Here's a recent post (among MANY) from a Tesla owner about his experience with "Autopilot"

The responses to this show this to be normal.

Tesla can't even safely take highway off ramps or notice barriers it's about to slam into when re-merging onto a highway. Meanwhile, GM and Waymo have full autonomy on any road in the city in complex conditions including obeying traffic lights, four way stop signs with traffic, avoiding pedestrians and double parked cars, no lines, construction zones, bikes, etc etc, with one human intervention in 3000 miles. An example from over a year ago:



That you think the research report is "paid for" and that Tesla might be ahead is ****ing ridiculous, Spurious. Do you not realize you're ridiculous here? What planet do you live on?

Here's another recent video of Tesla completely ****ing up when lanes disappear. It lacks any competent mapping of its environment and just blindly follows lanes and the car in front (and will sometimes swerve into oncoming traffic, confusing it for a car it's supposed to follow). It doesn't even warn when lanes are lost. Tesla are years behind the majors and have no viable path for going beyond level 2.



Tell me when you concede that Tesla are far behind and I'll stop. There are dozens of these videos. TMC forums are full of Tesla owners complaining about how dangerous and unsophisticated autopilot is anywhere outside of very well marked freeways.

Waymo and Cruise Automation also have trouble with disappearing lanes. It's probably the #1 reason for disengagements.

Also, that doesn't seem like that twitter video driver was driving on a highway.

Please. Tesla would disengage every 5 miles in the city roads that Waymo and GM drive on, at least, every 0.5 miles in the inner city, for which Tesla lacks anything at all. Tesla can't even handle the simplicity of highway off ramps. Meanwhile, Waymo is making fabulous progress:



In November they had it down to 1 disengagement in 30K miles navigating complex and busy city traffic, often without lanes. The idea that Tesla are anything but years behind Waymo and GM is ****ing absurd, and anyone who says otherwise should simply be mocked and ignored, because they have no clue what they're talking about.

GM Cruise has similar declining disengagement numbers, 3 and 4 for October and November 2017. And three of those four are "other road user behaving poorly", e.g. an imminent crash where the other user was unavoidably at fault. Disengagements that are an autonomous driving fail are zero or one out of 30K miles in later months.

I never said Tesla is even close to the level of sophistication of Waymo and Cruise Automation.

I merely said disappearing lanes are a problem for everyone. Comma ai ran into this problem in Nevada highways. You're essentially forcing the autonomous software to guess with disappearing lanes, and putting your car into the hands of machine learning.

Tesla would disengage way more often than Waymo and Cruise, complete agreement. Tesla isn't even meant to be used off the highway. You agree that twitter video you posted is probably not the highway, right?

TS,

I am not making up a conspiracy theory. I am stating the fact that Navigant primarily makes money from selling services. You don't understand this simple fact and rely so heavily on this that you'll end up being wrong.

I am not defending Tesla's AP, I attack your low quality posting.

Also, GM and Waymo have most likely problems as well. Waymo has been involved in a number of accidents where they allegedly weren't responsible but they had quite a few rear end accidents. GM has one video and you take it as the absolute truth. This is so embarrassing. Tesla's AP needs to navigate in a variety of situations whereas GM and Waymo restricts the area in which they operate. This vastly reduces complexity.

Here are a few examples:
https://www.reuters.com/article/auto...-idUSL2N1MF1RO


https://www.wired.com/story/ride-gen...f-driving-car/

The big difference is that GM is underselling super cruise while Tesla is overhyping autopilot to the level where they endanger lives.

This is total bull**** man. I post credible research from a billion dollar research company on the state of autonomous driving, showing Tesla dead last as every data point backs up, you come back with this horse ****:
Here, you imply the study showing Tesla dead last and GM and Waymo well ahead is invalid, because they are paid by "automotive companies" - a false claim, by the way, pure conspiracy. Navigant are a research company who publishes research across a large variety of industries, which is then sold to interested parties - media outlets, researchers, hedge funds, industry participants (like parts suppliers).

Now you backtrack from that absurd conspiracy theory claim (the report is suspect/wrong because it's paid for by autonotive companies), lying about what you said. How much more dishonest can you be? You should have stopped while you were behind.

You are though. Any criticism of anything to do with Tesla you bend into idiotic conspiracy theories for why Tesla aren't behind in spite of the clear facts. Any evidence of Tesla's horrendous, far behind state of autopilot - which is everywhere - you try to deflect. I mean, right here, you're trying to deflect again:

Dude, we have hard data on disengagements and incidents. They are incredibly low. I posted the table above.

No, dude. This is you being a dishonest, bad faith piece of **** again. I posted data on disengagements. I pointed out that they have a city wide fully autonomous ride hailing service actually running right now. I posted a study showing them far ahead of Tesla, along with most of the other majors. You rubbish and try to deflect every single data point, because you are a deranged cultist. Incredibly, you seem not to realize this.

This is like talking to someone on Yahoo! stock forums. I've rarely encountered such mouth breathing stupidity on 2p2. Congratulations on bringing down the quality of discussion on 2p2.

Tesla's AP only works on well laned freeways with on and off ramps. Waymo is navigating inner city traffic, pedestrians, bikes, cross traffic, traffic lights, four way stop signs, no lanes, etc etc.

Did you even read the article you linked? These accidents were the fault of other drivers, entirely.

How much of a disingenuous cult inductee are you? Here are the facts:

- Tesla autopilot is terrible and in a poor state of development. They are dead last of all major car companies in their level of developmentr, a fact which is shown by multiple data points. They are struggling with creating a safe level 2 and have massively oversold the capabilities of their "autopilot", as well as lied about the time frame (Musk promised Fully Self Driving coming in "maybe 3 months, definitely 6" in January 2017. This irresponsibility has killed people, including recently.

- GM and Waymo are far ahead and have a highly competent level 4 that on latest data does one disengagement every 10K miles - most of those for the faults of other drivers - and at most 1 in every 30K miles that are their own fault in the last few months.

That you can't accept these facts shows how unhinged you are. You're here to defend your cult leader with whom you are in love, and provide nothing of worth to anyone.

TS,

of course you never have accidents which are your fault if you drive ultra slow and abruptly brake. This is not a good argument. I read the article and that's why I linked it. I would have explicitly stated if it was their fault.

The more often you repeat multi billion dollar research company, the less I believe you that you yourself believe their stuff.

chytry,

I absolutely agree. I am only raising the point because TS is actually doing the opposite of what GM is doing.

I find it amusing there is "debate" on whether Autopilot (or other AES systems) prevent crashes.

https://static.nhtsa.gov/odi/inv/201...16007-7876.PDF

Guess there is a loose grasp on reality here.

exactly

None for nothing, if the system needs continuous user input, maybe calling it "autopilot" was a bit of an oversight.

There are three separate things here:

1. What Tesla classifies itself as in regulatory documents (level 2 - nothing but lane keeping and following and changing)

2. What Tesla represents and hypes the state of its "autonomous" driving as to customer, investors and the general public.

3. What Tesla claims about the state of development and what is coming (for example, "Fully Self Driving" hardware sales with the features coming in "3 to 6 months" in Jan 2017

The first is reality. The second and third are pure bull**** that Tesla is selling and people like Spurious are buying. Tesla has level 2 and nothing beyond that and little hope of going beyond it, since they can't even do level 2 + highway obstacle avoidance properly.

The tragedy comes when people don't use "autopilot" properly because Tesla has downloaded pure bull**** into their mind about the state of development, how safe and capable it is.

I don't know man, a brief perusal of the Model S owner's manual makes Autopilot's limitations very clear. Just look at page 65 and on. There are many very clear warnings of Autopilot's capabilities and limitations. I understand you have problems reading, but even you can probably understand.


https://www.tesla.com/sites/default/...rica_en_ca.pdf

So apparently the fact that its called "Autopilot" means everyone is a total idiot now and Musk is to blame. I guess we need larger warnings on toasters explaining not to use them in the bathtub for these same people.

I guess from a guy who's hung his hat on a fake trading record for years. (Has he even made a single trade in his prop bet?) It isn't really a whole lot to expect this sort of dishonesty. Apparently every expert is only an expert when they say Tesla is bad, never when they acknowledged they had the premier SDC technology deployed. Or the actual regulatory bodies.

Using something called "Autopilot" as an autopilot is a wee bit different than using a toaster as a bath warmer.

Just sayin', "requires constant user input" is literally the exact opposite of "autopilot." Seems like after the first guy died because he mistakenly thought the autopilot was an autopilot, the company should have considered renaming the feature. I get that the manual clearly explains that the autopilot is not to ever be used as an autopilot and will drive right into the nearest truck if it's on an unapproved road, but still you can appreciate why people get confused.

Of course the manual is clear. They're covering their ass like crazy. They do the same in their SEC filings, which use much more muted language than Musk does on Twitter or PR releases when say, talking about production targets or the date when fully self driving is coming.

How many people read a car's manual in full? Any? The fact is that the minds of most people, including drivers, are filled with the hype that Tesla puts out.

That's why Tesla drivers are watching Harry Potter and getting decapitated rather than the road. They've been sold bull**** rather than being told the system is incredibly dangerous and non-functioning outside of following the car in front and keeping in a lane in a well marked area.

I understand you have trouble following human nature and arguments, but this isn't hard, man.

Show me an example of this Autopilot hype you're claiming.

Airliner auto pilots don’t do collision avoidance either. Keeping Airways safe requires constant monitoring by air traffic controllers. Just sayin.

Naming the thing Autopilot was probably more a marketing move and they didn't anticipate people needing to put the hands on the steering wheel all that often. But people who can afford a $100+k car should probably be trusted to understand that Autopilot does not mean machine chauffeur.

Regardless of the name, turning on autopilot for highway driving should perform well enough to not slam into lane dividers.

https://arstechnica.com/cars/2018/04...viders-poorly/

I can show you dozens. The below are all from Tesla's autopilot promotion page.

How could anyone possibly get the wrong idea that it's safer to just let Tesla do the driving (my yellow highlight)?



How could anyone possibly get the wrong idea that Tesla isn't in fact capable of detecting fire trucks or concrete barriers and either giving ample warning or stopping itself?
How could anyone possibly get the wrong idea that Tesla can handle all aspects of highway driving without slamming into anything?
How could anyone possibly get the wrong idea that Tesla isn't merely an incompetent level 2 software with unreliable safety features?

The only disclaimer on their entire promotional page:

How could anyone possibly get the wrong idea and think that Tesla has their back in multiple ways on avoiding fire trucks in your lane, trucks blocking the entire highway, and lethal concrete divider crashes?

I think after reading that your average buyer is going to think that "Autopilot" can reliably find and warn about obstacles, follow lanes properly (and not into concrete dividers or fire trucks), and be safer than they are on highways. Tesla claims as "level 2" for legal purposes and then hypes the **** out of it, claiming far more and far more reliability than is justified for mere lane changing and leader following.

That's without getting into Musk's tweet hype or public statements or interviews/articles from friendly news outlet hyping their "autopilot"

Amazing that you guys are arguing with me on this. You're dead wrong and really stupid besides.

No ****, right? Not to deranged Musk fanboys though. It says it all when you have Spurious and SenorKeed in one corner.

In other news, Tesla at $305 now after being at $260 a few days ago.

Nice short squeeze and could even have a little more to run if there isn't bad news, $320 isn't out of play. Market helped a little but this 16% run is mostly bullish bias + short covering.

When the biggest bear is selling puts on a terrible quarter, you don't stay short. I would look to reshort/buy puts up here if any bad news comes out or the market turns, but mainly for a scalp. The carrot is in front of the donkey again and the market run off lows was the perfect timing for worsening the short squeeze.