NVIDIA Unveils "First" AI Computer for Level 5 Driverless Vehicles

[Az_Rael]

Second, while Tesla could just refund FSD purchasers their money instead of upgrading their hardware if they can't deliver, it would be VERY hard to save face while doing so. I don't think the Tesla brand would be willing to take on that amount of negative press. The only way I see it happening is if they give up, blame regulations. But, again, Tesla is a leader, so I don't see them giving up here.

I will bet what will happen is Tesla delivers a few "FSD" features that aren't available under EAP and calls that good.

I mean AP 1 was supposed to pick you up on private property and read stop signs. There was this jewel from the Summon announcement:

During this Beta stage of Summon, we would like customers to become familiar with it on private property. Eventually, your Tesla will be able to drive anywhere across the country to meet you, charging itself along the way. It will sync with your calendar to know exactly when to arrive.

The release of Tesla Version 7.1 software is the next step toward developing fully autonomous driving capabilities and delivering them through over-the-air software updates, keeping our customers at the forefront of driving technology in the years ahead.

Summon Your Tesla from Your Phone

While all the above statements ended up (may still end up) being true, no one at the time knew you would need to buy a new car to get said "over the air" updates. Or that "your" Tesla was actually the new one you would own several years from now, not the car getting the Summon Beta turned on with 7.1.

So I would not read much at all into Tesla's current FSD promises.

 

[Topher]

all I ever said is "not any time soon".

"I think maybe ten years" GO programmer on when a Go program would beat a human master (2014) actually happened in 2017.

The same thing happened in chess, e-sports, jeopardy.

If it were only Elon, one could put it off on his well-known optimism, but every company involved is giving a similar time-frame for their program.

Thank you kindly.

 

[gregincal]

I don't think anyone is arguing that it exists on the Model 3. The point is that once it does exist (Model 3 or other car), there likely will be a way for the car to get an over ride from law enforcement, or construction zone operator, or whatever option you want to plug in here.

That sort of thing doesn't happen overnight. Finalizing and deploying those sorts of government standards usually takes years.

 

[Max*]

That sort of thing doesn't happen overnight. Finalizing and deploying those sorts of government standards usually takes years.

I feel that we all agree on this, I'm not really sure what the point you're trying to make is though...

Considering it's almost 2018 and we're barely at L3 with any mass produced manufacturer. L5 is also a long ways out. And L4 has the option of the driver taking over.

As of today the Model S (or future 3, whatever) is not even L3. But it's L3-ready (Elon claims it's L4/L5 ready, lol), right? An OTA can take it there.

Model 3/the new Model S, as is, most likely will never hit L5. That doesn't mean that a hardware swap/software update can't take it to L5 (I don't think Tesla will do it, but I'm pessimistic); whenever lawmakers decide exactly what L5 is -- actually there is documentation already for autonomous cars, and I didn't see anything related to the question being raised here, but at the same time I wouldn't expect every corner case to be addressed either. And as we all know, the documentation/legislation will update as we get closer to having at least L3 cars on the road -- which is the first stage of real autonomy.

 

[gregincal]

I feel that we all agree on this, I'm not really sure what the point you're trying to make is though...

Considering it's almost 2018 and we're barely at L3 with any mass produced manufacturer. L5 is also a long ways out. And L4 has the option of the driver taking over.

As of today the Model S (or future 3, whatever) is not even L3. But it's L3-ready (Elon claims it's L4/L5 ready, lol), right? An OTA can take it there.

Model 3/the new Model S, as is, most likely will never hit L5. That doesn't mean that a hardware swap/software update can't take it to L5 (I don't think Tesla will do it, but I'm pessimistic); whenever lawmakers decide exactly what L5 is -- actually there is documentation already for autonomous cars, and I didn't see anything related to the question being raised here, but at the same time I wouldn't expect every corner case to be addressed either. And as we all know, the documentation/legislation will update as we get closer to having at least L3 cars on the road -- which is the first stage of real autonomy.

It's sounds like you and I agree on this, I think others are overly optimistic. I was just pointing out that there are some really tricky problems to get from Level 3 to Level 5, some of which may not be possible with just programming.

 

[Max*]

"I think maybe ten years" GO programmer on when a Go program would beat a human master (2014) actually happened in 2017.

The same thing happened in chess, e-sports, jeopardy.

If it were only Elon, one could put it off on his well-known optimism, but every company involved is giving a similar time-frame for their program.

Thank you kindly.

The difference is that those automotive manufacturers aren't selling cars today that have predefined hardware. Many companies plan to get there in 2020-2021.

That's a lot of potential hardware changes between now and then.

 

[Garlan Garner]

Nobody said it was easy but this is literally the stuff they have been working on now for many years... starting with primitive sensor systems on cars that were being built for trials at universities 10 or more years ago and now culminating in billions of dollars being spent to get the technology good enough to deploy on public roads.

object recognition is a big part of self driving but it's augmented by things like Lidar. The car doesn't necessarily have to know that it's a person on the road it just has to recognize that it's an object, start slowing down and then once it knows more about the object based on other sensor information it can decide what it is going to do.

In current cars with AP2 I believe that if a person was lying comatose on the road the car would not run over them, it would slow down recognizing there was an object on the road and then pull off and alert the driver to take control.

Did you see who I quoted and what he said? He said it was easy. So your statement isn't true. I was talking to someone else.

Now....Tesla hasn't been working on it for years. Tesla has been working on it since they got a divorce from their AP1 vendor and introduced AP2. Their AP1 vendor is now dancing with another car manufacturer and is making tremendous progress. It seems to be much easier for Tesla's AP1 vendor as AP2 still isn't up to snuff with AP1 ( as it was supposed to be in January).

Full autonomy - which is what we are talking about is NOT going to happen in November as Tesla / Elon stated long ago. Neither are those who purchased Full Autonomy in their lease will get a refund. Its just money lost. Oh well.


Again....the person I was speaking to said it was easy.

 

[Garlan Garner]

Everyone I know in the aviation industry says, "If they can do autonomous cars, planes should be no problem!" And they all consider cars the harder task because the environment they operate in is so much more variable.

Fulll autonomy in a plane is a ton less problematic. There is nothing to maintain in a plane but altitude and to stay away from any other squelch. That's it.

 

[Max*]

Did you see who I quoted and what he said? He said it was easy. So your statement isn't true. I was talking to someone else.

Now....Tesla hasn't been working on it for years. Tesla has been working on it since they got a divorce from their AP1 vendor and introduced AP2. Their AP1 vendor is now dancing with another car manufacturer and is making tremendous progress. It seems to be much easier for Tesla's AP1 vendor as AP2 still isn't up to snuff with AP1 ( as it was supposed to be in January).

Full autonomy - which is what we are talking about is NOT going to happen in November as Tesla / Elon stated long ago. Neither are those who purchased Full Autonomy in their lease will get a refund. Its just money lost. Oh well.


Again....the person I was speaking to said it was easy.

I know your majesty wasn't talk to me, but you're 100% wrong. He was telling you that it's EASY to solve that ONE particular problem you stated. And that problem is easy to solve.

No one said, or even implied, it's easy to get to FSD.

 

[Max*]

This is not to say that FSD will be easy.

Again....the person I was speaking to said it was easy.

Again.....I think you missed the "not"

 

[stopcrazypp]

I feel that we all agree on this, I'm not really sure what the point you're trying to make is though...

Considering it's almost 2018 and we're barely at L3 with any mass produced manufacturer. L5 is also a long ways out. And L4 has the option of the driver taking over.

As of today the Model S (or future 3, whatever) is not even L3. But it's L3-ready (Elon claims it's L4/L5 ready, lol), right? An OTA can take it there.

Model 3/the new Model S, as is, most likely will never hit L5. That doesn't mean that a hardware swap/software update can't take it to L5 (I don't think Tesla will do it, but I'm pessimistic); whenever lawmakers decide exactly what L5 is -- actually there is documentation already for autonomous cars, and I didn't see anything related to the question being raised here, but at the same time I wouldn't expect every corner case to be addressed either. And as we all know, the documentation/legislation will update as we get closer to having at least L3 cars on the road -- which is the first stage of real autonomy.

The thing about L3 is that it may make sense to skip (a few manufacturers have said they will skip, including Volvo and Ford). The issues of take over and response time is problematic. That's why Audi's L3 is so narrowly limited. It's not easy to give a large cushion in most driving conditions (as things can happen quite quickly).

 

[S3XY]

Did you see who I quoted and what he said? He said it was easy. So your statement isn't true. I was talking to someone else.

Now....Tesla hasn't been working on it for years. Tesla has been working on it since they got a divorce from their AP1 vendor and introduced AP2. Their AP1 vendor is now dancing with another car manufacturer and is making tremendous progress. It seems to be much easier for Tesla's AP1 vendor as AP2 still isn't up to snuff with AP1 ( as it was supposed to be in January).

Full autonomy - which is what we are talking about is NOT going to happen in November as Tesla / Elon stated long ago. Neither are those who purchased Full Autonomy in their lease will get a refund. Its just money lost. Oh well.


Again....the person I was speaking to said it was easy.

The person you quoted (me) said that determining if something was scenery versus an obstruction in it's pre-determined path was easy. You gave a tree as an example and I said it would be the same as if a person were on the sidewalk or in the way. In that same post I also said "This is not to say that FSD will be easy. There are an infinite number of variables". After which you responded "That's definitely NOT easy. I would almost assure you that if you lay a person down sideways in the middle of the street the computer won't know what it is."

I never said FSD would easily determine what was in the street. I'm not sure that it even needs to. As long as it determines that a sizable solid object is in it's path and either stops or maneuvers around it, it has done its job. In fact it does it already (DON'T TRY THIS AT HOME!)

 

[Garlan Garner]

The person you quoted (me) said that determining if something was scenery versus an obstruction in it's pre-determined path was easy. You gave a tree as an example and I said it would be the same as if a person were on the sidewalk or in the way. In that same post I also said "This is not to say that FSD will be easy. There are an infinite number of variables". After which you responded "That's definitely NOT easy. I would almost assure you that if you lay a person down sideways in the middle of the street the computer won't know what it is."

I never said FSD would easily determine what was in the street. I'm not sure that it even needs to. As long as it determines that a sizable solid object is in it's path and either stops or maneuvers around it, it has done its job. In fact it does it already (DON'T TRY THIS AT HOME!)

It has to judge the value of objects.

There is a difference between hitting a human vs. hitting a tree. FSD has to recognize objects on the fly to determine their value.

FSD will never be on trial in an accident. That's an in-insurable entity. The owner of the FSD vehicle will be the insured party. My FSD better choose to hit a tree vs. hitting a person - Just like humans do.

 

[Chewy3]

There is nothing to maintain in a plane but altitude and to stay away from any other squelch. That's it.

Umm... You might want to know at least the basic definition of things if you're going to try and talk about them as if your an expert....Squelch is a radio function.

 

[teddytoons]

So, then something as simple as adding storage to my Model S to play songs locally from the MCU (a feature promised to us prior to launch) should be trivial then eh? I'm still waiting.

So, unless you are going to duct-tape sensors to your car or some elegant aftermarket solution shows up it's just not realistically going to happen.

Understood. However, this is not trivial. My purchase agreement shows $5000 for Enhanced Auto Pilot and $3000 for Full self-driving capability. As far as I'm concerned, Tesla has 8 grand of my money to provide all the Hardware and AI required to fulfill their end of the bargain. Period.

 

[teddytoons]

It's sounds like you and I agree on this, I think others are overly optimistic. I was just pointing out that there are some really tricky problems to get from Level 3 to Level 5, some of which may not be possible with just programming.

Agreed. Apple created their "Bionic Chip" for the iPhone X as a way to insert more efficient AI processing instead of using solely software. It would be no surprise to me if Tesla was forced to do the same. Only time will tell.

 

[jhs_7645]

Understood. However, this is not trivial. My purchase agreement shows $5000 for Enhanced Auto Pilot and $3000 for Full self-driving capability. As far as I'm concerned, Tesla has 8 grand of my money to provide all the Hardware and AI required to fulfill their end of the bargain. Period.

Is it the $$ or is it the promise? If it's the money, then I would recommend *not* giving them the additional $3,000 and waiting and seeing. I understand that Tesla *says* that they will deliver FSD for an additional 3k, but I'm skeptical that it will rise to the level of 'your car will drive across the city to pick you up at work'. If it's the promise, then I'd say pay the $$ and wait and see.

Don't get me wrong.. I'm *paying* the extra 3k, I'm just doing it with the *hope* that they will deliver. I have no plans on suing anybody if they don't deliver.

 

[voip-ninja]

Is it the $$ or is it the promise? If it's the money, then I would recommend *not* giving them the additional $3,000 and waiting and seeing. I understand that Tesla *says* that they will deliver FSD for an additional 3k, but I'm skeptical that it will rise to the level of 'your car will drive across the city to pick you up at work'. If it's the promise, then I'd say pay the $$ and wait and see.

Don't get me wrong.. I'm *paying* the extra 3k, I'm just doing it with the *hope* that they will deliver. I have no plans on suing anybody if they don't deliver.

The only reason to give them the $3,000 today is that you believe they will eventually have to put a more than $3,000 option into your vehicle and you will be getting a bargain.

 

[jhs_7645]

The only reason to give them the $3,000 today is that you believe they will eventually have to put a more than $3,000 option into your vehicle and you will be getting a bargain.

That's not true (for me).

I don't even know what they might have to do to the car in the future, or what they might charge for FSD. I gave Tesla a 5K deposit a year and a half before they even made my Model S (VIN 3305) because I believed in what they were doing and hoped that they would *actually* be able to make the car and send me one.

The same is true (FOR ME) for FSD. The 3k is a bet, or a kickstarter. Other folks might or might not think of it that way, and that's fine. But that's what allows me to sleep at night knowing I may never see anything come from the 3k.

My main point is if you stress over the 3k and demand FSD you might be setting yourself up for disappointment.

 

[Garlan Garner]

Is it the $$ or is it the promise? If it's the money, then I would recommend *not* giving them the additional $3,000 and waiting and seeing. I understand that Tesla *says* that they will deliver FSD for an additional 3k, but I'm skeptical that it will rise to the level of 'your car will drive across the city to pick you up at work'. If it's the promise, then I'd say pay the $$ and wait and see.

Don't get me wrong.. I'm *paying* the extra 3k, I'm just doing it with the *hope* that they will deliver. I have no plans on suing anybody if they don't deliver.

I'm going to wait and won't mind spending $4k if it actually happens.

Its gonna take a lot to get FSD approved by the government and insurance companies.

Can Tesla push it out to the cars without government approval?