabling them to eventually close the gap and catch up the older player, and potentially over-take them (though I'm still betting that they'll put a couple more generations of sensors before being to deliver all that they've been promising).
A personal advantage that Tesla has is that their system is a bit more integrated, whereas other manufacturer tend to source some parts from 3rd parties (such as MobilEye), making it cheaper and faster for Tesla to integrate and iterate.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Hacker sez: "Attention all connected cars - turn hard left and accelerate to maximum now!"
Hilarity ensues.
Thatâ(TM)s not how shorting works...
Security needs to be built in, not something added afterwards or on top. If they have a separate security module they did it wrong.
I don't care about their source code, what I would like is the protocol between vehicles to be open and documented.
Nissan has an advanced self diving programme, for example.
Tell me more about this amphibious, submersible Nissan. :)
Meh, Lotus had one years ago, there was even a documentary of a man named Bond driving one into the ocean. Nissan is decades behind Lotus.
"That's the way to do it" - Punch
This is the year of the Linux soft-top.
"That's the way to do it" - Punch
Nope. They are all buying companies working on self-driving systems so that they can control the whole. They disagree with you on what "right" means. Except, of course, for the smallest automakers, or the ones rapidly circling the bowl like FCA.
So - they are letting companies that specialize in self-driving do the development of self-driving?
GM used to build their own fuel injectors for heavy vehicles, back when they owned Detroit Diesel.
So - GM had a different company (which they bought) design and build their fuel injectors?
Tesla has the best EV motor in the industry, literally. And they have it because they designed it.
And they lose over $17,000 on each unit they sell. How long will those "best EVs" be around?
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
Curious about when a $35K Tesla will finally ship, though...
And they're collecting $10k in profit from each sale right now, they aren't going to give that up any sooner than necessary.
Huh, I guess, based upon their loss of $717 million last quarter, Tesla deconstructed and removed from sale about 71,000 vehicles!
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
So - they are letting companies that specialize in self-driving do the development of self-driving?
Except they're not really separate companies any more, they're divisions or wholly-owned subsidiaries which is basically the same thing but with higher HR costs.
So - GM had a different company (which they bought) design and build their fuel injectors?
Absolutely not. Detroit Diesel was recently the "General Motors Detroit Diesel-Allison Division", which they sold to Daimler AG as "Detroit Diesel". Only the Allison part of that (which makes transmissions) was an external company which they bought and folded into the Detroit Diesel Engine Division, which they founded in April, 1938. DDED's first product was the Series 71 diesel engines, which were developed within the GM Research Division. GM actually bought Allison in 1929, and operated it as an internal division thereafter, not merging it with DD until September of 1970.
And they lose over $17,000 on each unit they sell. How long will those "best EVs" be around?
No, they make $10,000 on each Model 3 they sell, and they will make around $6,000 on each $35,000 Model 3 they sell, according to analysis by Munro and Associates. You know, the guys that all the Tesla shorters were applauding for saying mean things about Tesla? Guess it turns out that they are wholly scrupulous, which is why they have literally the best reputation in the industry for this type of product intelligence. There is no reason to doubt their numbers.
Granted, Tesla is reinvesting that profit in the company, and not running off with it or paying it as dividends, because their goal is to grow into profitability. Amazon lost money for a decade while getting their foothold, and achieved profitability through AWS. They achieved this through leveraging cachet. Tesla is losing money while getting their foothold, and will likely achieve profitability by becoming a powertrain supplier, or through some other related industry. And they are achieving this through cachet.
Every other automaker is seeking to develop the same type of hype around their products, and they are specifically looking at Tesla to try to figure out how to do it. This is what all their executives and department heads say when they get roped into a conversation about it. I'm going to trust them, since they have no incentive to lie about it.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
No, they make $10,000 on each Model 3 they sell, and they will make around $6,000 on each $35,000 Model 3 they sell, according to analysis by Munro and Associates
Ahh, so an "analysis" is better than actual published, SEC/GAAP/company approved financials which show a loss of over $17,000 per vehicle? Really? They LOSE OVER SEVENTEEN THOUSAND DOLLARS on every car they sell. That's the hard FACT. Provable numbers. Not some unknown-bought-by-whom analysis by someone else. Actual proven, traceable, hard numbers.
$717,000,000 lost in Q2 2018. They shipped 40,740 vehicles, meaning ($17,000,000 / 40740) $17,600 loss per vehicle. That's the fact.
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
Ahh, so an "analysis" is better than actual published, SEC/GAAP/company approved financials which show a loss of over $17,000 per vehicle?
Those financials include R&D. They do have to pay off the cost of the vehicles. That's why they have to reinvest, so that they can build more vehicles. See, the more you build, the more vehicles that cost is spread across.
Not some unknown-bought-by-whom analysis by someone else.
It doesn't matter who paid for it, only whether the analysis is correct. Do you have some reason to believe that it was not correct?
$717,000,000 lost in Q2 2018. They shipped 40,740 vehicles, meaning ($17,000,000 / 40740) $17,600 loss per vehicle. That's the fact.
The analysis doesn't estimate the R&D costs, only the production, part, and assembly costs. You're doing the math backwards, and the more vehicles they sell, the less they "lose" per vehicle. The R&D is already sunk, and they're not actually losing that money per sale. The research also applies to future vehicles, so it's not all lost on the one model. The fact is, when they sell Model 3s, they make money.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Volvo made not a lot of cars and now every car has a 3 point safety belt.
Don't fight for your country, if your country does not fight for you.
Take the gross revenue, subtract SG&A (sales and general admin - kind of required to do any sales/delivery process) and you're already at a loss. Add in REQUIRED debt servicing, and you're at more of a loss. You want to talk just gross margin but ignore the costs related to more than just making the product. Go ahead - add "R&D" back into the mix, it's still a loss.
And aren't you the one trumpeting how they are great for doing their own R&D to integrate their systems? Wouldn't that make R&D a key spend in their system?
The fact is, when they sell Model 3s, they make money.
Provably wrong. The hard numbers in the financial statement prove as much. SG&A has been scaling with revenues, and that includes Model 3s. They LOSE MONEY on every vehicle they sell. Even if Tesla set R&D to zero - they would still lose money by the hundreds of millions of dollars per quarter. That's from their published financial data - straight out.
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
But look at the trend.
Your not looking at the trend though.
Yes, Volvo, an international car company at rhe time, had one good idea 55 years ago.
The fact that you have to go back 55 years to find an example of a single idea (not selecting and establishing overall protocol standards) tells us something, doesn't it.
To be fair, he promises all kinds of shit that happens after many delays from his initial promise.
The shit does actually happen, eventually; just nowhere near the time table he pulls out of his ass when he promises it.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
I don't understand the logic of "OMG hurry up and make a car that has less margin on it, while I'm constantly bleating about thin margins and profitability!"
Baking analogy (as we're talking about cars, and car analogies about cars don't really work):
I can make a plain cake donut and charge a price that gives me a 15% markup over materials and labor and sell 100% of what I can possibly bake per unit time, or I can make a fancy custard-filled donut that gives me a 30% markup over materials and labor and still sell 100% of what I can possibly bake per unit time.
Hmm, which one I'm going to opt for?
Please stop contradicting yourself.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Yeah. Why don't they just wish for billion-dollar assembly lines to sprout from the ground like weeds and use those?
How come Slashdot's unending capacity for stupidity increases year after year?
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Because we all know that auto assembly factories can just be wished into existence, and cost absolutely nothing to build, and that the zero cost of building that line won't be spread across the future revenue of every car you create with the line after you wish it into existence with no capital costs whatsoever.
Did you have to practice to be this dumb?
If you buy shares of something, you're "long" on it. Shorting involves borrowing shares someone else bought, selling them, and then buying them back at a lower price to give back to the lender once the "loan" is due.
Also, spend 5 seconds on wikipedia / google before commenting on something you clearly know nothing about.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Really.
Google has both the Hyundai Ioniq and the Kia Soul EV at 124 and 110 miles of range, or less than half of the not-yet-existing standard range Model 3, and a little over a third of the shipping for 6-months Model 3 long range.
Yeah, Hyundai has also "announced" the Kona EV with a 250 mile range, but there's exactly as many of those on the road as Model 3 standard range: zero.
So how is around 50% "similar" ? Or is it that they are similarly not available anywhere except a web site?
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Only a fool would think "we're a huge entrenched player so we clearly know everything there is to know about $TOPIC"
Where are Nokia and Blackberry today? How is Sears doing? Enjoying that IBM-manufactured PC you posted this on?
Everyone started somewhere, including Toyota and Volkswagen AG.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Because we all know that auto assembly factories can just be wished into existence, and cost absolutely nothing to build
Doesn't that more or less describe the planning and investment that went into their last assembly line?
'Latest', dammit, not 'last' - I'm not holding a short position!
You really have no idea what you are talking about.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
You just don't understand the concept of capital costs being amortized over the lifetime of the manufacturing line, do you?
Do you run a business where you constantly run into chicken-and-egg problems? The manufacturing line has to exist before you can build the cars, and the profit from selling those cars pays back the cost of building the manufacturing line. At the beginning of the product life cycle, you will have negative cash flow. If you've done your pricing and product planning properly, there will be a crossover point when you stop building manufacturing lines and instead just sell the stuff you make from them. At that point, your capital spend goes down, but your margins stay the same, and magically you go cash flow positive.
This isn't a hard concept, but it's one that continually eludes you, somehow.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
The trend is that gross margin is pretty fixed - SG&A hasn't dropped really any, as a percent of revenues. It is SLOWLY towards the positive - but with 3 quarters of cash left, and a current trend that becomes positive in about 17 quarters - there's a big, 14 quarter gap in between.
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
You seem to think that they will have to keep building billion dollar factories every quarter. No other auto manufacturer has do do this, why would Tesla?
No, they make $10,000 on each Model 3 they sell, and they will make around $6,000 on each $35,000 Model 3 they sell, according to analysis by Munro and Associates. You know, the guys that all the Tesla shorters were applauding for saying mean things about Tesla?...They achieved this through leveraging cachet.
Sorry Mr poo, exactly when did Kia automobile of the 1990s(fast forward to 6:55) struck you as "cachet"?
The only thing Tesla shills are leveraging is bullshit.
Your man crush Sandy Munro had also said the Chevy Bolt(1:03:25) at half the price of a Model 3, had a superior designed body than the "too heavy...cost too much" Model 3.
Munro Associate analysis is only useful for the parts pertaining to the mechanical aspects.
Electronic, finance/economics he's fish out of water.
For example(3:30), why is Munro comparing the cost of a Model 3 rear view mirror (which is just a dumb mirror), to the cost of the mirror of a Chevy Bolt (which has a built in rear camera display)?
Complete idiotic apples and oranges comparison.
BTW, which youtube video did you dig out the $10,000/$6,000 profit numbers from?
Link your source poo.
Sixty years from now Telsa might beat Ford and GM (and Toyota and Volkswagen and Honda and BYD) in the same way that Walmart beat Sears. There's no telling what the industry will look like in 60 years, except to say that based on the lessons of history, most likely about half of the big players will still be there and about half won't.
Sixty DAYS from now, Tesla will still be trying to figure out how to run a production line, not bossing Volkswagen and Toyota around. In 60 years, if by chance Tesla becomes a significant automaker, they can reasonably start talking about drafting standards that other automakers might follow.
So you think Tesla has a "vast quantity" of cars compared to Toyota, VW, or General Motors, out there gathering real-world experience?
What exactly does Tesla have a "vast quantity" of, other than ego and hype?
BTW, there IS an answer to that question. There IS one other thing Tesla has a vast quantity of.
It starts with "L" and ends with "osses".
Supposedly they are at 100% capacity of production right now, so they will need to build more factories...
Browsing at +1 - no ACs, I ignore their posts. So refreshing!
He promises many things, and - usually with delay - delivers.
- we have landing Earth orbital rockets (supersonic retro-propulsion is vital for Mars human exploration)
- we have reusable rockets
- we have the most powerful rocket available for fraction of the price of the next one in launch cargo capabilities
- we do have model 3
- power grid stability: https://www.washingtonpost.com...
His rocket technology can truly open space for sustain human exploration with ~$2000/kg to LEO (FH) through asteroid mining. I remember very well the time, when people were saying that landing an orbital rocket in Earth gravity well is not possible, then that relaunching is not possible, then that it is possible to relaunch, but it will never be profitable - now they are quiet.
People might not like Mr Elon Musk personally, I do not know the guy, but I am happy he peruses his dreams.
Glad you agree, the trend is good so it's moving in the right direction.
But they wont have to design new factories or design new cars. Costs will be less, the trend is good.
supersonic retro-propulsion is vital for Mars human exploration
you actually believe all the spacecraft landings to date have been subsonic?
we have reusable rockets
why do you think NASA didn't even bother to reuse the space shuttle main engines?
we have the most powerful rocket available for fraction of the price
lay off the kool-aid dude
we do have model 3
apparently with the build quality of a 90s era Kia
power grid stability
I'll give you that, assuming you live in some some 3rd world shithole
~$2000/kg to LEO (FH)
I said lay off the kool-aid, you're over fellating Elon by a factor of 5x.
Falcon9=~$100m to put 10,000kg to LEO=~$10,000/kg
Falcon Heavy(FH) is just 3x F9 strapped together, cost & lift are just proportionally scaled.
You should probably stop guessing...
Uh, what about all of the things that DO happen? How much of the crap you spew out of your mouth actually happens? Probably far less than you think, and you don't have millions of people analysing everything you do and say either. I think you might find yourself ashamed if you did. Most people would.
supersonic retro-propulsion is vital for Mars human exploration
you actually believe all the spacecraft landings to date have been subsonic?
Do you know what you're talking about (keywords:"super", "sonic", "retro", "propulsion")?
~$2000/kg to LEO (FH)
I said lay off the kool-aid, you're over fellating Elon by a factor of 5x. Falcon9=~$100m to put 10,000kg to LEO=~$10,000/kg Falcon Heavy(FH) is just 3x F9 strapped together, cost & lift are just proportionally scaled.
"As of March 2013, Falcon Heavy launch prices were below $2,200/kg ($1,000/lb) to low-Earth orbit when the launch vehicle is transporting its maximum delivered cargo weight.[76] The published prices for Falcon Heavy launches have changed somewhat from year to year, with announced prices for the various versions of Falcon Heavy priced at $80–125 million in 2011,[59] $83–128M in 2012,[60] $77–135M in 2013,[77] $85M for up to 6,400 kg (14,100 lb) to GTO in 2014, $90M for up to 8,000 kg (18,000 lb) to GTO in 2016,[78] and $150M for 63,800 kg (140,700 lb) to LEO or 26,700 kg (58,900 lb) to GTO (fully expendable)or $95M for 90% of its maximum capacity in 2018" https://en.wikipedia.org/wiki/...
Do you have any factual arguments to support your statements besides:
lay off the kool-aid dude
?
No, I was referring to you specifically talking about BMW, and some of the stuff they have on their cars.
I own a 2015 BMW 435 that has the "driver assistance" package with lane departure warnings, frontal collision warnings, and all that stuff. I get more false positives out of it than actual warnings to the point of finding it annoying and thinking about reaching for the button to the lower left of the steering column to turn it off - it vibrates the steering wheel like I'm leaving the lane when going down a road that has crack sealant running parallel to the lane striping. I've had rain bring up the frontal collision warning on the HUD.
If the data that they are getting back is no good for the simple "you're going over a dotted line" or "there's something really close to your bumper" then what makes it any good for sorting out autonomous driving? And please cite your source showing that BMW is collecting any of this data to begin with.
Until you can answer those two questions, I'm staying with "You really have no idea what you are talking about."
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
No, I was referring to you specifically talking about BMW, and some of the stuff they have on their cars.
I own a 2015 BMW 435 that has the "driver assistance" package with lane departure warnings, frontal collision warnings, and all that stuff. I get more false positives out of it than actual warnings to the point of finding it annoying and thinking about reaching for the button to the lower left of the steering column to turn it off - it vibrates the steering wheel like I'm leaving the lane when going down a road that has crack sealant running parallel to the lane striping. I've had rain bring up the frontal collision warning on the HUD.
If the data that they are getting back is no good for the simple "you're going over a dotted line" or "there's something really close to your bumper" tasks being asked of it today, then what makes it any good for sorting out the amazingly more complex tasks of autonomous driving? And please cite your source showing that BMW is collecting any of this data to begin with.
Until you can answer those two questions, I'm staying with "You really have no idea what you are talking about."
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
Not really. Lidar basically just gives you a depth map. You can get the same thing from a pair of cameras, which Teslas have, at least on the front (the only direction that really matters that much). That doesn't require sci-fi-level strong AI; that's literally what the iPhone does every time you take a photo in portrait mode. It's basically a glorified version of an MPEG encoder's P-frame motion vector computation. It doesn't even require ML, much less strong AI. It is just a bunch of motion vector computation between two reference frames.
Check out my sci-fi/humor trilogy at PatriotsBooks.
When I develop software, the valuable information that can help me improve a system is information about under what conditions it's not yet correct, what's going wrong. It's not too helpful to say "works for me in my prestine lab environment". I also don't develop many improvements based on "I haven't noticed any issues". We improve things by characterizing what's not working well.
If BMW has a lot of real-world examples of things that cause problems with a naive implementation, such as distingusiing between a crack sealant running parallel to the lane marking vs the lane marking itself, that's exactly the kind of data I would expect to be most useful for making improvements. I don't work for BMW, so I don't know exactly what kind of data they are gathering, either summary data or detail, or samples, but they certainly have a lot of sensors out in real-world conditions, getting experience with real-world problems. According to you, each and every one of their cars likely has some example problems they can work with. Millions of cars generating billions of data points is obviously much more "vast" than thousands of Tesla cars.
balh blah blah..
No witty comeback for rocket re-usability is marketing bullshit with no economic benefit, and Tesla Model 3 is a piece of shit.
Thanks for the implicit agreement with those claims.
As of March 2013, Falcon Heavy launch prices were below $2,200/kg
Falcon Heavy first flight, Feb 6,2018.
Maybe you can soothsay the launch price of Spacex BFR?(assume the first flight is like 2028?). Better yet, let's have the winning lottery numbers next week.
Falcon Heavy launch prices... $150M for 63,800 kg (140,700 lb) to LEO
Falcon 9 CANNOT, repeat CANNOT even lift 10,000kg to LEO
How is Falcon Heavy (FH), being 3x Falcon9 strapped together, going to lift 63,800 kg to LEO?
Spacex math: 10,000+10,000+10,000=63,800?
Not to mention Spacex is charging $152m per F9 flight(rising to $200million by 2020) to supply the ISS.
What are the chances FH (3x F9 strapped together) will be priced at $150m?
Try extracting Elon's tiny penis out of your anus, it's affecting blood supply to your brain.
That's wonderful that as a software developer you know how to sniff out edge cases and account for them. What do you do if you have an edge case that you can't sniff out, because your data is garbage? As the saying goes, garbage in - garbage out.
You have no idea if they are even collecting this data, by your own admission. You don't know what sensors they have on the cars, what the quality of data is that they get via those sensors, what kind of processing they are doing in order to do what the cars do today, or if that data is even recorded anywhere for further analysis. Which basically reinforces my statement that you have no idea what you are talking about - it doesn't matter if you have a million microphones to record the sound of a tree falling in the woods if none of them are plugged in, or if they are the cheapest microphones possible because you designed the system to only record if a sound happened, and not record any fidelity at all, so those microphones chop the hell out of the frequencies so as to make the recorded data inadequate for any form of analysis past "there was a pressure wave of X dB." Sure, some data is better than no data, but if you have tremendous data quality issues you may be better off with no data.
The point someone above was making: we know that Tesla, even though their fleet of vehicles is smaller than other manufacturers, actually is retrieving billions of vehicle-miles travelled worth of data from actual hardware that can do the job. That is likely to give them an advantage. There's a reason why Apple, Waymo, et. al. are driving around collecting data on minivans - anything they could get their hands on is just inadequate for the job.
Slashdot still doesnâ(TM)t support Unicode after it was added to the HTML standard in 1997.
> Tesla, actually is retrieving billions of vehicle-miles
Billions? Tesla's production numbers are in the thousands, not the millions. Does each one of their cars drive a million miles per year?
We don't know exactly what data each company is collecting. What we do know is that Tesla is a *small* and hopefully therefore agile company. Nothing about Tesla is "vast". BMW, Freightliner, or other actual auto industry companies could have been collecting only 1% as much data as Tesla per mile for the last ten years and they'd STILL have more data than Tesla. That's even assuming, for no reason at all, that Tesla captures a hundred times as much data per mile, a very questionable assumption.
We can also be fairly certain that Freightliner trucks drive a lot more miles per year, in more widely varying conditions, than your average Tesla.
There are plenty of good things to say about Tesla. You can talk about their vision. Pretending they have a huge advantage in "vast numbers" and "years of experience" just makes one look comically ignorant.
"Once again, LynnwoodRooster fails to understand how costs work, despite having it repeatedly explained".