Tesla "Full Self Driving" not capable of autonomous driving

[Dadhawk]

Originally Posted by soundman98

ever notice everyone's either a 'tesla hater' or a 'tesla lover'? why is there so much passion for a car company?

Actually, I'm neither, as stated above.

I'm also a sports car enthusiast that has the exact opposite of a sports car on order (Cadillac LYRIQ). If Tesla built something similar from a design prospective, I would consider it.

My LYRIQ build...

[NoHaveMSG]

Originally Posted by Irace86.2.0

Yeah, that is one reason why I could see the adoption of autonomous driving happening sooner rather than later. In the meantime, those sulfur or any solid state battery will likely help, but these things could be the solution. Push the EV in the container and hit it with water that recirculates so you aren't using a lot of water.

https://brandogsikring.dk/en/news/20...electric-cars/

I have heard of those. I just don't see a lot of departments being able to budget for them or have access to them for a while. Maybe someone will come up with something more compact and cheaper in the future that is on site deployable. Now I am kinda curious what temp the water would get up to.

[NoHaveMSG]

Originally Posted by soundman98

ever notice everyone's either a 'tesla hater' or a 'tesla lover'?

why is there so much passion for a car company?

I don't care either way. I have driven a few, I don't think they are bad cars and they drive fine. The UI I find a bit annoying in some everyday situations. The company and the guy behind it, I don't really care for. I am sure if I found out more info on the CEO's of other car companies and their dealings, I wouldn't care for a few of them either

[Capt Spaulding]

Originally Posted by Irace86.2.0

At least I am using statistics, or even trying to make an argument. It seems like most of what everyone else is engaged in is baseless conjecture, fear mongering or making false statements. I get it. Most sports car owners don't like EVs and are looking for fault with them in any way possible, but the special pleading and baseless claims get old. I'm just trying to apply some reason to irrational arguments. If anyone can make a rational argument or provide some evidence then I am good to agree with it.

I've ridiculed Musk enough throughout this forum on more than one occasion on a number of different issues to escape the fanboy title, but if you want to throw that around, I could care less. I've stated that my go-to personal EV would be the Hyundai Ioniq 5 because I like the looks the most. The Honda Urban EV and Sports EV Concepts were also two EVs I would take over a Tesla any day of the week, purely for exterior design because that is almost chief in any car I choose. My wife would be good with the Audi E-Tron Q4 or Tesla Model Y/X, as I've also stated numerous times.

Using statistics and misusing statistics are not the same thing and I know you know that. From what I can tell - the "10 times safer" claim is a classic example. They compare ducks with oranges and conclude that their car is the Flying Spaghetti Monster's gift to automobile safety. Doing stuff like that is IMO just plain wrong.

I have nothing against Tesla as an entity. They had a number of QC issues early on and a majority of the few people I know who own them have had service issues that turned painful. I wouldn't buy one - probably ever. Why? Elon Musk. I don't hate the car, but as a corporate figure head, he is more toxic than the dust from a Kryptonite/Plutonium alloy. I suspect that NASA pinning their Artemis landing on SpaceX will end in tears. But, they appear to have a group of talented engineers. Perhaps if Musk will keep completely out of the way, they'll pull it off.

There's not much point in my shopping eVs at this point. By the time I would be in the market, that market will look nothing like it does now. I look, but that's it. And just so you know it's me, I'll go pedantic for you. The proper term for something NOT mattering is "I could NOT care less." LOL

[Irace86.2.0]

Originally Posted by NoHaveMSG

I have heard of those. I just don't see a lot of departments being able to budget for them or have access to them for a while. Maybe someone will come up with something more compact and cheaper in the future that is on site deployable. Now I am kinda curious what temp the water would get up to.

They shouldn’t be too expensive because they are basically a shipping container retrofitted with water suppression. Maybe in rural areas, but a city like the one I live in could get away with one or several without issue and be a five to ten minute drive to the scene.

[Irace86.2.0]

Originally Posted by Capt Spaulding

Using statistics and misusing statistics are not the same thing and I know you know that. From what I can tell - the "10 times safer" claim is a classic example. They compare ducks with oranges and conclude that their car is the Flying Spaghetti Monster's gift to automobile safety. Doing stuff like that is IMO just plain wrong.

You say this about statistics, yet you aren’t saying anything about people stating that the system is more dangerous than not using the system. How are people coming up with the belief that the system is more dangerous? A few stories about Teslas crashing or a person dying while abusing the system and people seem to quickly draw the conclusion that the system is more dangerous than someone driving without these aids. They have zero statistics. Zero.

Some have argued that Autopilot is engaged in highway driving, so numbers will be higher. Some argue that premium car owners are safer drivers. Maybe, but do you think it is likely that these statistics are the opposite, and even if they could be, the conclusion would be conjecture at best. We just don’t have the statistics besides the hard numbers. As it is, Autopilot appears to be safer when engaged than when not. The only way it could be less safe is if the safety of the average vehicle is soooooo much better on a highway, but considering the average safety of a vehicle and the average miles traveled on highways versus roads, that seems like an impossibility.

Same with the fire statistics. Those seem like hard numbers to manipulate in such a way that the situation is the reverse. Someone would have to believe that Tesla is just lying, but this data is probably tracked by government data. Where are the fact checkers if that is the case? Again, if “we just don’t know” is your conclusion then why is it that people can continue to make claims that Teslas are a greater fire hazard, as if they are spontaneously catching on fire often? The only thing we can conclusively say is that battery fires are harder to put out, but this is not what people are claiming when they say Teslas and EVs are a significant fire risk compared to ICEs. This is a baseless claim, and at least this data set suggests the opposite. What sources are other people using?

https://www.tesla.com/VehicleSafetyReport

[Capt Spaulding]

Some have argued that Autopilot is engaged in highway driving, so numbers will be higher. Some argue that premium car owners are safer drivers. Maybe, but do you think it is likely that these statistics are the opposite, and even if they could be, the conclusion would be conjecture at best. We just don’t have the statistics besides the hard numbers. As it is, Autopilot appears to be safer when engaged than when not. The only way it could be less safe is if the safety of the average vehicle is soooooo much better on a highway, but considering the average safety of a vehicle and the average miles traveled on highways versus roads, that seems like an impossibility.

You make the point yourself and then you subtly distort mine. My point is that use case matters. If you base your conclusion on a comparison of AP's performance on the highway with the performance of human drivers in ALL driving cases, that is a cheat. I'm not saying people are better on the highway than AP. They are probably not. AP appears, for the most part, to fall asleep less often. But that is not my point. My problem with their numbers is that to claim AP is 10 times safer than human drivers based on the comparison they did is a lie. a LIE. I really don't like cheating. Nor lying. Nor those who practice either. Particularly when it come to safety.

I suggest that this discussion illustrates the point well. Based on observed behavior, some people, perhaps even you, seem to have bought Tesla's specious safety claims. That appears to have encouraged some of those humans to do REALLY stupid things behind the wheel. I never said human drivers are better than AP/FSD or that AP/FSD is worse than humans. I said, and say, that the data don't support the claims some make for it. And that, making those claim as fact has really pernicious consequences.

Edit: as to the vehicle fire issue - that is a different conflagration. All I can say for sure is that my son, who commands ARFF operations at a midsized airport, seems leery of LI battery fires because they are difficult to put out.

[NoHaveMSG]

Originally Posted by Irace86.2.0

They shouldn’t be too expensive because they are basically a shipping container retrofitted with water suppression. Maybe in rural areas, but a city like the one I live in could get away with one or several without issue and be a five to ten minute drive to the scene.

Don't forget the truck to haul and deploy it. Our population is nothing like CA's, the majority of it is fairly rural.

[Irace86.2.0]

Originally Posted by Capt Spaulding

You make the point yourself and then you subtly distort mine. My point is that use case matters. If you base your conclusion on a comparison of AP's performance on the highway with the performance of human drivers in ALL driving cases, that is a cheat. I'm not saying people are better on the highway than AP. They are probably not. AP appears, for the most part, to fall asleep less often. But that is not my point. My problem with their numbers is that to claim AP is 10 times safer than human drivers based on the comparison they did is a lie. a LIE. I really don't like cheating. Nor lying. Nor those who practice either. Particularly when it come to safety.

I suggest that this discussion illustrates the point well. Based on observed behavior, some people, perhaps even you, seem to have bought Tesla's specious safety claims. That appears to have encouraged some of those humans to do REALLY stupid things behind the wheel. I never said human drivers are better than AP/FSD or that AP/FSD is worse than humans. I said, and say, that the data don't support the claims some make for it. And that, making those claim as fact has really pernicious consequences.

Edit: as to the vehicle fire issue - that is a different conflagration. All I can say for sure is that my son, who commands ARFF operations at a midsized airport, seems leery of LI battery fires because they are difficult to put out.

I don't think you mean to say it is a lie. The data Tesla is providing are statistics. They can be interpreted in different ways. As far as I can tell, Tesla didn't state the car is 10x safer with Autopilot; that was what some articles said and that was what people like me quoting those articles said when I quoted those articles. I've admitted that the numbers aren't likely an apples to apples comparison, but I've also said that they are not suggestive of the opposite, which is what people are suggesting here, that Autopilot/FSD is more dangerous than the average driver. What people here are claiming is baseless because they are providing zero data when making these claims, and the only data we have from Tesla suggests Autopilot is safer. You called me a "cool aide fueled Musk fanboi", criticizing my analysis, while not criticizing everyone else for making baseless claims that Autopilot is more dangerous than drivers. That seems biased. Then you went on a Musk name-calling rant, which also suggests you might be biased.

But the numbers are misleading. Autopilot is used mainly for highway driving, which is generally twice as safe as driving on city streets, according to the Department of Transportation. Fewer crashes may occur with Autopilot merely because it is typically used in safer situations.

From: https://www.nytimes.com/2022/06/08/t...fety-data.html

Looking at Tesla's fine print, they state that the data they provided is conservative:

To ensure our statistics are conservative, we count any crash in which Autopilot was deactivated within 5 seconds before impact, and we count all crashes in which the incident alert indicated an airbag or other active restraint deployed. (Our crash statistics are not based on sample data sets or estimates.) In practice, this correlates to nearly any crash at about 12 mph (20 kph) or above, depending on the crash forces generated. We do not differentiate based on the type of crash or fault (For example, more than 35% of all Autopilot crashes occur when the Tesla vehicle is rear-ended by another vehicle). In this way, we are confident that the statistics we share unquestionably show the benefits of Autopilot.

From: https://www.tesla.com/VehicleSafetyReport

This is in contrast to the national data, which is likely an underreporting:

Two factors complicate the national crash data. First, states have different requirements concerning what incidents are reported as crashes. Second, many crashes go unreported. Estimates of unreported rates of crashes have ranged from as little as 15.4 percent to as much as 59.7 percent (Blincoe et al., 2015; M. Davis & Co, 2015). The result is that the current national crash rate is essentially a low estimate of the actual crash rate.

From: https://featured.vtti.vt.edu/2016/01...-city-streets/

Unless Tesla outright lied about the data, the data suggests that a Tesla with its normal driver safety aids is safer than the national average (some of which also include these similar aids from other manufactures, so that only betters Tesla's position relative to cars without driver aids). Even if highway driving is twice as safe as city driving like the NYT article stated, there is no way to tease out highway miles from the national average and not have Autopilot still being magnitudes safer. It very well could be 10x safer. We don't know.

If anyone is suggesting Autopilot is more dangerous or that fires are happening more often than the national averages then I would love to see their data sets, so we can have a real debate. Yes, BEV fires consume more water and are more difficult to put out, but the incidence of fires is blown up in the media.

Lastly, you say:

I suggest that this discussion illustrates the point well. Based on observed behavior, some people, perhaps even you, seem to have bought Tesla's specious safety claims. That appears to have encouraged some of those humans to do REALLY stupid things behind the wheel.

To review, the claim from Tesla is that using Autopilot (a level 2 autonomy system that requires driver engagement) is safer than not using Autopilot. Yes, people probably believe this. I too think it is highly likely that it is safer, so I'm assuming others do too. Some people might think the opposite, that it is more dangerous, but people know it is more dangerous to speed, drive too close and change lanes aggressively, and they still do those things, so I don't know what conclusions we can draw based solely on human behavior or use and purchase of the Autopilot and FSD systems, respectively. Perhaps everyone thinks it is more dangerous; isn't that what the media seems to be suggesting all the time? The safety of Autopilot and FSD may have absolutely zero influence on people doing "REALLY stupid things behind the wheel", and it could have everything to do with the capabilities; it isn't that people feel safe to use the system contrary to how they should, but rather, they can use the system that way. You are assuming the people who jump in the back seat of a Tesla during Autopilot/FSD and who take a quick TikTok video, or whatever stupid thing they are doing, are doing it because they believe they can do it safely. This is a huge assumption on your part. People do all sorts of stupid things in cars and on motorcycles that have the capabilities, but that increase the risk of morbidity and mortality. In fact, it is far more likely people post these stupid stunts because they know they are risky and inherently not safe like stunting a motorcycle on the highway or driving a Hellcat at 190mph. If it was safe, and they believed it to be safe, then what is the big deal?

[soundman98]

it's official. my attention span is shorter than any of these posts.

[spike021]

Originally Posted by soundman98

it's official. my attention span is shorter than any of these posts.

mine ended when i posted the gif

[Spuds]

Originally Posted by soundman98

ever notice everyone's either a 'tesla hater' or a 'tesla lover'?

why is there so much passion for a car company?

I don't hate Tesla. I will argue against technology, practices, and marketing that I believe are wrong but that's not what I hate.

I don't hate Elon Musk either. He's just a quite flawed human like many others. He has a modicum of talent, decent business sense and a lot of "personality" whatever that means. Oh and a hyperinflated ego and complete lack of ethical fortitude.

I do hate deception, hypocrisy, and manipulation. There's a difference between sharing a vision of the future, however much I disagree with it, and outright lying to people and altering technical facts for the purpose of getting money into your bank account.

[Spuds]

Originally Posted by Irace86.2.0

--He pretty much said that exact thing, and when they removed radar, they announced that features would be removed temporarily, as they updated the algorithm/AI to perform radar-only-based functions. “Haha, FSD 9 beta is shipping soon, I swear! Generalized self-driving is a hard problem, as it requires solving a large part of real-world AI. I didn’t expect it to be so hard, but the difficulty is obvious in retrospect. Nothing has more degrees of freedom than reality.”--Elon Musk
Elon Musk admits self-driving is harder than he thought as Tesla owners troll him over missed deadlines
https://electrek.co/2021/07/05/elon-...sed-deadlines/
Tesla Vision Update: Replacing Ultrasonic Sensors with Tesla Vision
https://www.tesla.com/en_eu/support/...g-tesla-vision

--There could be some bias built into the statistics, but autopilot and FSD is available at night and in the rain. Who knows if people activate it more often or less often in those situations. Most accidents happen at intersections, changing lanes aggressively in blind spots or from driving too close; all of those situations are not likely to happen with Autopilot and FSD based on the grandma driving style of the systems.
Tesla’s radar-less Autopilot performance improves in heavy rain
https://www.teslarati.com/tesla-visi...model-y-video/

-Well, it seems like Tesla wants people with >10k miles on FSD beta to be able to opt out of the steering wheel nannies because they have proved themselves responsible, and the camera will still be there (probably covers them more than alerts the driver). I don't know if that works for any driver of the car like maybe I drove it 5k miles and need the nannies, but my wife drove 5k miles in the car with FSD and is more responsible. Overall, Musk has always been more about personal freedoms and personal responsibilities. He wouldn't expect a Hellcat to have an accelerator limiter on the gas pedal and different than a car needs to monitor your driving. If I can turn off traction control and stability control then maybe people should be able to turn off the nannies.
Tesla plans to remove a Full Self-Driving Beta driver monitoring feature, regulators are concerned
https://electrek.co/2023/01/09/tesla...ors-concerned/

-The difficulty was obvious from the get-go. His early "demo" wasn't a demonstration of the actual algorithm driving the car, he just had a bunch of stuff hard coded in. As an engineer with some experience in autonomy, I was certain Musk/Tesla was full of shit back in 2017 when he said it would only take a year.
-I would need to look closer at the data to discuss this further and I just don't have the time or energy. I still have a high amount of suspicion though.
-He's all about personal freedoms until the "free speech absolutionist" finds somebody saying something he doesn't like. Also a Hellcat does have a limiter, it's called traction control. You can defeat it if you want to wind up wrapped around a tree but that's a bit like falling asleep at the FSD wheel. The difference is you don't have the CEO saying this car will drive itself.

[Capt Spaulding]

Without resorting to a wall of text response -

Edit: it appears I was mistaken.

Things I did NOT say.

An average human driver is demonstrably safer in the highway use case than AP/FSD/LSMFT/AMF.

That numbers themselves lie.

That you ARE a “cool aide fueled Musk fanboi.”

That data ARE statistics.


What I DID say or imply.

Tesla is dishonest. They have created the misleading impression/narrative that their product can do things it cannot. In doing so they have created a false sense of security among drivers which arguably leads some of those drivers to engage in extremely unreasonable and dangerous behavior.

Further, to create this narrative Tesla cherry picked their data in such a way as to misrepresent the level of safety afforded by their “driver assist” features. To use the euphemism “driver assist” in a post hoc attempt to distract from their use of the term “Full Self Driving” is further evidence of their duplicity/dishonesty and evinces a disregard for either the intelligence or the safety of their customers - or both.

Statistics are NOT data. I know that. You know that. Statistics are tools to analyze data and make inferences about real world phenomena. The operationalization of one’s conceptual constructs and the measurement processes used in the data collection process have a direct bearing on the quality of the data one collects. These are up to the investigator and can be used honestly or dishonestly. Likewise, the choice of what, if any, data manipulations to make and what statistics methods/tools to employ in analyzing the data are up to the analyst. The responsibility for those choices lie squarely with the investigator.

The quotes from Tesla’s study you furnish persuade me that their research design was developed to support specific claims, which they make, that an honest analysis of the data do not support. They choose to make those claims anyway. In doing so they are lying.

You see, data don’t lie. Statistics don’t lie. Data analysts, armed with the tools of statistics lie all the time. They should be called out when they do.

Now you are making this a much more convoluted discussion than is necessary.

To simplify things let’s take several questions.

Did Tesla’s analysis of the relative safety of their so called FSD draw a comparison between accident data from vehicles so equipped in a narrowly defined use case with those of vehicles not so equipped in a much broader use case?

Would a reasonable and prudent professional data scientist realize that?

Could conducting such a comparison lead to biased and misleading results?

Would/should a reasonable and prudent professional data scientist realize that?

Did Tesla nevertheless make claims suggesting that their product was many times safer than an average human driver?

It seems to me the answer to each question is yes.

Therefore, Tesla lied.

Now lots of makers of lots of products lie. I guess what I find interesting is the insistence on defending them.

And you have me pegged on one thing. I loathe Elon Musk. As I said before, i consider him a juvenile, narcissistic, egotistical, petty, overbearing, selfish, bullying, megalomaniac. I altered my list of adjectives.

With regard to yourself, I think I said that you were coming across as a cool aide fueled Musk fanboi. Not that you were/are one.

In the final analysis, this isn’t about me or about you or about intellectual **** measuring. It’s about honesty and dishonesty. In my considered opinion, Tesla is and has been dishonest.

That is it.