sleepydoc
Well-Known Member
Like he says at the end of the video, it's pretty clear DoD went out of his way to make the car fail in his video. If that's the case Tesla likely would have a basis to sue for libel.
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Will Tesla sue Dan O'Dowd?
- Thread starter Daniel in SD
- Start date
I'd like to consider this test, but right from the beginning when the car does not identify the dummy he adds a hat and adjusts the model to try to improve detection. It still does not identify it with FSD Beta disengaged. However when he engages FSD Beta it does identify it. All I've seen so far is an attempt to game the test, he never tests the dummy without hat again.
Question:
Why did FSD Beta=off not see the child dummy at all but FSD Beta=on did see the child dummy?
For the later tests he adjusts the position of the adult dummy between low speed test and high speed tests. To improve the test result? Who knows, but his statement from the start of the video is that he considers FSD Beta is safe, so his test methods can be considered biased.
But for what he's shown, it's interesting nonetheless.
Last edited:
EVNow
Well-Known Member
If that was the idea - there was no reason to show that initially it (non-FSD version) wouldn't recognize the small dummy.
ps : We could also ask - how many trials did Dan & Co do before getting a dummy that they could hit (apparently by pressing on the accelerator).
Well exactly. It's almost like things posted on the internet may not always be true. I don't believe either of them.
sleepydoc
Well-Known Member
If he were trying to be biased he wouldn’t have shown the first tests at all. I agree it would have been interesting for him to try the tests again without the hat (I doubt that actually made a difference) but that doesn’t negate his tests.
Well sure it negates his tests. He showed he improved the dummy from the initial conditions and achieved a desired result. He said at the outset that he considers FSD Beta to be safe, and lo he showed us the videos where the car didn't hit the dummy.
Apart from the entertainment value of seeing the car drive around the dummy and hit some cones all we learned was that he was prepared to alter the tests and perhaps selectively edit the videos. Far from blind or double-blind, this is a biased test. Zero scientific value, and proof only of him showing us what he wanted to show us.
I'm not saying FSD Beta did not avoid the dummy in the demonstration, just that this has no bearing on overall FSD Beta safety, and hardly satisfies his video title's claim "Is Tesla's FSD Beta Software SAFE? This video Settles it Once and for All - v10.12.2"
I don't understand the controversy over "Is FSD safe?" Tesla says FSD is not safe by itself, and that there must be an alert driver and that driver must intervene when necessary, even without any warning from the computer.
FSD is at present an unsafe development system that requires constant supervision and occasional intervention. This puts it in the class of all cars, which are unsafe if not operated properly. Operated properly, most cars are relatively safe, but any car can be in an accident due to mechanical failure or driver error.
Why is there any discussion?
FSD is at present an unsafe development system that requires constant supervision and occasional intervention. This puts it in the class of all cars, which are unsafe if not operated properly. Operated properly, most cars are relatively safe, but any car can be in an accident due to mechanical failure or driver error.
Why is there any discussion?
mainly due to lack of consistent messaging on the topic from the company. We as Tesla enthusiasts know all of the info, details, real life reality about it, etc.
But to the average consumer, one can see how it can be confusing messaging when they hear the name of the product, what their definition in their minds of what "FULL SELF driving" means to them, what they interpret to mean by messages of "NYC to LA with NO human input" and "will drop you off and park itself" and "will reach level 5 by end of 2021", etc.
Those messages typically get more traction in the press/public than the wording customers accept in the user agreements when you select "enable FSD" on your screen that basically conflicts with those aforementioned statements.
AlanSubie4Life
Efficiency Obsessed Member
I’ve tried this argument. It doesn’t really work, and people don’t even address it. I find that to be odd but it probably makes sense somehow.
Anyway, FSD is entertaining, has a long way to go, and you are driving the car when you use it, and are responsible for hitting things, just like you always are when driving a car. The car is not to blame - it is not designed to avoid all objects and should never be expected to do so.
Obviously Tesla will not sue - this video has done them a great service.
EVNow
Well-Known Member
Because Dan is trying to get some cheap publicity by disparaging Tesla. FSD is just a useful kludge.
Tam
Well-Known Member
When I bought FSD, I was in denial, and I rationalized that Tesla wouldn't sell anything unsafe.
But FSD beta may be safe. 35M miles without a severe collision is quite good.
Of course wide release may show that it’s not safe in the hands of people who can’t or don’t want to achieve a 95+ safety score (beta).
EVNow
Well-Known Member
Wait .... when you bought FSD, there was no FSD Beta. So, you bought something that wasn't available.
If you wait for production release of FSD (or atleast fleet-wide release of FSD Beta), it will not be "unsafe".
But all this doesn't matter - because Dan is not making any kind of arguments in good faith. You can't reason with Alex Joneses of this world ...
Tam
Well-Known Member
I thought "wide release" meant very narrowly released to the existing 100,000 beta testers and not to the general population who bought FSD without being selected into the beta tester pool.
Tesla might have different meanings for "soon, this year, full, self, wide..."
Last edited:
There are as many definitions of wide release as there are of FSD itself. I’m this context wise release means beta to everyone who bought FSD.
The car is not unsafe. It's extremely safe when driven as intended: By a driver who understands that they are fully responsible, and who drives in a safe manner.
Of course, and I've said this many times for a long time, it was unconscionable of Tesla to sell a product that did not yet exist, and to claim that their cars had all the needed hardware, when it was impossible to know what hardware would be needed before the software existed. And it's unconscionable of them to use the name "FSD" for a product that is still just Level 2 driver assist. And it borders on criminal that they quietly changed the definition of "FSD" from "driverless car" to "Level 2 driver assist" without ever acknowledging the 180° turn-around in what they were promising.
And it's incomprehensible to me why they did all this when they already had the best cars a consumer can buy. And the safest, when driven responsibly.
Typos!
In this context wide release means beta to everyone who bought FSD except those with MCU1, they'll have to wait until non-beta wide release and visualizations are no longer necessary to safely test the software.
Similar threads
- Article
