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

Elon ducking behind bogus "regulation" for $10K Level 5 bet this year

This site may earn commission on affiliate links.
Also, as I said, level 5 is not actually defined. Some people will interpret level 5 as some sort of general intelligence (for example if a cop uses a megaphone to tell the car what to do). Others will simply see it as statistical human level in terms of safety. Elon can say level 5 if he wants because there's no consensus about what it actually means.

I agree L5 has poor definition. L5 as defined be SAE says "intent". So how is intent defined? As long as it is intended to be L5 than it is L5? That sounds bogus. Can be L5 with intent and not be real L5 for 20+ years.

Here is the actual definition of L5 with notes and an example. I think it is very clear. I find the example to very clear. I am not sure where the confusion is.

p 25-26:

5.6 LEVEL or CATEGORY 5 - FULL DRIVING AUTOMATION

Definition: The sustained and unconditional (i.e., not ODD-specific) performance by an ADS of the entire DDT and DDT fallback without any expectation that a user will respond to a request to intervene.

NOTE 1: “Unconditional/not ODD-specific” means that the ADS can operate the vehicle under all driver-manageable road conditions within its region of the world. This means, for example, that there are no design-based weather, time-of-day, or geographical restrictions on where and when the ADS can operate the vehicle. However, there may be conditions not manageable by a driver in which the ADS would also be unable to complete a given trip (e.g., white-out snow storm, flooded roads, glare ice, etc.) until or unless the adverse conditions clear. At the onset of such unmanageable conditions the ADS would perform the DDT fall back to achieve a minimal risk condition (e.g., by pulling over to the side of the road and waiting for the conditions to change).

NOTE 2: In the event of a DDT performance-relevant system failure (of an ADS or the vehicle), a level 5 ADS automatically performs the DDT fallback and achieves a minimal risk condition.

NOTE 3: The user does not need to supervise a level 5 ADS, nor be receptive to a request to intervene while it is engaged.

EXAMPLE: A vehicle with an ADS that, once programmed with a destination, is capable of operating the vehicle throughout complete trips on public roadways, regardless of the starting and end points or intervening road, traffic, and weather conditions.
 
Here is the actual definition of L5 with notes and an example. I think it is very clear. I find the example to very clear. I am not sure where the confusion is.

p 25-26:

5.6 LEVEL or CATEGORY 5 - FULL DRIVING AUTOMATION

Definition: The sustained and unconditional (i.e., not ODD-specific) performance by an ADS of the entire DDT and DDT fallback without any expectation that a user will respond to a request to intervene.

NOTE 1: “Unconditional/not ODD-specific” means that the ADS can operate the vehicle under all driver-manageable road conditions within its region of the world. This means, for example, that there are no design-based weather, time-of-day, or geographical restrictions on where and when the ADS can operate the vehicle. However, there may be conditions not manageable by a driver in which the ADS would also be unable to complete a given trip (e.g., white-out snow storm, flooded roads, glare ice, etc.) until or unless the adverse conditions clear. At the onset of such unmanageable conditions the ADS would perform the DDT fall back to achieve a minimal risk condition (e.g., by pulling over to the side of the road and waiting for the conditions to change).

NOTE 2: In the event of a DDT performance-relevant system failure (of an ADS or the vehicle), a level 5 ADS automatically performs the DDT fallback and achieves a minimal risk condition.

NOTE 3: The user does not need to supervise a level 5 ADS, nor be receptive to a request to intervene while it is engaged.

EXAMPLE: A vehicle with an ADS that, once programmed with a destination, is capable of operating the vehicle throughout complete trips on public roadways, regardless of the starting and end points or intervening road, traffic, and weather conditions.
What I find most entertaining about this definition is how people interpret it.
This definition basically says that if it can't proceed and the user decides not to intervene then it will pull over and wait.
Far too many people have an unrealistic expectation that and autonomous car will just carry on blindly like too many human drivers do.
Those same folk bleat about a cautious AP system might cause accidents by slowing down or not plowing through an unsafe condition like a human would.
 
  • Like
Reactions: GeoffEV
Quote the SAE part that talks about intent.

p 30: "8.2 Levels are assigned, rather than measured, and reflect the design intent for the driving automation system feature as defined by its manufacturer."

This line is not saying that manufacturers can just assign any level they want. That would be ridiculous. And it would contradict the fact that the SAE clearly defines what each level is.

This line is saying that the level can't be measured but is assigned by the manufacturer. But the manufacturer still needs to follow the definitions of the levels.

For example: we know L5 means the car can complete entire trips with no human supervision, regardless of starting and ending points or traffic, road or weather conditions and can automatically pull over when necessary. That's the definition of L5.

So we cannot do tests to measure a system to be L5. The definition is qualitative, not quantitative. But, the manufacturer will classify it as L5 because they designed a system that meets that definition of L5.
 
  • Like
Reactions: GeoffEV
Here is the actual definition of L5 with notes and an example. I think it is very clear. I find the example to very clear. I am not sure where the confusion is.

p 25-26:

5.6 LEVEL or CATEGORY 5 - FULL DRIVING AUTOMATION

Definition: The sustained and unconditional (i.e., not ODD-specific) performance by an ADS of the entire DDT and DDT fallback without any expectation that a user will respond to a request to intervene.

NOTE 1: “Unconditional/not ODD-specific” means that the ADS can operate the vehicle under all driver-manageable road conditions within its region of the world. This means, for example, that there are no design-based weather, time-of-day, or geographical restrictions on where and when the ADS can operate the vehicle. However, there may be conditions not manageable by a driver in which the ADS would also be unable to complete a given trip (e.g., white-out snow storm, flooded roads, glare ice, etc.) until or unless the adverse conditions clear. At the onset of such unmanageable conditions the ADS would perform the DDT fall back to achieve a minimal risk condition (e.g., by pulling over to the side of the road and waiting for the conditions to change).

NOTE 2: In the event of a DDT performance-relevant system failure (of an ADS or the vehicle), a level 5 ADS automatically performs the DDT fallback and achieves a minimal risk condition.

NOTE 3: The user does not need to supervise a level 5 ADS, nor be receptive to a request to intervene while it is engaged.

EXAMPLE: A vehicle with an ADS that, once programmed with a destination, is capable of operating the vehicle throughout complete trips on public roadways, regardless of the starting and end points or intervening road, traffic, and weather conditions.
Sorry, things being clear is nonsense. My experience with standards is in the area of programming languages. Saying an implementation is compliant with the standard is easy. What that actually means can take years of argument and still nobody will agree.

I can only imagine that something as fuzzy as driving will be subject to far more disagreement and argument. Nobody can even agree on what a good driver is when that driver is human.

So, no, I'm a gambler and I would never take a bet that involved any sort of agreement on what level 5 means in practical terms. Elon's no fool.
 
... This line is not saying that manufacturers can just assign any level they want. That would be ridiculous.
Yeah, Elon has been rediculous since 2015 saying driverless is coming in two years every year since then, and now is saying a year away.
And it would contradict the fact that the SAE clearly defines what each level is.
You are contradicting what the standard says. If it says intent, it means intent. Doesn't matter what the earlier L5 definition spells out, as long as the design is "intended" to meet it, doesn't matter if it is in the next century, then it is still level 5.
... reflect the design intent for the driving automation system feature as defined by its manufacturer."
Defined by the manufacturer, so Elon can define what ever he wants to be L5.

There needs to be a new definition, one without "intent". Should we call it L5 real and the existing definition L5 fake / intent?? :)
 
Last edited:
You are contradicting what the standard says. If it says intent, it means intent. Doesn't matter what the earlier L5 definition spells out, as long as the design is "intended" to meet it, doesn't matter if it is in the next century, then it is still level 5.

There needs to be a new definition, one without "intent". Should we call it L5 real and the existing definition L5 fake / intent?? :)

Yes, it does matter that what the earlier definitions spell out. Why did the SAE put the definitions in the document then? You can't just ignore an entire section of the SAE document, especially the section that actually defines what the level is.
 
  • Like
Reactions: GeoffEV
In court when reviewing documents such as laws written, any later passage over rules any earlier passage. So intent passage carries greater weight than any earlier passage. If there is a conflict, then the later passage rules. The L5 definition can be viewed as guidance to manufactures, but manufactures ultimately decide the level because the spec says that.
 
NOTE 1: “Unconditional/not ODD-specific” means that the ADS can operate the vehicle under all driver-manageable road conditions within its region of the world. This means, for example, that there are no design-based weather, time-of-day, or geographical restrictions on where and when the ADS can operate the vehicle.

I'd settle for vanilla cruise control working in all weather conditions. Spoiler: it doesn't. Salty water disables the radar, makes the sonar false positive, and blocks all the camera's.
 
In court when reviewing documents such as laws written, any later passage over rules any earlier passage. So intent passage carries greater weight than any earlier passage. If there is a conflict, then the later passage rules. The L5 definition can be viewed as guidance to manufactures, but manufactures ultimately decide the level because the spec says that.

If the manufacturer clearly mislabels what the level is they would be sued and lose.
 
Defined by the manufacturer, so Elon can define what ever he wants to be L5.

I wouldn't interpret it like that. He shouldn't be able to define whatever he wants as L5, but rather design it to be L5.

Hence why I think liability will be a determining factor. I hope AV makers, namely Tesla and its CEO given his propensity towards bending his followers' perceptions of reality in his favor, are taken to task for false claims in this matter.

The design intent line leads me to understand that the system may fail without losing the L5 label. This makes sense to me as we're not clairvoyant, and otherwise (without mentioning intent) L5-intended systems would still be labelled L5 until downgraded due to failure. What I hope it also means, is that if an AV maker designs a system intended for L5 automation, then it must accept that this intention implies that the vehicle is meant to be driving itself, and that the maker accepts liability for faults. No ifs, no ****ery.

Continuing, the design intent line should also help clarify practical issues with level distinction. A particular L5 system that fails in a given condition, is still L5, and the maker should now work on the matter, as the system is meant to function anywhere, in any condition (within reason). A system that is designed and stated to be meant for a particular set of conditions or area is L4.

Edit: "Design intent", if my interpretation is correct, should be useful in preventing Elon from potentially claiming that his L2 system is "L5-capable" (drink if he tries that), as L5 depends on it being intended for unsupervised general use, with liability to match, and so if capable, nothing would stop the maker other than unwillingness to adhere to their claim of L5 (and a very telling unwillingness to be responsible). I quite like it.
 
Last edited:
I wouldn't interpret it like that. He shouldn't be able to define whatever he wants as L5, but rather design it to be L5.

Hence why I think liability will be a determining factor. I hope AV makers, namely Tesla and its CEO given his propensity towards bending his followers' perceptions of reality in his favor, are taken to task for false claims in this matter.

The design intent line leads me to understand that the system may fail without losing the L5 label. This makes sense to me as we're not clairvoyant, and otherwise (without mentioning intent) L5-intended systems would still be labelled L5 until downgraded due to failure. What I hope it also means, is that if an AV maker designs a system intended for L5 automation, then it must accept that this intention implies that the vehicle is meant to be driving itself, and that the maker accepts liability for faults. No ifs, no ****ery.

Continuing, the design intent line should also help clarify practical issues with level distinction. A particular L5 system that fails in a given condition, is still L5, and the maker should now work on the matter, as the system is meant to function anywhere, in any condition (within reason). A system that is designed and stated to be meant for a particular set of conditions or area is L4.

You get it.
 
Safer than human has been implied since beginning, it's not a moving goalpost (if anything, people are becoming worse drivers with their smartphone distractions).

They probably want to be backed by statistics (more data) before they unleash the FSD where may be.

In the past Elon has repeatedly stated that we will be able to sleep in the backseat while our cars take us to our destination.

Now it's just going to be "safer than human drivers". Goal posts = MOVED
p 30: "8.2 Levels are assigned, rather than measured, and reflect the design intent for the driving automation system feature as defined by its manufacturer."

This line is not saying that manufacturers can just assign any level they want. That would be ridiculous. And it would contradict the fact that the SAE clearly defines what each level is.

This line is saying that the level can't be measured but is assigned by the manufacturer. But the manufacturer still needs to follow the definitions of the levels.

For example: we know L5 means the car can complete entire trips with no human supervision, regardless of starting and ending points or traffic, road or weather conditions and can automatically pull over when necessary. That's the definition of L5.

So we cannot do tests to measure a system to be L5. The definition is qualitative, not quantitative. But, the manufacturer will classify it as L5 because they designed a system that meets that definition of L5.

When, exactly, did you become the authority on automated driving? Watching you write walls of text, drawing conclusions that are only your own, is really annoying.

Sprinkle in "orders of magnitude".:rolleyes:
 
When, exactly, did you become the authority on automated driving?

I used to be very ignorant about automated driving. A few posters on this forum actually called me out on it. So I decided to educated myself on the topic. I read a lot and tried to learn as much as I can. I am still learning.

I would not say that I am an authority on automated driving but I enjoy sharing what I know.

Watching you write walls of text, drawing conclusions that are only your own, is really annoying.

I am sorry that sharing my thoughts on automated driving is annoying to you. I try to share facts and information. I do share my opinions too of course which I have every right to do.
 
So it's ok to lie to customers about your product as long as your company is successful?

How much money has Elon snookered out of people with his claims about what FSD will be able to do and when those things will happen?
My parents taught me to be suspicious of advertising. Not to say that was a long time ago, but they got it from the romans: “caveat emptor”. They also taught me to RTFC “read the contract”. Anyone who feels hornswaggled regarding FSD because of the optimism in Tesla’s advertisements (aka Musk’s tweets, because that IS their advertising) should apply the lesson to their future pursuits so they do not fall victim to the legions of companies out there promoting their product.

I’m not saying it is ok to lie. Just that it is standard in advertising.
 
My parents taught me to be suspicious of advertising. Not to say that was a long time ago, but they got it from the romans: “caveat emptor”. They also taught me to RTFC “read the contract”. Anyone who feels hornswaggled regarding FSD because of the optimism in Tesla’s advertisements (aka Musk’s tweets, because that IS their advertising) should apply the lesson to their future pursuits so they do not fall victim to the legions of companies out there promoting their product.

I’m not saying it is ok to lie. Just that it is standard in advertising.


Oh I've made it my mission to talk as many people out of buying FSD as I can. They've definitely lost more money because of me than they got out of me at this point.