Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

CNN denouncing FSD and Tesla

This site may earn commission on affiliate links.
Not sure how they can say it indiscriminately mows down children, all the children in the video are white.

Yeah. Though mostly they were covered up, so maybe it was profiling the choice of clothes.
Seems like a pretty big failure of FSD Beta 10.12 (more details: The Dangers of Tesla's Full Self-Driving Software - The Dawn Project )

But a little detail: slowed from 38 or so to 25 (and perhaps 18 at time of impact) with the first collision (they claim 25 but I am not sure about that). And was telling driver to take control (too late). Of course the audio from the car was not included in the videos.

They seemed unable to afford a proper in-car video mount, or a camera with sufficient depth of field to keep both the screen and the distance reasonably in focus. Makes it hard to see what is happening. And no audio as mentioned.

It veered to the side on the last two tests.

I wonder how they had FCW set (not sure if it would matter for in-cabin alerts, but presumably would not change the driving behavior).

The setup with cones seems a bit non-optimal to me. Why not do it with standard road markings? Not that this is an excuse - FSD Beta 10.12 should not be plowing into obvious obstacles. I just wonder (given that it seemed confused prior to the impact) whether it’s actually a great test and may actually be uncovering some sort of poorly-coded corner case.

To be clear, I suspect it would hit the mannequin with standard road markings, unless they were in a crosswalk.

But I do wonder what is wrong with the VRU identification here. I’m slightly surprised it fails so badly. I’m not at all surprised it hit the VRU though.

I wonder how many tests they did?
 
If its an ad, how is it any different from political ads that stations HAVE to accept?
If its an ad, how are you equating an ad to journalism?

We need answers..
As I have worked in the broadcast industry for 58 years, I am reasonably familiar with paid ads vs programming material.
I only caught the last 75% or so of this tripe and could not tell whether it was a paid ad or something they thought was "news". There was a lower third graphic with an 800 number to call and some websote, but I was not in a position to jot either down.

A broadcast station can indeed refuse to air any ad they deem not to air. However, comma, there are a number of regulations under the Fair Standards Act with which licensed broadcasters must comply. a station can refuse to air any political ads, but must then refuse to air all ads regarding a particulat subject or candidate.

Cable TV networks, however, are a different kettle of fish. They have certain regulations with which they must comply, but not all fall into lockstep with the FCC rules and regs.

For all I know they may be "reporting" a story about someone running this crap, in defense of Tesla. Knowing CNN's track record, however, I highly doubt it. I've asked an associate who works for CNN to see if he he can track it down.

I only brought it up to see if anyone else had seen it, or if it was just the result of a bad burrito.
 
  • Informative
Reactions: DukeofURL
I wonder how many tests they did?
My guess is they did a bunch of tests with different road markings and mannequins until they found a failure. That’s the beauty of neural nets, you can get any output you want if you try enough different inputs.
FSD beta is optimized for zero intervention drives. Swerving around the road for phantom or real children would not be enough for Whole Mars Catalog to disengage. However if you’re boxed in by cones and it might be a phantom child then there’s no reason to stop and ruin a zero disengagement drive! That’s why you have a safety driver.
They really need to get that alien spaceship recognition NN working. Anyway that’s why it’s still beta!
 
Last edited:
Did anyone else notice that each case had a figure dressed in loose clothing with a large floppy hat or hoodie. They went out of there way to disguise the human form of the mannequin until the car no longer saw it as a VRU. It's no accident that the mannequins were not dressed in proper-fitting clothes.

Look closely at a closeup shot of the mannequin. The face is very poorly defined with no eyes and just a hint of a nose. This was also intentional so as to make it hard to identify the object as a human. Combine that with the fact that the mannequin does not move and you have a clear attempt to trick the car.

It is also intentional to have placed cones for a lane. They made the lane intentionally too narrow for the car to steer around the mannequin.

Of course, FSD beta shouldn't run over immobile kids standing in the middle of a one lane street wearing their dad's clothes and floppy hats with stockings over their heads.
 
As I have worked in the broadcast industry for 58 years, I am reasonably familiar with paid ads vs programming material.
I only caught the last 75% or so of this tripe and could not tell whether it was a paid ad or something they thought was "news". There was a lower third graphic with an 800 number to call and some websote, but I was not in a position to jot either down.

A broadcast station can indeed refuse to air any ad they deem not to air. However, comma, there are a number of regulations under the Fair Standards Act with which licensed broadcasters must comply. a station can refuse to air any political ads, but must then refuse to air all ads regarding a particulat subject or candidate.

Cable TV networks, however, are a different kettle of fish. They have certain regulations with which they must comply, but not all fall into lockstep with the FCC rules and regs.

For all I know they may be "reporting" a story about someone running this crap, in defense of Tesla. Knowing CNN's track record, however, I highly doubt it. I've asked an associate who works for CNN to see if he he can track it down.

I only brought it up to see if anyone else had seen it, or if it was just the result of a bad burrito.
The ad was shown at least twice on the 11PM news by my local CBS affiliate. Seems like libel to me, I'm not a lawyer
 
Did anyone else notice that each case had a figure dressed in loose clothing with a large floppy hat or hoodie. They went out of there way to disguise the human form of the mannequin until the car no longer saw it as a VRU. It's no accident that the mannequins were not dressed in proper-fitting clothes.

Look closely at a closeup shot of the mannequin. The face is very poorly defined with no eyes and just a hint of a nose. This was also intentional so as to make it hard to identify the object as a human. Combine that with the fact that the mannequin does not move and you have a clear attempt to trick the car.

It is also intentional to have placed cones for a lane. They made the lane intentionally too narrow for the car to steer around the mannequin.

Of course, FSD beta shouldn't run over immobile kids standing in the middle of a one lane street wearing their dad's clothes and floppy hats with stockings over their heads.
Fan boy alert
 
Did anyone else notice that each case had a figure dressed in loose clothing with a large floppy hat or hoodie. They went out of there way to disguise the human form of the mannequin until the car no longer saw it as a VRU. It's no accident that the mannequins were not dressed in proper-fitting clothes.

Look closely at a closeup shot of the mannequin. The face is very poorly defined with no eyes and just a hint of a nose. This was also intentional so as to make it hard to identify the object as a human. Combine that with the fact that the mannequin does not move and you have a clear attempt to trick the car.

It is also intentional to have placed cones for a lane. They made the lane intentionally too narrow for the car to steer around the mannequin.

Of course, FSD beta shouldn't run over immobile kids standing in the middle of a one lane street wearing their dad's clothes and floppy hats with stockings over their heads.
Hmmmm
 
Sooooo....this morning here in Colorado our local NBC affiliate put up an commercial about Telsa and Full Self driving. I was like WOW....Tesla cars kill children....what????

So I looked at the TEENY TINY print at the bottom and found out it was from "The Dawn Project"....

I figured I would look into them because "why not?".

Here is a link to the basis of their "study".....

https://dawnproject.com/wp-content/uploads/2022/01/Tesla-FSD-Safety-Analysis.pdf

Take it for what it is worth. To me, it seems like a competitor is trying to make waves....

Not a FSD person, heck we pick up our M3LR this week, but these are the types of commercials I simply tune out. Food for thought.
 
You can’t tell from that screen (as I mentioned above they failed to get sufficient depth of field).

Anyway if you look at their document (at the link above) they have a picture from inside the car of the screen and you can tell at least when that picture was taken that AP was engaged. That picture in the article was at the beginning of a run. I recall from watching the video that on at least one run there appeared to be blue on the pathing. (Detailed videos linked in link above. See screen captures below.). Video link: https://dawnproject.com/wp-content/uploads/2022/08/raw-footage.mp4

I think it’s a terrible test to be clear. But we don’t have to assume AP is off. We definitely expect FSD Beta to hit the mannequin - that is how it works in some situations. FSD Beta will hit VRU sometimes!!!! We KNOW this, people - anyone who uses FSD Beta knows this!
 

Attachments

  • CFDE1123-4AD8-4A3B-BDC3-76C8EE0C13AC.png
    CFDE1123-4AD8-4A3B-BDC3-76C8EE0C13AC.png
    1.1 MB · Views: 57
  • 328845C9-7D7F-440F-8AE8-290C4BCB3075.png
    328845C9-7D7F-440F-8AE8-290C4BCB3075.png
    1 MB · Views: 41
  • 73CBAF18-C1AF-4715-B10C-4A3AF3398971.png
    73CBAF18-C1AF-4715-B10C-4A3AF3398971.png
    1 MB · Views: 42
Last edited:
See above. Be realistic about the limitations of the system if it is placed in a contrived situation designed to make it fail (and perhaps only the failures are shown - the affiants do not declare they did only three tests, in the one affidavit I read). Neural nets are nowhere near as good as humans at the current time. Of course it will also sometimes fail in non-contrived scenarios!
 
Last edited:
But it does raise question, can you initiate Autopilot on an un mapped road without lines like a race track? FSD needs a Navigation track to follow so would that even work if you could not put in a destination.

if all of it does work could I get in a Plaid with FSD and have it take me on a loop of the Nurburgring? cause that would be scary awesome