-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
I don't own stock in Tesla, I've not sold stock I don't own either, but it's not difficult to spot those that do have stock as they defend the indefensible. If you want to know what pathetic is, it's to throw the shorting line as the only argument you have left.
I have Tesla stock and I have said critical things about their new software. People's safety is more important than my share values. It just happens by coincidence that lesser fatalities in the Tesla fleet will improve the overall outlook of the stock.
People make money where they can but shorting Tesla stock is not good natured with what all of us are trying to accomplish.
I also don't own any stock in TSLA and am not shorting it. I'm just not convinced that missing their expected timeline for EAP by, so far, less than a week is the catastrophic betrayal many others are portraying it as.
Nor is it appropriate to blow up the thread discussing the rollout of some of the features(along with every single other thread even remotely related to the rollout) with complaints about how they sold the software or the timelines.
stopcrazypp
Well-Known Member
Ditto, I feel sorry to the others that this thread got hijacked (and I participated) and went on a tangent on other things that have discussed at length since the release of autopilot 1.0 (or even when Tesla decided to do OTA updates). Perhaps that discussion should be moved elsewhere and this thread be kept solely about the release of AP 2.0 and its features, which is what most people were expecting coming into this thread.
stopcrazypp
Well-Known Member
Didn't I hammer home the point that Tesla does just as much in-house and closed beta testing as other automakers and that the product is just as "ready" as any other semi-autonomous product out there (and in fact is more reliable when talking about the autosteer component you complained about)?
As for the claim that Tesla is the only automaker to use fleet learning, that is false (as is your claim that they are using it to release unready products). Toyota announced that they are doing so during the end of 2015:
Toyota to Display New Map Generation System at CES 2016 | Corporate
In 2015, Audi, BMW, Daimler also bought out a company that does the same thing:
Highlights of 2015 | Daimler > Annual Report 2015
In September 2016, they announced that their cars would be gathering such data:
Audi, BMW and Mercedes cars to communicate via data sharing network | Autocar
Mercedes had a demo vehicle that did this also in 2013:
High-precision maps for self-driving cars.
The fleet learning right now that Tesla is using is the same method to generate high res maps. This is the only way to gather enough updated road data to make semi-autonomous and autonomous vehicles operate reliably (even with faded lines), which is why all the large automakers are looking to do the same. Driving with only manufacturer vehicles is not enough and will leave you with outdated data even if the test driver was able to get full coverage (Mercedes article linked above pointed that out).
thegruf
Active Member
But is it fair to say Tesla were first to actually implement fleet learning?
... and in fact have they shown the ROW the way (oops = slightly messianic there!)
(I) You really have no way of knowing that Tesla does as much in house and closed beta testing as others. Note that, for example, Google's been doing in-house testing for more than five years.
(II) Even if Tesla is doing as much in-house testing as others, that shouldn't really be the goal. At a minimum, they should be doing enough in-house work so that, before a release is made to consumers. the obvious problems are solved (think truck lust and keeping in a right lane when there is an exit) and so are any other problems of which they become aware after reasonable testing.
(iii) I'm not giving any opinion about whether other company's auto-steering systems are ready for consumers. Maybe they aren't either (in which case I'd oppose them too). But it seems pretty clear to me that Tesla is doing consumer releases before the software sufficiently works.
Joe F
Disruption is hard.
I have no problem with that. They just need to collect the data from the installed sensors on AP2 cars. No need to turn on flaky AP2 software in order to build a map.
Craptastic. Another update email from Tesla ServiceHelpNA. I guess no cruise control of any kind for the foreseeable future. I inquired if I can roll back to 2.50.180 but I'm not hopeful.
"During the initial roll out, we have identified conditions that may prevent some cars from successfully completing their calibration. If you have driven your car for over 6 hours and the calibration is still in process, it is likely you will need a new software release before you can use the new features. We are working on an update that will resolve this but do not have an exact ETA on when it will be released at this time."
I was asked if I wanted to be part of EAP (Early Access Program) since I am a SpaceX employee. If that means installing pre-beta software on my car, judging by my experience with 2.50.185, I think I'm going to decline.
Ah, that explains my >9 hours with no calibration complete. Hopefully they get that fixed fast. I'm with you on being annoyed at losing cruise control.
Needless to say, I'm no longer so happy about having received the update.
Bet TSLA
Active Member
I don't understand this. Your situation specifically helped them identify a bug, which they'll fix, and it will be in the next attempt. What's not to like? You can complain that they shouldn't have had any bugs, but that's never true. You can complain that some other poor bastard should have gotten stuck with a failure to calibrate and it should have been fixed then, but maybe your situation is new and unusual.
So sorry, this is just how it goes when you're trying to move fast. People who object to this kind of thing need to realize that what they are asking for is that Tesla's software, with all its features, get released more slowly. That's right, you want to AP 2.0 software a year from now instead of a couple months. Yeah, didn't think so. You want it soon, you get it rough and it improves over time.
I think the complaint is that before the camera is calibrated you should still have regular non-Traffic Aware cruise control available. (Even if it meant going in and turning off the AP features so that there is no question about if you are going to enable CC or TACC.
First, I'll note that I said I'm no longer so happy, not that I'm angry or anything. I was ecstatic that I was one of the 1000, but for obvious reasons less so when I discovered that I didn't actually get any of the new features because of said bug.
Second, if I were to be upset about anything it'd be that the software disables normal cruise control until camera calibration is complete. It's a little baffling that people at Tesla saw that calibration was required and figured people would be fine with driving several hours without even the basic cruise control that was there before.
That aside, I knew this is what I was getting into by buying a Model S and am not upset.
ZeroDarkSilver
HEPA cleaning your atmosphere, you're welcome
Someone from your branch of the Elon Empire would know better than anyone else, the risk that with early attempts at launching stuff you may experience some slight turbulence and then - explode.
3Victoria
Active Member
It is pretty easy to surmise that it is not possible to have normal cruise control while calibrating the camera because the two are incompatible. Its not the end of the world that you do not have it. I think that people who are upset by these things (not you) should probably not accept the earlier version of the software. One needs to be a bit of a nerd type with a sense of adventure, and want to help out the effort to be a suitable candidate for this. In the past Tesla had a set of people who tested new stuff ... the supply of people is likely more limited with the new hw, since the total number of so equipped cars is less. This is a bit of a chicken-and-egg problem for Tesla.
Similar threads
