Why AP 2.0 Won't Be Here Soon, and It Won't Be What You Think It Is

[HankLloydRight]

I am saying that the answer to 'how is a car going to do this?' is probably 'Same way that you do.'

Here's your actual claim.

Are you going to provide any proof to back this up, that humans will most likely react in exactly the same way as a computer would given the same inputs?

 

[Stoneymonster]

I predict this thread, or one like it, will still be going on while there are actual, fully autonomous non-research vehicles driving on US roadways.

 

[RogerHScott]

I predict this thread, or one like it, will still be going on while there are actual, fully autonomous non-research vehicles driving on US roadways.

So you're making a prediction about 5+ years out in the future?

 

[Stoneymonster]

So you're making a prediction about 5+ years out in the future?

Quite possibly! But I promise I don't work in marketing

 

[MarkS22]

I don't see how you can support the theory that humans, 100% of the time, without a single fail or falter, in the entirety of mankind, will always make the same decision given the "exact same inputs". If that doens't happen just once, your theory fails.

This is becoming philosophical. If exact same inputs means literally exact, and that results in the same result every time, there's no free will and everything that ever happened and ever will happen is predetermined.

The question is if there are truly any "random" reactions that would shift the impulses in the brain enough to react differently based on identical stimuli.

So, about those radar sensors on AP 2.0...

 

[Ohji]

I was really excited about the Drive PX 2 platform when I read about it -- with Elon Musk giving talks at Nvidia events and the use of the Tegra in the Model S/X, I was hoping an enhanced Tesla/Nvidia partnership was in the works with a Nvidia-based solution coming to AP 2.0. But one article I read discussing Drive PX 2 stated the cost of the board is $15,000! $15k before the cameras, sensors, etc is a LOT of money. With the increasingly public and unpleasant breakup of Tesla and Mobileye, I am wondering where they will go next. The enhanced radar is a nice interim step, and there have been rumors here on TMC that maybe a Tesla/Bosch partnership for AP is being developed. The job postings for chip engineers/designers suggests an in-house chip may be in the works, but I would imagine that would be years off.

It will be interesting to see when AP 1.5/2.0 really does hit the streets...

 

[HankLloydRight]

Not only isn't that a proof, it isn't even an example, nor would your caveats even apply to your original claim. I won't bother explaining card counting to someone who thinks holding on an 11 is something people might do.

Ok, since my gambling examples seem to confuse you, how about this. I take random people off the street, and flip a coin ten times and ask them to "call it" but not show them the result of each flip. Unless every person tested always says "heads" or always says "tails", I think that's pretty good proof that given the exact same input ("the flip of a coin") that people will do different things given the exact same input.

If you are going to again claim that this isn't a proof, or even an example, don't just say it isn't a proof, please explain why, I'm willing to listen and learn. Just saying it isn't a proof doesn't mean it isn't a proof. It just means you reject it.

 

[RogerHScott]

Can we send this off on a nice little vacation on its own thread? Though I can't begin to imagine what the title would be...

 

[msnow]

Ok, since my gambling examples seem to confuse you, how about this. I take random people off the street, and flip a coin ten times and ask them to "call it" but not show them the result of each flip. Unless every person tested always says "heads" or always says "tails", I think that's pretty good proof that given the exact same input ("the flip of a coin") that people will do different things given the exact same input.

If you are going to again claim that this isn't a proof, or even an example, don't just say it isn't a proof, please explain why, I'm willing to listen and learn. Just saying it isn't a proof doesn't mean it isn't a proof. It just means you reject it.

I can't believe you've indulged this for as long as you have. It seems to me that for him it's about not losing the logic argument even though you have proven it several times.
Thank you kindly.

 

[kevinf311]

Can we send this off on a nice little vacation on its own thread? Though I can't begin to imagine what the title would be...

"snippiness"

 

[Todd Burch]

I agree and disagree with the OP. Agree that L3 or L4 autonomy is quite a few years away. There are so many scenarios that are tricky to navigate that it would take some seriously good sensors/cameras and software to be successful. However, I disagree that the AP 2.0 hardware is that far off. I think you'll see the hardware sometime next year, but the software will be almost identical to the current AP at first and then slowly add features to improve.

So then you agree and agree. That's pretty much what I said .

 

[edb49]

Hi everyone, long term lurker here - have spent many, many hours reading this forum. Sorry to drag this back on thread too!

I put down a reservation on an X in 2014, and have ordered a P100D. Tesla haven't started producing RHD Xs yet, as far as we know, and even if they did start making them tomorrow then it would probably be towards the end of November I would get my car due to shipping, etc.

When I say "Autopilot 2.0", I am talking about the next iteration of hardware. Although far from certain, it seems likely that will be dual/triple front cameras and more radars around the car.

I am really keen to make sure I don't miss out on the next iteration of hardware. The software doesn't really matter to me; I will get that when everyone else does, but potentially I could take delivery of my car on a Monday, and the guy who takes delivery on the Tuesday has paid the same as me but has got the next iteration of hardware, which will inevitably be "better".

I think Troy's post was a good summary of the indicators towards hardware changes:

◘ In THIS video (after 39:35s) Elon basically confirmed that the Model 3 will have autopilot 2.0 hardware.
◘ Elon also said, "The Model S and the X are always going to be our technology leader." (source)
◘ And we know that the deadline for Model 3 suppliers is 1st July 2017. (source)

Therefore it is obvious that Model S and X will have autopilot 2.0 hardware before 1 July 2017. My best guess is January 2017.

There are a few other data points I'd like to add though, that seems to have been missed so far.

1. USA Today reported Mobileye's CTO saying that it is best to end the relationship with Tesla by the end of this year. Source:
Tesla and key Autopilot supplier split up

Shashua said moving cars to higher levels of self-driving capability "is a paradigm shift both in terms of function complexity and the need to ensure an extremely high level of safety." He added there is "much at stake" in terms of Mobileye's reputation, and that it is best to end the relationship with Tesla by the end of the year. He didn't elaborate.

2. The design studio in the UK for a X says "early 2017" delivery. In August, the last time I checked, it said "Feb 2017".

3. Tesla DS in the UK have said that Tesla were making the engineering changes necessary for RHD at the same time as some other general engineering changes.

4. The UK is one of the most conservative markets Tesla are in, in terms of pre-ordering. (A Tesla employee told me this.) To corroborate this, I was reservation #103 in the UK and I made it in Dec 2014 - compare this to the USA (albeit population 5 times higher).

5. Various sites have reported the reveal #2 of the Model 3 will be before the end of this year. As Troy pointed out above, it's likely this will have next iteration of hardware, and Tesla don't want the Model 3 to be superior in this respect to the S/X, so anything they bring to the 3 they'll bring to the S/X first.

So the analysis of this all:
The timing of (2) would indicate cars being produced end of Dec/early Jan. I don't think there is much of a queue for Xs in the UK due to (4). From (3), it would seem to be they're making changes that are more significant than just switching to RHD. From (1), Mobileye have have hinted that the supply (but not support) agreement with Tesla will come to an end this year.

My interest in this all is to make sure I don't miss out on new hardware by a matter of days/weeks. There is a certain safety net I have with the transit time - e.g. using as an example:
1) 1st Oct - Tesla build my car
2) 8th Oct - Tesla switch to next iteration of hardware
3) 9th Oct - Tesla build car for someone in California on TMC user johndoe1234
4) End Oct - johndoe1234 reports on forum new hardware

So before my car has reached the UK, I know it is obsolete. Obviously only applies for a certain window, I am not sure what I'd do in this scenario!

 

[HankLloydRight]

Buying a Tesla means that no matter what you do, it will be obsoleted in some way in about a six month horizon. It's just the way it is. Trying to time the AP 2.0 hardware upgrade to the week is a fool's errand (IMHO, no offense intended). The only sure fire way to be sure to get AP 2.0 hardware is to wait for it to be released. And nobody knows when that will be. In my opinion, it surely won't be anytime this year, and not anytime soon (i.e. "early" 2017). Elon said they're working hard to redesign/maximize the software given the current AP 1.0 hardware suite. I don't think they'd be doing that AND working on new sensors, which would require an entirely additional level of sophisticated software engineering.

And at some point, I'd think we'd start to see a few spy photos pop up of mules with a new hardware suite, just like we did with AP1.0.

 

[Chopr147]

I could take delivery of my car on a Monday, and the guy who takes delivery on the Tuesday has paid the same as me but has got the next iteration of hardware, which will inevitably be "better".

When I ordered my S in June this forum was full of members saying "anyone who orders now is nuts" . Mostly because the perception was AP 2.0 would be released in September and maybe even before that. Or at least order in September with a delayed delivery because the hardware would be in by January 2017. Now, i'm not knocking the thinking. I just decided to order because I can make myself crazy worrying about Tesla's next big advancement. I took the leap.
Having said that I am relieved and quite happy with my S which will get 8.0 next week and did not miss out on any hardware changes.

 

[Darthamerica]

There is a lot of speculation that Autopilot 2.0 is imminent--perhaps coming within a quarter or two. There's also a lot of speculation that Autopilot 2.0 will be capable of level 4 autonomy.

I'm going to put myself out there and say: Not a chance.

I should "color" this opinion, as the financial analysts like to put it, by stating that:

1. I am a huge Tesla fan
2. I am a huge Elon Musk fan
3. I am a fan of technology
4. I am a happy early Model S adopter (reservation January 2010, first S delivered Dec. 2012).

But this forum tends to be the king of manufacturing unrealistic expectations. I'll summarize this post right off the bat by saying the following: If you are putting off a Tesla purchase waiting for autonomous driving, you're wasting your time.

You will probably not see AP 2.0 hardware for another few quarters at the very earliest. And that's being optimistic.

You will not see AP 2.0 (what I am using to characterize level 4 autonomy) for about 5 years. And that's being VERY optimistic, even keeping in mind the blindingly fast pace Tesla is moving with this technology.

You will not see level 3 autonomy from Tesla for at least 2 years. And that's being optimistic.

Let me explain.

Autopilot has been out for a year now. While improvements have been remarkable, think about the basic highway driving scenarios that aren't handled yet.

1. Every time I crest a hill, my car dives for the left or right side of the road unless I'm following another car.
2. Every time the lane markings fade, the car drifts and I have to take over.
3. Every time I pass an entrance ramp with cars merging, I have to take control. AP does not handle sequencing itself with merging vehicles.
4. Every time I'm merging myself, I have to take control. The car cannot sequence itself to merge onto the highway.
5. The car does not automatically change lanes to maintain a target speed.
6. The car does not avoid large road obstructions.
7. The car does not move laterally to avoid parked cars on the shoulder, bicyclists, or pedestrians.
8. The car is not always clear about which lane a car ahead is in. Sometimes TACC slows for a car in an adjacent lane.
9. If a car cuts in front of you, the braking is later than most would consider to be comfortable.
10. The car cannot stop as smoothly or gradually as a human would (though something close to this may be coming in 8.0).

These are just a few scenarios in the simplest driving environment--on a highway. While some of these scenarios would clearly benefit from additional hardware, some of them should be perfectly doable with existing hardware: notably 1, 6, 7, 8, 9, and 10. There are more examples of improvements that can come with existing hardware. For example:

-Stoplight and stop sign recognition.
-Automatic speed limit adjustment.
-When following a vehicle, the car should maintain itself over the path that the lead car took over the ground. (This is useful when following a car on a road without lane markings, or when passing through an intersection without lane markings). Instead, it tends to somewhat "cut the corner" and head straight for the lead car, which could put you into the curb or an adjacent lane (or an adjacent car!)
-If following a car into an unmarked intersection and the lead car changes lanes, your car will follow it into the adjacent lane! This will greatly surprise the car next to you and could lead to a bad day.
-In stop and go traffic, if the lead car alternates between moving a few feet forward, then stopping...then moving forward, then stopping...then your car will annoyingly accelerate, brake, etc. Car should be able to recognize that the time average speed of the lead car is low, and glide along gradually at a very slow speed, using little energy or brakes and leading to a smooth slow ride.

These are additional scenarios that are perfectly achievable with existing hardware.

So over a year of autopilot, we've seen improvement in lane holding, smoothness of steering (and, to a lesser extent, braking). contrast, and other things. But there is a *LONG* way to go before we've exhausted AP1.0 hardware capabilities.

So while it's possible that, within a few quarters, Tesla could put out cars with hardware for full autonomy (or at least level 3) and then update them via software over time, I wouldn't expect fully autonomous hardware anytime soon. Why?

1. Achieving level 4 is clearly an interative process. Starting out, Tesla thought they could rely on cameras. After the Joshua Brown accident, they realized the limitation of using cameras as a primary sensor (I believe this caused Tesla to end the relationship with Mobileye) and changed their focus. This will probably happen again. For instance, Elon might be against LIDAR, but he may come around to it if radar processing doesn't turn out to work as well as he'd hoped.

2. Each sensor that's added makes the processing and software that much more involved and complicated.

3. Level 3 and certainly 4 will require an enormous amount of testing and validation. By Elon's benchmark, it has to be an order of magnitude safer than a human driver at the very least. Probably several years worth of testing data once they have a Level 4 system before Tesla says you can ride as a passenger while the car drives. But certainly even longer before the government says it's ok..

4. Building a bunch of cars with level 4 hardware and selling them to customers with a promise that they might be able to use them for level 4 driving 5 or more years down the road doesn't make sense. That would be a money-losing proposition for Tesla, unless there are enough foolish buyers out there to pay for the feature many years before it could potentially even be usable.

Yes, Elon mentioned the car being able to drive from NY to pick you up in LA in about 2 years. (Guess what? As awesome as he is, did anyone ever notice that he's a bit overly optimistic when it comes to time frames?)

Yes, what Tesla is doing is awesome.

Yes, Tesla's gathering FAR more data than *anyone* out there.

But if you extrapolate the improvements we've seen since 7.0 and project that out toward even Level 3 autonomy, you should be able to recognize it will be several years before you can chill out and watch your Tesla automatically merge with traffic and autonomously change lanes. And even longer before it will navigate an intersection and make a turn for you.

Yes, there are "marketing" and "research" videos out there showing what approaches Level 3 and 4 technology. But while you might see something near Level 3 or 4 autonomy in the video, that's a very narrow subset of the curveballs the world can throw at you.

I hate to burst any bubbles, but putting a Level 4 car out there in the real world is many orders of magnitude harder.

Long story short, if you're waiting for the "fully autonomous Tesla" before you put down your deposit, I recommend either buying now or moving on. It's going to be awhile.

It's going to be interesting as hell.

But it's going to be awhile.

I think this is reasonable. I've got about 20,000 AP miles and have seen the gradual improvements. I think there's a lot of room for improvement with the current system before full level 4 autonomy is made available. I won't go as far as to speculate on timing for Autopilot 2.0 because no one but Tesla knows what will be included in that system. Simple things like being able to recognize curbs for example. There just so many things Tesla can do leading up to full autonomy. I badly want level 4 but I want Tesla, others and legislators to do it the right way because this is bigger than just one OEM deciding to just go for it on public highways. However Tesla should have a huge advantage with the amount of real world data they have already collected to date.

 

[RogerHScott]

"Not current" is not synonymous with "obsolete". I think the word "obsolete" gets thrown around with abandon here. Something is
only obsolete if it is no longer reasonably well-suited to its purpose. That seems like an absurd claim to make about a six-month-old
Tesla -- or even a two-year-old Tesla.

 

[u00mem9]

"Not current" is not synonymous with "obsolete". I think the word "obsolete" gets thrown around with abandon here. Something is
only obsolete if it is no longer reasonably well-suited to its purpose. That seems like an absurd claim to make about a six-month-old
Tesla -- or even a two-year-old Tesla.

For many here, obsolete is the correct term. Nobody bought a P90DL because they needed a 2.8 sec 0-60 time and 3.2 wasn't enough. They bought it to have the best. That purpose is no longer served when the neighbor rolls in with his P100D!

 

[edb49]

We know for certain that Tesla are developing "Tesla Vision", a replacement to Mobileye. What incentive would they have for not deploying that as soon as possible to new builds? What functionality do they get from Mobileye?

1) Detect the lane markings
2) AEB for pedestrians/other cars
3) Reading speed limits

I think that it's only actually (1) that is the bit on the critical path. The AEB now involves radar with the V8 update, and I can imagine Tesla releasing Autopilot 2.0 without (3) being in place, promising it will be there in a future update.

Autopilot 2.0 "just" (1) to be done I think. They've already determined what cameras and other sensors to use, and just need to wang them in the car. They'll put in a CPU that has plenty of capability for future improvements to Tesla Vision and the radar.

I understand the "your car will always be obsolete, don't wait" argument. Everyone has their own personal situation. I have already waited almost 2 years, and there is a lot of noise that indicates a significant upgrade is coming soon. If I missed out on Autopilot 1.0 hardware, I would be trading my car in and taking a big depreciation hit - I don't want that to happen with my current order.

 

[RobertF]

I think new hardware will be added sooner than later. What new capabilities are implemented and when, however, will probably be piecemeal. I remember we got one of the first S's with autopilot hardware, but it took a year to implement even the first stage of its use. My major caveat is what effect will the Mobeleye divorce have on this process.

 

[edb49]

I think new hardware will be added sooner than later. What new capabilities are implemented and when, however, will probably be piecemeal. I remember we got one of the first S's with autopilot hardware, but it took a year to implement even the first stage of its use. My major caveat is what effect will the Mobeleye divorce have on this process.

Agree 100% on this. The hardware will be added in a bang, functionality will depend on the software and have significant lag. From my POV of spending 6 figures on a car, I want to have the new hardware as that will make a big difference to me, but it doesn't impact the value of my purchase when the software capabilities come out.