Thứ Bảy, 24 tháng 12, 2016

Tesla is dumping Mobileye??? part 1

  • Jul 26, 2016
    jkliu47
  • Jul 26, 2016
    Xenius
    Interesting, wonder if this was their doing, or if BMW orchestrated it, or maybe even Tesla. Hmmm.
    �
  • Jul 26, 2016
    ShadowR55
    So how will this affect the second iteration of the hardware and current AP1.0 customers?
    �
  • Jul 26, 2016
    plankeye
  • Jul 26, 2016
    ShadowR55
    The suspense is killing me. Hopefully they don't leave current AP customers in the cold.
    �
  • Jul 26, 2016
    Jtrader
  • Jul 26, 2016
    h_corey
    Who initiated this move? Intel/BMW? Mobileye because of the bad press? Tesla because they can recreate what Mobileye is doing? Did this really happen when Elon made a big call for software engineers? Just lately?

    This is certainly one way to make sure Tesla stays a generation behind in autopilot starting ~2018 when eyeQ4 comes out. It feels like BMW was in big trouble and paid Mobileye for an exclusive on eyeQ4.

    Not good in my view. Mobileye was the path of least resistance to full autonomy.
    �
  • Jul 26, 2016
    Vitold
    There was a rumor in April that relationship was over. Musk visit to MBLY was probably to get Tesla back on-board - obviously did not succeed.

    Going with MBLY eteQ4 solution was never in the cards because Tesla would loose the spot as a leader and would be dependent on eyeQ4 delays which are likely.

    My guess is that Tesla was planning to continue using eyeQ3 for front facing camera combined with Tesla/Nvidia silicone (Tesla hired Apple chip designer) and more sensors. After confusing trailer with sky fiasco Tesla maybe moving away from the MBLY platform altogether.
    �
  • Jul 26, 2016
    Vitold
    This is exciting news as it confirms that Tesla is developing their own system.
    �
  • Jul 26, 2016
    Mo City
    I hope that is the case and Tesla is the party that decided to end the relationship.

    I also hope this was already in the works instead of being a reaction to Joshua Brown's tragedy.
    �
  • Jul 26, 2016
    h_corey
    So who initiated the move? Tesla or Mobileye? From Elon's tweets I get the sense that he likes Mobileye a lot.
    �
  • Jul 26, 2016
    Vitold
    There are two threads about this. As I wrote elsewhere Tesla likely ended relationship in April. Apparently Musk's recent trip to MBLY HQ was to decide between doing it alone or sign up for eyeQ4.

    Considering that MBLY said they need one year to come-up with a solution for detecting cross traffic it may have rubbed Musk a wrong way as well.
    �
  • Jul 26, 2016
    WarpedOne
    You don't to help us anymore?
    OK, we will do it without you.

    Why did Panasonic join the GF story? They saw that if they stayed out, they would be left behind.
    MBLY does not see that they just sealed their death.
    �
  • Jul 26, 2016
    ecarfan
    That would be my guess. I think it is clear that Elon recognizes just how important autonomous driving is. He has authorized hiring a big team to make it happen, and the head of that team reports directly to him. He wants Tesla to have an autonomous system that is second to none. It would be advantageous if Tesla totally controlled such a system and did not have to rely on outside sources for hardware or software.

    I doubt that Mobileye's decided to end its relationship with Tesla. They are are valuable, high profile client. It is more plausible that Elon decided to end the relationship with Mobileye so that Tesla has full control and ownership of all aspects of Tesla's autonomous driving system when it becomes a reality, which Elon has stated he believes will be in two or three years (but regulatory approval will take longer).
    �
  • Jul 26, 2016
    h_corey
    That is my thought as well. If Mobileye forces Tesla to produce an autopilot without them, they just opened up major competitor number 1. Monopoly gone by 2018-2020.
    �
  • Jul 26, 2016
    Vitold
    I think it was Tesla that made the move. Mobileye said that they need one year to detect cross traffic - that's a lifetime in EM years.

    EDIT: It's also possible that MBLY is reserving some of the features for eyeQ4 and Tesla, having invested heavily in hybrid eyeQ3 saw no other way as doing it alone (probably with Nvidia's help).
    �
  • Jul 26, 2016
    ecarfan
    Agreed. I think it is clear that Elon recognizes just how important autonomous driving is. He has authorized hiring a big team to make it happen, and the head of that team reports directly to him. He wants Tesla to have an autonomous system that is second to none. It would be advantageous if Tesla totally controlled such a system and did not have to rely on outside sources for hardware or software.

    I doubt that Mobileye's decided to end its relationship with Tesla. They are are valuable, high profile client. It is more plausible that Elon decided to end the relationship with Mobileye so that Tesla has full control and ownership of all aspects of Tesla's autonomous driving system when it becomes a reality, which Elon has stated he believes will be in two or three years (but regulatory approval will take longer
    �
  • Jul 26, 2016
    blablub
    Who knows. Maybe it has to do something with the new partnership between Mobileye, Intel and BMW.
    �
  • Jul 26, 2016
    ivengo
  • Jul 26, 2016
    Vitold
    I actually think that Intel-BMW-Mobileye alliance is in response to Tesla AP 2.0
    �
  • Jul 26, 2016
    Darren F
    Woah!
    �
  • Jul 26, 2016
    Mo City
    OK, this seems to be the main thread for the subject.

    Agreed it is very unlikely the vendor (Mobileye) ended this relationship and it is quite possible Mobileye's agreement with Intel & BMW resulted from Tesla's decision to move on.

    Obviously Elon Musk will have something to say about this very soon. Curious how much he reveals.

    We now know why he mentioned AP on current Tesla's will be improved going forward.
    �
  • Jul 26, 2016
    Vitold
    Tesla and Mobileye parted their ways three months ago, before Intel-BMW-MBLY announcement:
    Mobileye NV (MBLY) Dips on Rumors of Losing Tesla Motors Inc as Client

    Also, there would be no point for EM traveling to MBLY HQ recently to demonstrate their tech if MBLY/MBW wanted Tesla out.
    �
  • Jul 26, 2016
    JohnSnowNW
    I don't know enough about this to form an opinion...but I'd kinda like to hear what Tesla is doing instead.
    �
  • Jul 26, 2016
    xkwizit
  • Jul 26, 2016
    Shawn Snider
    Sounds like BMW and Intel payed em off lol
    �
  • Jul 26, 2016
    SabrToothSqrl
    Yes, they did. It's so sad. I wonder who gets the dog. :( And the CD collection. Are CDs still a thing?
    �
  • Jul 26, 2016
    brkaus
    Well, really not sure what that means for my future model s. Maybe time to just lease, although my preference is to drive cars 15 years.
    �
  • Jul 26, 2016
    Vitold
    If anything, "Tesla doing it alone", moves likely-hood of Tesla AP 2.0 to a later time.
    �
  • Jul 26, 2016
    ivengo
    Thinking about it I think it is a good thing, because Tesla will have control of more components from scratch so they can provide tighter integration and more superior final solution. They hired a person who worked in AMD on processor manufacturing and it means that they started panning to bring everything in house some time ago.
    �
  • Jul 26, 2016
    ItsNotAboutTheMoney
    With Musk's recent tweet about changing how the Autopilot sees and MobilEye's hook-up with BMW and IBM, I think that both companies could be looking in different directions.
    �
  • Jul 26, 2016
    omarsultan
    Bringing it in house makes sense. Seeing as AP technology is so central to Tesla's brand and vision, I can see them wanting to have great control over direction and pace of the technology's evolution, plus, there are cost benefits to bringing it in house. I wonder if there is a MobieEye competitor out there that Tesla is eying for an acquisition.

    UPDATE:
    Reading the article, it sounds like maybe the MobilEye technology did not make the cut for AP 2.0
    �
  • Jul 26, 2016
    Cosmacelf
    Wow, this is almost insane. Mobileye has very sophisticated chips that run complex neural network algorithms to understand what the camera is seeing. Tesla must now source (or build, gulp) chips from someone else and develop equivalent neural networks to recognize cars, paths, people, etc. Tesla did hire a microprocessor design guru less than a year ago. I sure hope Tesla's engineers aren't doing what engineers normally do, look over someone else's work and say "Pshh, that's so trivial, I could build that in a weekend".

    By bringing AutoPilot hardware and software completely in house, I fear this will delay the introduction of AP 2.0 by quite a bit.
    �
  • Jul 26, 2016
    tentonine
    Another possibility is that Tesla could use NVIDIA's Drive PX instead.
    �
  • Jul 26, 2016
    brkaus
    Agree...
    �
  • Jul 26, 2016
    Gizmotoy
    Back in May there were some articles that said MobileEye was telling customers not to expect EyeQ4 until 2018. Engineering samples of EyeQ5 are expected at around the same time. Knowing Musk, it's likely he decided Tesla can do it faster. Or perhaps more specifically, that they must do it faster to get them into Model 3.

    So I suspect we'll be waiting a bit for a substantial AP upgrade.

    Cross-traffic is hard to do with forward and rear-facing cameras. There's a reason many of the existing implementations use corner-mounted radar units (some also fuse data from the radar with wide angle camera data to determine object type). The ultrasonics Tesla chose for proximity detection are particularly bad for cross traffic, so they must either fit new hardware (radar) or rely solely on the camera.

    It's not an impossibility, but it is a hard problem. I don't know how wide the front camera is, but the rear is nowhere near as wide as that on some other vehicles that use the rear camera for cross traffic detection.
    �
  • Jul 26, 2016
    Dynastar
    Tesla hired Jim Keller, a microprocessor engineer, six months ago which is well before the autopilot fatality. I think Tesla has been planing on going it alone for some time, and this is a very SpaceX-ish move by them. Elon has said before that SpaceX commonly outsources parts for their rockets while developing the same part in-house. Once the in-house part has shown that it is as capable as the outsourced one SpaceX switches to it. This is one of the big reasons SpaceX is able to thump the competition when it comes to costs and features.
    �
  • Jul 26, 2016
    xav-
    Time to short Mobileye? The stock is super pricy..
    �
  • Jul 26, 2016
    privater
    I believe their CEO's stupidity is the reason: they openly blame on Tesla autopilot after the crash to promote their future product. That backstabbing made Tesla embarrassing and cut the loose in quick. Now tesla might fully cooperate with Nvidia, they have better hardware design and long relationship with Tesla plus it's easy to work with someone in close distance.
    Ps: Nvidia CEO have a good relationship with Elon in person.
    �
  • Jul 26, 2016
    brkaus
    Having a cpu guy for 6 months doesn't get them a solution they can put in a car, probably not even in time for the 3.
    �
  • Jul 26, 2016
    Sir Guacamolaf
    This tells me that eventually Tesla's autopilot will be about as good as their media and phone integration is today.
    �
  • Jul 26, 2016
    J1mbo
    Maybe explains why the MX and facelift MS only got one camara when they were clearly intended to have at least 2.

    They probably would want to concentrate on software and integration rather than building a new AP hardware platform inhouse.

    I am guessing that the simpliest option would be to move to Bosch for the camera suite and either Nvida or Qualcomm for the DNN (self-driving).
    �
  • Jul 26, 2016
    voyager
    Why would a partnership with Intel and BMW be a deal breaker to Musk anyway?
    �
  • Jul 26, 2016
    Matias
    Let's hope this doesn't delay AP 2.0.
    �
  • Jul 26, 2016
    Tam
    Tesla has a large deposits for Model ?.

    For Mobileye, that's a loss of 400,000 chips in a near future.

    The loss is even more staggering as Tesla plans to produce 500,000 in 2018 and 1 million vehicles annually starting 2020.

    But something is worth much more than money--reputation:

    Mobileye NV (MBLY) Stock Tumbles As Partnership With Tesla Motors Inc (TSLA) Ends

    "Without directly citing the recent fatal accident, he suggested reputation concerns about safety of the company�s technology, clearly referring to the crash which took place while the autonomous driving technology was activated."

    As mentioned by @tentonine, Mobileye is not the only company, Nvidia is a good alternative too.

    How Our Deep Learning Tech Taught a Car to Drive | NVIDIA Blog
    �
  • Jul 26, 2016
    shokunin
    Not surprising, especially when the mobileye roadmap was hinting to 2017-2018 for EyeQ4 chipset. Buying someone else's autonomous driving hardware and chip processing only gives you a small differentiator.

    I'd rather see them license google's autonomous driving tech, even if that means giving in to LIDAR.

    I guess it's highly unlikely to retrofit AP1 to AP2.
    �
  • Jul 26, 2016
    zambono
    I had the same thought
    �
  • Jul 26, 2016
    anticitizen13.7
    Drive PX is more of a general purpose computing solution than Mobileye's EyeQ series, which are ASICs tailored to run Mobileye's algoritms. In terms of overall power consumption and performance per watt, I'm not sure that Drive PX is better than Mobileye.


    I remember from other news articles that Mobileye had expressed reservations about how quickly Tesla was advancing the Autopilot system in its cars. I got the impression that Mobileye's leadership viewed Tesla's actions as risky.

    While Mobileye's EyeQ visual processing unit is just one component of Autopilot (front radar, surround sonar, are from other suppliers, and Tesla has to write the software to tie everything together), I can understand how the general public might conflate Mobileye with Autopilot.
    �
  • Jul 26, 2016
    RDoc
    You didn't mention their Most Excellent implementation of the Navigation system!

    If Tesla is going to rely on their in-house software expertise, I'd suggest everyone get used to the current Beta and plan on their kids using it as well. Grandchildren might be another story if you're still young.
    �
  • Jul 26, 2016
    shokunin
    Mobileye is making it sound like they ended the partnership. Possible that Elon was forcing their roadmap to improve and enhance the Q3 platform since EyeQ4 is still a year+ out and mobileye didn't want to invest in an aging platform.
    �
  • Jul 26, 2016
    Fallenone
    Power consumption and performance per watt may be issues for ICE with limited electric power to support these, but not an issue for EVs.

    Also I agree MBLY is worried about the pace of TSLA's progression in autonomous driving. At least this is the main reason they claimed in today's statement ("our reputation").
    �
  • Jul 26, 2016
    brkaus
    I would guess nvidia. They already have them in the center display, don't they?. Of course, that hasn't been updated recently either.
    �
  • Jul 26, 2016
    jhm
    Mobileye's technology was not going to be able to detect lateral transvers movements of vehicles until 2018. This was the sort of movement in the recent fatality. Tesla took all the heat, but the Mobileye's technology was not up to the task. So I wonder if this has something to do with the break up.
    �
  • Jul 26, 2016
    ecarfan
    Agreed. Elon has been clear that autonomous driving capability is central to his vision of an EV future. He has authorized hiring a large engineering group just to work on AP and the person in charge of that group reports directly to Elon.

    It is a big blow for Mobileye.
    �
  • Jul 26, 2016
    Johan
    Elon on CC:

    "One of the challenges we face is that for a lot of supply chains they're impedance-matched to the timeframe of the big OEMS, and Tesla just moves a lot faster than the big OEMS. And so, if they're impedance-matched to a typical sort of 6-year development cycle and we're on a 2- or 3-year development cycle, it just doesn't connect properly. Some suppliers can handle that, and some can't."

    Mobileye couldn't deliver the next gen chip in time, so Tesla cut them out. It's that simple.
    �
  • Jul 26, 2016
    Vitold
    It's other way around - BMW would probably not like Tesla getting the latest tech. MBLY needed a partner after Tesla's departure, BMW needed autonomous tech and Intel to get "foot in the door".

    I think the alliance is positive for MBLY and puts pressure on Tesla, however, just like MBLY-Tesla, MBLY-Intel are temporary allies.
    �
  • Jul 26, 2016
    dixonpete
    I have to think this is a big thing. All the AI that's been developed, isn't it dependent on the Mobil-Eye hardware?
    �
  • Jul 26, 2016
    Stoneymonster
    Probably not, if they've done it right.
    �
  • Jul 26, 2016
    andrewket
    This isn't a surprise. Tesla already uses their own software on top of mobil-eye's hardware when other manufacturers are using their software. Tesla sees AP as critical to their success, and clearly likes to in-house critical dependencies.

    I'm still bullish on Mobil-eye, though.
    �
  • Jul 26, 2016
    malcolm
    Key AP employees were appointed by Elon and he keeps very close tabs on its development.

    Fair to say that he has little patience for any perceived bottlenecks and/or suppliers who may move at a different speed.
    �
  • Jul 26, 2016
    voyager
    Are OEMs like BMW able to claim vendors?
    �
  • Jul 26, 2016
    stopcrazypp
    Anything is possible. You just need to sign a contract. I don't think there are any laws against signing exclusive contracts with a vendor.

    As a general comment, I can see justification for both reasons (whether Mobileye initiated or Tesla initiated). I can see Mobileye ending from pressure by BMW/Intel, Mobileye wanting to get away from Tesla's autopilot controversy, or disagreeing with the extensive customization Tesla has done. I can see Tesla ending from wanting to move things in house, reducing the development time, or adopting a different company with better technology or deal (like Nvidia maybe). A reason for mutual ending of the relationship would be disagreements over ownership of data (this had ended other autonomous car partnerships before).
    �
  • Jul 26, 2016
    NicoV
    'A person' is a bit of an understatement. That's like saying they hired 'a musician' if they would hire Mick Jagger. They hired Jim Keller, Jim Keller (engineer) - Wikipedia, the free encyclopedia , and with him, a lot of his collegues from AMD. I'm pretty confident they will build a world class chip.
    �
  • Jul 26, 2016
    Tam
    George Hotz, Comma.ai CEO says: "Without a doubt it�s Tesla who parted ways.�

    "Mobileye isn�t innovative enough for Tesla."

    "�Their stated mission is to lower the safety rating of cars that don�t have their technology,� he toldRecode. �So when you have a company like that versus a company like Tesla who actually wants to build incredible cars, there�s absolutely no reason Tesla needs Mobileye.�"

    In the wake of a fatal crash, Tesla will quit using Mobileye's chips for Autopilot vision
    �
  • Jul 26, 2016
    Vitold
  • Jul 26, 2016
    Vitold
    Mobileye CEO said that Tesla did not want to work with them like BMW and Intel are willing to (MBLY wants their data probably):

    Since BMW are beyond 'typical OEM' they probably will get the latest and greatest before anyone else (OEM's). Tesla does not want that special relationship, obviously thinking they can do better.
    �
  • Jul 26, 2016
    scottf200
    Aside: This doesn't seem to make sense in an X subforum vs Tesla overall one.

    Well who will Tesla turn to? Mobileye was the king and various other companies are using them.

    related keyword didn't help me search.

    related:www.mobileye.com link related:www.mobileye.com - Google Search
    �
  • Jul 26, 2016
    scottf200
    Also do this search as Mobileye is doing crowd sourced maps with GM, VW, Nissan, etc, etc. With all those players they will get a lot more data than Tesla one would imagine ... I mean they will add it to some of their ICE cars too.

    crowd sourced maps mobileye - Google Search
    �
  • Jul 26, 2016
    anticitizen13.7
    I'd rather have lower power consumption and performance per watt though. Even in a car with a huge battery, every bit of saved energy helps.

    Also, electronics that have high heat dissipation require more cooling, which means more wasted energy and the need to accommodate more parts in the car.


    I would take anything that George Hotz says with a grain of salt. First, he's not working for Tesla and probably has no idea what is really going on there. Second, he says things that are a combination of arrogance and irrelevance. From the linked re/code article, he said:

    Hotz boasted back in December than his tricked out Acura ILX would self-drive better than the Tesla Model S in a short amount of time. That claim hasn't panned out.
    �
  • Jul 26, 2016
    hockeythug
    Kind of saw this coming when Tesla wasn't going to share their data with Mobileye's pool.
    �
  • Jul 26, 2016
    Tam
    Could he mean Tesla doesn't have a free all-you-can-eat cafeteria :)
    �
  • Jul 26, 2016
    Tam
    Tesla Motors Inc (TSLA) Doesn't Require "Obsolete" EyeQ Chips: Chowdhry

    Our popular analyst Trip Chowdhry from Global Equities Research "claimed that Mobileye�s technology is �obsolete�

    There's a need for processing a field of vision between 150-180 degrees but Mobileye can offer only 50 degrees.

    Its processors do not offer required dynamic range, thus it cannot distinguish objects in high light or dark conditions.

    He believes Tesla will use LIDAR in future.
    �
  • Jul 26, 2016
    JohnSnowNW
    Not according to Musk, but people can change their mind.
    �
  • Jul 26, 2016
    sillydriver
    I assume Tesla is bringing in house what they were doing with Mobileye for the same reason Apple brought in house what they were doing with google on maps: they want to control what they see as a core platform/competence. I just hope Tesla does better than Apple has with maps!
    �
  • Jul 26, 2016
    3Victoria
    I think EM believes that they can achieve a low resolution 3D image (point cloud) from the radar unit. While this would be lower resolution than using Lidar, it has the advantage of seeing through weather (rain, snow).
    �
  • Jul 26, 2016
    anticitizen13.7
    I'm astonished that whatever company Trip Chowdhry works for hasn't fired his ass yet. EyeQ3 won't be "obsolete" until something better reaches the marketplace.

    The EyeQ3 platform might not have these items, but that's not to say Mobileye won't offer improvements in the future.
    �
  • Jul 26, 2016
    sillydriver
    I'm sure the decision to go in house was made a while ago, given the hiring of chip designers. I think the consequences are

    1) the transition to 2.0 is not imminent but at least a while off, and

    2) there will be a greater discontinuity (as opposed to incremental change) between 1.0 and 2.0 -- obviously in hardware but also in software algorithms and functionality. The likelihood of upgradability for existing owners is lower.
    �
  • Jul 26, 2016
    Johan
    I agree, but with regard to your point #2 there really was never much likelihood of upgradability even if they stayed with Mobileye as hardware supplier.
    �
  • Jul 26, 2016
    sillydriver
    Agreed. But we could hope!

    Less hope now.
    �
  • Jul 26, 2016
    sillydriver
    If that's Chowdhry's reasoning it seems bizarre to me. Mobileye's core hardware does multiply-accumulate operations for neural networks, just as GPU hardware can can do the same. Changing the field of view is a matter of changing a cheap lens and perhaps some trivial parameters in the software. Changing the dynamic range is a matter of changing the camera, which I'm sure is outsourced, to one with a different spec. Not a big deal.

    I suspect Tesla going in-house has more to do with their wanting to control what they believe will be a core differentiating technology.
    �
  • Jul 26, 2016
    S4WRXTTCS
    Over the long haul I think this will work out great.

    One of the problems right now with the Tesla Model S is it's really a bunch of third party solutions mashed together.

    In Model S/X

    You have Mobile Eye for the Image Processing for Autopilot
    You have Google for the online maps
    You have Garmin/Navigon maps for the offline maps
    You have a really old Nvidia Tegra solution for the Infotainment
    The rear camera doesn't go to anything, but the infotainment 17 inch display
    The front camera isn't viewable in the infotainment display
    There is no dashcam capabilities

    It's really quite limited so it would be nice to see them switch to a more integrated solution. You still want to keep separation between infotainment and critical functionality, but where things were duplicated (like the signals from the cameras going to both the infotainment system and the autopilot computer).
    �
  • Jul 26, 2016
    Alketi
    I just watched an interview with George that covered a number of topics, but one interesting tidbit was that he said the MobileEye camera (that's in Tesla vehicles) is a RCC (Red-Clear-Clear), which sheds more light on why it wasn't able to discern a white truck against a bright (possibly blue) sky. Since the camera is monochrome it sees shading and contrast as opposed to color.

    There's just no way Tesla can be reliant on a single, 3rd party vendor for such a critical component like it's autopilot hardware suite and processors. No question in my mind it was Tesla who walked away.
    �
  • Jul 26, 2016
    Tam
    It might be the processing capability.

    A camera may supply all the beautiful colors of the light spectrum but Mobileye can only process red and gray intensity which explained why it can't see a big white semitrailer against a brightly lit sky.
    �
  • Jul 26, 2016
    dakh
    Looks to me Tesla's plan was to end the partnership at some point. The accident probably put some real issues into focus that made it necessary to have this announced publicly. I hope what Elon is saying is true, that this doesn't throw Tesla's progress back.
    �
  • Jul 26, 2016
    laorient
    I think there are 3 main reasons for the breakup:
    1. MobileEye wants to keep the TDP low while increasing processing power, to cater to ICE customers also. But Tesla has less concern on TDP and want to go full autonomous fast. This means their goals are not perfectly aligned.
    2. MobileEye has a lot of "legacy" platforms to support, which means they probably can not turn to things such as Full Color Camera, Stereotypical Vision quickly since these will render "legacy" systems to be obsolete. But Tesla is willing to sacrifice their older platforms, as they have repeatedly done.
    3. Tesla has the unique advantage of high speed, constant connection with their cars, with rapid update capability, which none of the other MobileEye customers has. This again allows Tesla to collect more data and evolve status quo quicker, while MobileEye is having a hard time catching up to them.
    Elon Musk is very far-sighted, and he had wisely decided to NOT share the data collected from Model S/X with MobileEye, which means the in house DNN Tesla trains currently maybe leaps and bounds ahead of MobileEye now. And that's when Elon says: "So long, and thanks for all the fish.".
    �
  • Jul 26, 2016
    h_corey
    As an aside, I believe intel wants to buy mobileye to get a bigger foot in the door of the next 'mobile' device trend.
    �
  • Jul 26, 2016
    AmpedRealtor
  • Jul 26, 2016
    Lex
    Yep it seems that MobilEye are the ones making the move to not continue "past the current line of products" according to Amped's cited article, which could mean all of Autopilot 1.0, and as everyone seems to think that Autopilot 2.0 is autonomous anyway, it seems they've left the door open to re-enter... ???

    So if they continue to supply for AP 1.0 and are then interested in AP 2.0 then I'm not sure what changes. Seems like they're trying to distance themselves from a single tragedy from what is "possibly / allegedly" looking like negligent drivers and/or an incapacitated driver.

    But if they're pulling out of what they're already in, then I surely see why Elon and Co. are so interested in keeping everything in-house, as "turning off" MobilEye in our AP systems today turns it back into just TACC. MobilEye is a big part of AP from what I know. I wonder if the "surge" in AP related coders was all about replacing MobilEye. Maybe the writing was already on the wall.
    �
  • Jul 26, 2016
    182RG
  • Jul 26, 2016
    Alketi
    Interesting. I would not have expected MobileEye would be the one to turn away business. Usually the vendor doesn't leave the parent.

    Reading between the lines, MobileEye's story seems to be that they don't like the fact that Tesla is controlling the implementation (re: software).

    I wonder how this impacts AP 2.0. If Tesla anticipated this or it was mutual, then they'd have already been spec'ing other non/MobileEye hardware.

    Lastly, this certainly puts the emails between Elon and George Hotz in a new light. I wonder if the MobileEye relationship became strained when those emails leaked publicly.
    �
  • Jul 26, 2016
    Twiglett
    I would suspect the fact that MobileEye were not getting any of the telemetry and mapping data from Tesla was a big issue for them. Tesla has a huge advantage that would be holding back MobileEye and the others.
    �
  • Jul 26, 2016
    Waiting4M3
    In 2015 Mobileye announced that EyeQ4 wasn't going to be ready until 2018
    Moving Closer to Automated Driving, Mobileye Unveils EyeQ4� System-on-Chip with its First Design Win for 2018

    Tesla hired Jim Keller and Peter Bannon in Jan-Feb 2016, who were with PA Semi, which Apple bought to design A4/A5 chips for their iPhone
    The chip guru who built Apple�s Ax microprocessors joins Tesla to lead the Autopilot Hardware Engineering team
    Tesla hires yet another chip architecture titan out of Apple�s PA Semi, feeding the rumor that it plans to design its own silicon

    Then rumor had it that Elon visited Mobileye in early March 2016
    Elon Musk reportedly visited Mobileye to test tech for next gen Tesla Autopilot

    I think it's pretty easy to guess what happened. Elon wasn't happy with the EyeQ4 2018 roadmap, hired guys for in-house plan-B, then met with Mobileye to start the divorce proceedings, we just now learned that all the divorce papers have been signed.
    �
  • Jul 26, 2016
    plankeye
    From that Bosch video, looks like Tesla doesn't need them anyway. Plus, ME might have been slowing them down. They most likely just need a fairly high horsepower processor to go with the software they already have and are continuing to improve.
    �
  • Jul 26, 2016
    Odebek
    I am sure it is just MobileEye saving face, part of the negotiation was that they get to announce that they are cancelling the contract. Tesla is a small part of their overall sales, because Tesla only buys the chips and not the software/services; but the contract with Tesla is what brought them to prominence, so it is important that they be seen as the entity ending the contract, otherwise they would have taken a bigger stock hit.

    Also, TSLA has been busy buying up AP hardware talent so it was part of the strategy to part ways on their part: Tesla poached a team of chip architects and execs from AMD to develop the next gen Autopilot
    �
  • Jul 26, 2016
    hybridbear
    This seems quite surprising considering past statements about how crucial Mobileye's technology was. In one article I researched again this afternoon, Elon said that Mobileye's technology is the best in the world & that is why Tesla uses their tech. I wonder who will supply the AP camera & chip now.
    �
  • Jul 26, 2016
    hybridbear
    It's Mobileye that ended the relationship according to this article.
    �
  • Jul 26, 2016
    3Victoria
    It is unclear who did the leaving, perhaps it was mutual. Both parties would like to save face. MobilEye felt they were at risk for bad publicity when they were not in control of how their systems were used. Tesla wants to move quickly, and doesn't what to use a generic system that is shared by other manufacturers. I think Tesla will push things along for their cars and maintain their lead, and MobilEye will provide reliable systems to the other manufacturers. Win win. (I am impressed by both companies.)
    �
  • Jul 26, 2016
    Hodginator
    Maybe they will switch to the Nvidia DRIVE PX solution. Hasn't that been a rumor for a while?
    �
  • Jul 26, 2016
    sillydriver
    I think your view is exactly right. Tesla sees autonomous driving as one of their core differentiators and won't tolerate being generic. But there are plenty of established car companies that see it as a checklist capability to be outsourced. Mobileye's potential sales into that group are quite large.
    �
  • Jul 26, 2016
    Tam
    BRIEF-Tesla's Musk on Mobileye our parting ways was inevitable


    Elon Musk comments on Tesla discontinuing Mobileye�s Autopilot system, says it doesn�t affect timeline

    �This was expected and will not have any material effect on our plans. MobilEye�s ability to evolve its technology is unfortunately negatively affected by having to support hundreds of models from legacy auto companies, resulting in a very high engineering drag coefficient. Tesla is laser-focused on achieving full self-driving capability on one integrated platform with an order of magnitude greater safety than the average manually driven car.�
    �
  • Jul 26, 2016
    Alketi
    Some addition details on the MobileEye camera from George Hotz, who's famously not a fan of MobileEye.

    He said the MobileEye camera is an RCC digital camera, meaning RED-CLEAR-CLEAR. Implying it's a grayscale camera that only sees the luminosity in the RED channel. He stated that THIS was the reason why it couldn't distinguish a white truck against a bright blue sky. The camera in comma.ai's system will be a full color camera, and I would think Tesla may use a color camera in the future as well.

    The interview is worth watching if you're a fan of autopilot and technology in general.

    �
  • Jul 26, 2016
    Vitold
    Except that they did not - you're quoting bizjournals who is quoting marketwatch. If you go to the original source says that both companies did not see a way forward and both did not want to say much about dissolution of their partnership. Marketwatch as well other sources note however that Tesla was not happy with Mobileye.
    �
  • Jul 26, 2016
    Tam
    Not much but Elon Musk seemed to comment about Mobileye's ability as well as its speed to keep up with Tesla's pace as mentioned on a duplicated thread at:

    Tesla Ends Relationship with Mobile Eye
    �
  • Jul 26, 2016
    Cosmacelf
    Yes, that bizjournal article made an incorrect assumption that Mobileye was the one terminating the partnership. I suspect it was much more mutual.
    �
  • Jul 26, 2016
    theslimshadyist
  • Jul 26, 2016
    msnow
    This will alter people's expectations of AP 2.0 from early next year to late next year IMHO.
    �
  • Jul 26, 2016
    theslimshadyist
    My thoughts exactly.
    �
  • Jul 26, 2016
    brkaus
  • Jul 26, 2016
    ElectricLove
    They may be jumping onto the Cruise Automation system! Fully autonomous vehicles are already on the road operated by Cruise in the Phoenix area. GM owns Cruise, purchased for $1B in March of this year!

    The founder is the same kid that challenged Elon 2 years ago that his system was better and Elon said to him that if he proves it then he'd be seeing TESLAs business immediately! I'd say GM spending $1B on them may just be the "proving it" part...
    �
  • Jul 26, 2016
    JohnSnowNW
    You're thinking of George Hotz...so different guy.
    �
  • Jul 26, 2016
    AZ Desert Driver
    Based on Google news:
    In the wake of a fatal crash, Tesla will quit using Mobileye's chips for Autopilot vision
    The car company will likely work on its own image-recognition technology.
    BY JOHANA BHUIYAN @JMBOOYAH JUL 26, 2016, 2:23P
    Weeks after a fatal Tesla accident involving the company�s semi-autonomous tech sparked probes from federal agencies, Tesla parted ways with Mobileye, the startup that supplies the chips that enable Teslas to recognize images and to work with the car�s Autopilot system.

    This also means Telsa is likely working on its own chips to improve the semi-autonomous features.

    Mobileye CTO Ammon Sashua made the announcement during a call discussing the company�s second-quarter financial results.

    �We continue to support and maintain the current Tesla Autopilot product plans. This includes a significant upgrade of several functions that affect both the ability to respond to crash avoidance and to optimize auto-steering in the near term, without any hardware updates,� the company said in a statement. �Nevertheless, in our view, moving toward more advanced autonomy is a paradigm shift both in terms of function complexity and the need to ensure an extremely high level of safety. There is much at stake here, to Mobileye�s reputation and to the industry at large. Mobileye believes that achieving this objective requires partnerships that go beyond the typical OEM / supplier relationship, such as our recently announced collaboration with BMW and Intel. Mobileye will continue to pursue similar such relationships.�

    In the hours after the announcement, Mobileye�s stock dropped more than 7 percent.

    When asked why the two companies parted ways, Sashua pointed to the companies� respective responses to the fatal accident. Mobileye�s technology is only capable of helping to avoid accidents with cars in front of it, not trucks crossing the highway laterally, as was the case in this accident.

    �This incident involved a laterally crossing vehicle, which current-generation AEB systems are not designed to actuate upon,� the company wrote.

    Tesla said Autopilot, which combines proprietary and third-party technology, is supposed to be able to recognize �any interruption of the ground plane in the path of the vehicle� but �the high, white side of the box truck, combined with a radar signature that would have looked very similar to an overhead sign, caused automatic braking not to fire.�

    While it�s not clear which party ended the relationship, Comma.ai CEO George Hotz � whom Tesla CEO Elon Musk once challenged to build better software than Mobileye�s � said: �Without a doubt it�s Tesla who parted ways.�

    Hotz, who also challenged Musk to a race of their respective semi-autonomous cars when his was done, said Mobileye isn�t innovative enough for Tesla.

    �Their stated mission is to lower the safety rating of cars that don�t have their technology,� he told Recode. �So when you have a company like that versus a company like Tesla who actually wants to build incredible cars, there�s absolutely no reason Tesla needs Mobileye.�

    It�s likely Tesla � which recently tapped former Advanced Micro Devices chip engineer Jim Kellerto be its vice president of Autopilot � will begin working in-house on the technology for which it once depended on Mobileye. And according to Hotz, it won�t be hard.

    �Tesla has a lot of trouble attracting machine-learning talent. They don�t have a cafeteria, for example. So people are going to Google and Apple,� he said. �But to be honest, the Mobileye system is so easy to reproduce; you don�t need the best talent. I did it in a couple of months. Our software can already do more than Mobileye�s.�

    According to the company, Mobileye will continue to pursue partnerships like its recently announced deals with BMW and Intel to create autonomous cars. BMW has yet to ship any semi-autonomous tech. But Hotz says its proclivity for working with established players is exactly why he thinks Mobileye isn�t concerned with innovating.

    �I met with some automakers and all they seemed to be about was their brand,� Hotz said. �They say, �Well, no people buy � insert crappy car Y here � because their parents had crappy car Y. That�s not what sells cars anymore. What sells cars really is the car experience. When Apple comes out with their car and even the Tesla Model 3, [automakers] will see they�re not in the same ballpark. And then a company like Mobileye wants to work with these established players.�

    �It�s like you see the same thing in mapping � all the losers are trying to group up together,� he said.
    �
  • Jul 26, 2016
    bhzmark
    yes whatever mobileye used to train their DNN, Tesla has much better data now with so many cars on the road. they could build a much better DNN with that data that no one can touch.
    �
  • Jul 26, 2016
    Tam
    It sounded credible until the cafeteria is mentioned as discussed on the twin thread:

    Tesla Ends Relationship with Mobile Eye
    �
  • Jul 26, 2016
    Tam
    Current Mobileye system is too dependent on road markings.

    As shown in youtube by another thread, the Autopilot was centering the lane fine until there's a break of lane markings coming up.

    Look at the left red mark that locates the left yellow lane marking before the road marks are erased (with unmarked fresh concrete):

    [?IMG]

    After the Autopilot encountered the missing lane markings and passed over them, look at the left yellow lane marking and the left red mark, Autopilot then veered toward the right and hit the upcoming right traffic cones.


    [?IMG]


    Other system like Nvidia can learn how to steer even in the absence of road markings.

    It can understand the rules of engagement between vehicle and road.
    �
  • Jul 27, 2016
    182RG
    Sometimes, the little things matter, especially to Millennials. Don't discount free gourmet food.

    It helps that Google and Apple are wildly profitable and stable. Tesla, not so much.
    �
  • Jul 27, 2016
    theslimshadyist
    Ooops, sorry. I have a bad habit of only staying in the MS forums.
    �
  • Jul 27, 2016
    voyager
    Is Verizon perhaps a good match for Tesla?
    �
  • Jul 27, 2016
    golfingBuddha
    The little things matter to everyone. Do you know how many 60 year old doctors I have seen leave the comfort of their home to go get free food at the hospital cafeteria?

    Everyone likes to feel appreciated and if you commit to working on Tesla's autopilot team you are likely committing to 70-90 hour work weeks. You will want to be very comfortable at work and the lack of a comfortable environment which you will spend the next years of your life in would be a big deal to anyone. :-D

    George is a smart guy but probably not the best at articulating things. The broad point he was trying to make is valid though. What he said about recreating MobilEye's software in a few months is also true. It is not absurdly complicated.
    �
  • Jul 27, 2016
    Krugerrand
    Free food wasn't mentioned in the article, but Tesla does have a cafeteria and they do offer coffee, milk, soda, juice etc... for free along with things like cereal. Nope, not gourmet cafeteria food (if one could ever call cafeteria food 'gourmet') but let's at least get the facts right.
    �
  • Jul 27, 2016
    zambono
    There won't be any new AP hardware till at least the release of the 3, and that doesn't even mean it will be activated. There is a chance that the model 3 will be released with the current hardware and they will try to push it to the limits.
    �
    • Người đăng: vào lúc

    Không có nhận xét nào:

    Đăng nhận xét

    Đăng ký: Đăng Nhận xét (Atom)