Sorry. Eight and a half inches (George Stephenson)
Don't I wish...
You can install our site as a web app on your iOS device by utilizing the Add to Home Screen feature in Safari. Please see this thread for more details on this.
Note: This feature may not be available in some browsers.
Sorry. Eight and a half inches (George Stephenson)
I see what you did there...Don't I wish...
They're most likely not even running the chips at full potential performance yet as usually the first couple of iterations of your design have issues with yield, and it takes a couple of respins to nail it. Though I wouldn't expect a doubling or anything here, just another 5-20%.
As for switching from ARM to x86, I wouldn't bet on it. Maybe from ARM (cheap licensing) to RISC-V (zero cost licensing), but going to x86 doesn't make a lot of sense when you're already building custom silicon. They can put the ARM or RISC-V core(s) on the ASIC and eliminate both the Parker SoC and Pascal GPU, but if they go x86 they still must have two separate devices (or pay even more money for custom silicon from AMD with NN logic embedded into an x86 SoC - not impossible but likely not the optimal solution here). Plus most likely ARM or RISC-V have more than enough performance for the task of keeping the NNs fed and operating, x86 is overkill.
The x86 CPU on the newer AP HW cars I believe is running the UI/UX, not the AP decision making. I think that's still happening on the Parker SoC. It's possible that Parker isn't enough for FSD decision making, but that doesn't mean they would move to x86, you can get higher performing ARM cores and they may move that into the NN ASIC itself anyways.
Don't assume anything from Nvidia presentations of DrivePX and related systems to have any bearing on how Tesla is doing things, that only applies to using Nvidia's software/hardware solutions and even though Tesla is using Nvidia hardware currently, they're doing it their own way. Since the screens run on separate SoCs they can potentially render all the feedback there and not on the AP SoC, so the AP SoC doesn't have any need for any GPU resources at all (Parker has a small GPU in it) if the image input processing functionality can be handled in the ASIC.The arm core will be needed to handle normal looping and higher function logics (the basics needed to run a firmware) like calling a function, memory addressing etc. So some type of general purpose core has to be there. From what I've read of the Nvidia presentation, the Parker SoC is used for displaying what the autopilot is seeing as well as these higher functions. When the NN is at its infancy, the display of what the NN is doing is important. The rest of the discreet GPU is most likely used for preprocessing the image to find objects, orientation, angles with the tensor cores programmed for the Neural Net decision making.
Tsla can go ahead and engineer their own general purpose cpu core, Peter certainly has the know how to do it since he comes from the Apple chip design team. However, that's another 2~3 years of testing and a lot more complicated than the Neural network chip. I wouldn't be surprised if they just stick to licensing Arm core inside their ASIC as a black box and later swap it out with their CPU design once the design team is done with the Neural Network chip.
And I agree x86 is just overkill for this. Maybe for the infotainment system so people can play games.
Makes me think they should create a subsidiary to partner with Nvidia to build specialized chips and code for driving. It would be a more bet mill. They could keep it a private 50/50 partnership or potentially sell a partial stake to invest back in the business. This could be worth more than Mobileye and Tesla would still have a lead with system integration. This could provide a market value of fsd and let Tesla focus on their other market advantage in power supply, motors, controls and DESIGN!Reply in a different thread:
Tesla Autonomous Driving H/W
I think Elon was pissed about covering the flying Tesla. Enough people covering every accident and idiot driver.umm. who cares?
and no, we don’t think it would be good for tesla to go through a bankruptcy like gm did. how about you? maybe you should go through a bankruptcy. do you think that would be good? great idea. we should all try int.
stop trying to spread your nonsense propaganda
TSLA volume continued its downward trend today with 5.25 M shares traded. Interday SP range also narrowed slightly to $6.71 or 2.25% of the low. Summary stats in the table below cover period May-07-2015 to today. Data from NASDAQ.
View attachment 330826
Do you want the ironic answer?Hmm, now if we could only organize and get tax exempt status..
As someone else noted several pages up, it seems whenever there isn't any big news for a few days, there's a slow drop. Any serious explanation for this? People just being nervous and impatient? Shorts pushing price down?
Disagree, but it could solve cash flow problems and create a cash cow subsidiary that would be worth more then the parent company is worth today.Makes me think they should create a subsidiary to partner with Nvidia to build specialized chips and code for driving. It would be a more bet mill. They could keep it a private 50/50 partnership or potentially sell a partial stake to invest back in the business. This could be worth more than Mobileye and Tesla would still have a lead with system integration. This could provide a market value of fsd and let Tesla focus on their other market advantage in power supply, motors, controls and DESIGN!
Well, there was a point where someone said "free upgrade" but no detail about what that required. Was is EAP or EAP and FSD (already paid for) or EAP plus FSD (not paid at purchase but paid for after the fact). Or if one has EAP and now buys FSD, they will put in the new hardware for free as well. That's what I'm referring to hasn't been confirmed. If there is a link with absolute definition I'd love to read it. Thanks.
If you are young, are well paid, have little to no debt or other obligations (college for kids) then I agree that going 'all in' on one stock/company that you believe in, while risky, is certainly a strategy that can make you incredibly wealthy, though with great risk.
Not lecturing here, just indicating that we all have different risk/reward/obligations. If you are happy, I am happy for you.
Young, no - 52 - but not exactly old either.