I am now planning to get the full self-driving option and here is why

[Dave B]

...and to confirm that I did what I said I was going to do, FSD option checked

 

[ModelNforNerd]

I know the consensus is that getting the FSD option now does not make sense. The argument goes that we have no idea when FSD will arrive, it could be years or maybe "never". Then there is the argument that we don't know if the hardware is even good enough so why spend money on an option if we later find out, it requires new hardware. So the argument is it is better to wait and see. I used to agree with that consensus. But the recent conference call has changed my mind. I am now planning to get the FSD option on my Model 3 unless something changes between now and when I get my car.

Here are my reasons:
1) Tesla is planning the FSD coast to coast demo in 3-6 months. Musk also said the neural net is solid and will learn exponentially. Both those statements taken together make me optimistic that they are close to getting FSD working. Now, I know that Tesla misses deadlines. After all, Musk said "3-6 months" for AP2 to diverge from AP1 and we all know how that panned out. So, yes, it is possible that the FSD demo will get pushed back again. But even if they do miss the 6 month deadline to do the demo, I think they are pretty close now to getting FSD working. I definitely don't think FSD is decades away as the skeptics claim.
2) I plan to finance my Model 3, pay it off and keep the car for as long as it lasts. So even if FSD is still a few years off, I think I will see it happen in the lifetime of my car. So getting it now will save me $1000.
3) When Tesla does release FSD, I think it will be AMAZING! Call me a fanboy, but I think Tesla will impress when they do finally release FSD. And I want to be ready to get it as soon as it is released!

Having said all that, I am still about 10 months away from getting my Model 3 according to my estimated delivery date. I will pay close attention to any FSD news in the coming months. There is still plenty of time for things to change for better or for worse. And if bad news comes out about Tesla's FSD in the next coming months, I will of course re-evaluate my position. But right now, count me as back in the "planning to get FSD" camp.

While I will also be getting FSD, it will come down the road for me....when it's available for download.

Here in MA, I would actually pay sales tax both on the initial purchase of the car, as well as in my annual excise tax on the residual value of the vehicle.

However, if I purchase after the fact, I only pay sales tax on the transaction of purchasing the software. I figure, over time, the difference will be made up somewhere in there.

 

[ModelNforNerd]

That's the biggest part of the level 2 to level 3 leap, and likely the reason no one has really gone that far yet.

On a level 2 car, the driver is clearly responsible for any accident, even if the level 2 system is active at the time of crash.

On a level 3 car, the manufacturer is responsible if the accident occurs while in self driving without an alert - or even within the alert period specified.

It'll be interesting to see how the insurance piece of that works - Tesla buys insurance for the fleet for those conditions, and you get cheaper insurance for the rest of the time?

Actually, this is the sticking point right now.

NO ONE is responsible if a car in Level 4 or 5 (according to SAE Levels of Autonomy) crashes. It still needs to be sorted out.

 

[RedEagle]

My point of delight in the future will be when our X says, "I'll be turning left at the next intersection onto xxx st." and I have the option to let it without any intervention or override the turn by intervening (voice override?)

 

[xav-]

FSD will take a decade plus to come to fruition. What we can possibly expect in a year or two is some additional capability to drive in more complex situations than what AP does today.

I am *extremely satisfied* with the current state of AP today. For $5K AP is worth every penny. More important than all other options - AWS, PUP, SAS - combined, except LR.

For those who are gambiling on getting FSD - later don’t tell you have not been warned. Don’t come back and complain here if you never see it within the lifetime of your car.

Musk is visionary and highly intelligent, but I am not willing to dismiss others - almost everyone chasing FSD - who are pursuing LIDAR. This argument of to-LIDAR-or-not is not an open and shut case like the fool cells Toyota is pursuing.

It should be obvious that the problem that Tesla is trying to solve would be 100 times easier to solve if they used Lidar.

It should be pretty obvious that having a full 3D cartography of your surrounding would make the system 100 times more reliable and safe than relying on image recognition to vet objects.

Everybody agrees.. mobileye agrees and they should know more than anybody else on cameras. That's why they canceled their contract with Tesla. Google agrees. The only ones who disagree are Elon Musk and Elon Musk hardcore fanboys.. that's it.

If you guys were coders in the Autopilot team I am pretty sure you would have a very different view on this. The several Autopilot executives who resigned for sure did.

And by the way... I am not making the claim that cameras or radar are useless. That's not the claim I am making.. they are essential as well. What I am saying is Lidar is not optional.. unlike what Elon Musk is saying. You can't do self driving with Lidar alone, and you can't do it without it.

Elon Musk has been making this claim for years by the way.. and where have we been exactly?

 

[Brian Anderson]

That's the way I look at it too, and also why I'll be getting FSD when I get my 3 "between March and June 2018."

If you want to see something happen sooner - contribute towards that goal. When it comes to FSD, I know that my programming skills to make that happen would lead to more deaths, not less. That's why I'll give them my money instead.

I seriously think the EAP is overpriced for what it is, and the FSD (if features actually working) could/should be 5K instead of 3K.

My reasoning is no other car manufacturer charges such a huge premium (5k) for features like EAP. Lets be honest, its basically traffic aware cruise control at this point. Its nice, and yes probably the best implementation of any manufacturer; but recent cars I've owned like the Mazda CX-9 and the BMW i3 have just come with that tech as part of tech packages of which the entire bundles that included tons of other physical upgrades were way less than 5k.

Assuming all features were up to par and working; EAP should be a 2-3K option and FSD should be the 5K option.

Disclaimer: When I bought the Model X last Sep, I opted for the EAP, Its fun to play with, but I'm seriously considering dropping it when I configure my Model 3, I just don't know if I can justify paying that much again for like I say essentially traffic aware cruise control. Or at the very least I could put it this way, I'm happy to be paying the 49,000 for the model 3, because I think it is worth the value; but I'm not happy as I'll be paying the 5,000 for the EAP, because I think its overpriced.

The Bolt Premium trim is over 4k more and basically adds basic AP-type features (not as good though).

 

[Brian Anderson]

Funny how FSD is either available later this year or not available for decades. As in all polarized arguments, reality will play out somewhere in the middle. I believe Tesla will have a version of FSD that can be put before regulators in late 2019. It doesn't have to be perfect (and won't be), but upgrades will follow an exponential curve as all technological developments have throughout history.

I am back to buying FSD with my initial order because I agree that Elon's comments implied that the cost of upgrading later will be higher than the $1000 currently in place, and that owning the feature will give preference for HW upgrades, if needed, that make it work optimally in the future.

 

[essmd]

OMG, this thread demonstrates the perpetual mental illness afflicting "believers" that will continue to look to Elon and Tesla with hope and optimism, despite over 3 years of amassing thousands of frustrated owners, mostly purchasing a Model X.

Perhaps it's a self-inflicted delusional syndrome, but just review the topics of threads in the Model X section.

What started with the announcement in October of 2016 of HW 2.0, offering Enhanced Auto Pilot 2.0 and Full-self Driving capability, every Model X delivered after that is an example of False Promises, Vaporware, and Broken Dreams.

Despite the extra hardware, (stronger radar, more cameras, 7 more cameras, faster CPU), there still is no E, or Enhanced feature of the AP 2.0, in fact, it still is not as good as AP 1.0, (owned both for almost a year each) and certainly no FSD...probably will NEVER be for 2.0, or perhaps even 2.5

However, if Elon promised free trips to Mars for all Tesla buyers after March 1, there would be a 6 month delay for all models and thousands packing their spacesuits and Tang (reference to us over 50).

Eerily, while each Gallery is like a Tesla Temple, filled with young folks spreading the Tesla gospel, the Superchargers are populated by owners all too eager to share their aggravation with Tesla Ownership, but the perspectives are widely variable based on the equally variable Tesla build quality and commitment to owner satisfaction.

Such a potential Love-Hate relationship sold with each Tesla... I love driving mine, but the owner experience can really vary widely....

..... but don't let me spoil your Model 3 dreams, may The Force be with Your Tesla....

 

[xav-]

Anyway I'm looking forward to my model 3 with Autopilot... not because I expect the thing to drive me around.. but because I think it will be a cool assist feature.

I just would put a grain of salt on any predictions about FSD. My take that trip from LA to NYC won't happen or it will be highly staged (just like they staged their AP20 video at the end of 2016)

 

[dss33]

It should be obvious that the problem that Tesla is trying to solve would be 100 times easier to solve if they used Lidar.

It should be pretty obvious that having a full 3D cartography of your surrounding would make the system 100 times more reliable and safe than relying on image recognition to vet objects.

Everybody agrees.. mobileye agrees and they should know more than anybody else on cameras. That's why they canceled their contract with Tesla. Google agrees. The only ones who disagree are Elon Musk and Elon Musk hardcore fanboys.. that's it.

If you guys were coders in the Autopilot team I am pretty sure you would have a very different view on this. The several Autopilot executives who resigned for sure did.

And by the way... I am not making the claim that cameras or radar are useless. That's not the claim I am making.. they are essential as well. What I am saying is Lidar is not optional.. unlike what Elon Musk is saying. You can't do self driving with Lidar alone, and you can't do it without it.

Elon Musk has been making this claim for years by the way.. and where have we been exactly?

When you're manually driving a car, do you have some form of organic LIDAR?

That's the point Elon's made again and again. We have a ton of experience and understanding of our environment as humans, however we drive with a CPU/GPU/neural net (brain), two cameras (eyes), and two microphones (ears). That's it. If we can't teach a car to do that with eight eyes, sonar, radar, a CPU/GPU/neural net, and lots of mapping data from cars that have driven that same route before, then we're doing it wrong.

They're different in many ways, but remember when Elon decided we needed reusable rockets and everyone else was doing it differently and thought he was crazy? My money's on Elon here.

 

[xav-]

When you're manually driving a car, do you have some form of organic LIDAR?

That's the point Elon's made again and again. We have a ton of experience and understanding of our environment as humans, however we drive with a CPU/GPU/neural net (brain), two cameras (eyes), and two microphones (ears). That's it. If we can't teach a car to do that with eight eyes, sonar, radar, a CPU/GPU/neural net, and lots of mapping data from cars that have driven that same route before, then we're doing it wrong.

They're different in many ways, but remember when Elon decided we needed reusable rockets and everyone else was doing it differently and thought he was crazy? My money's on Elon here.

But humans have something computers do not: intelligence

I know Elon is all concerned about computers taking taking over humans, but sincerely judging from AP20's behavior I think robots outsmarting humans is not anytime soon. Joking aside, Artificial intelligence is still at its infancy.

At the end of the day.. it's a far simpler job to do with Lidar than without it. By the way a 360 degree Lidar/radar/camera system is far safer than a human.

What we know:
- google car works
- AP 20 is a nice assist feature

I trust google more on software than Elon Musk.

again... where have we been exactly in 3 years?!

 

[dss33]

But humans have something computers do not: intelligence

I know Elon is all concerned about computers taking taking over humans, but sincerely judging from AP20's behavior I think robots outsmarting humans is not anytime soon. Joking aside, Artificial intelligence is still at its infancy.

At the end of the day.. it's a far simpler job to do with Lidar than without it. By the way a 360 degree Lidar/radar/camera system is far safer than a human.

What we know:
- google car works
- AP 20 is a nice assist feature

I trust google more on software than Elon Musk.

again... where have we been exactly in 3 years?!

Have you ever driven/ridden in a Google Car? Or even seen one for that matter? I haven't and don't know anyone who has (and I live about 20 mins from where they test them here in Washington). I'm sure they're doing good work, but they've been at it for over eight years now...

Remember that in those three years, they had a nasty breakup with MobilEye and had to basically recreate in a year or two what MobilEye had built over the course of more than 15 years. And they had a number of their main AI folks leave or get poached, so it's not like they haven't had huge obstacles and setbacks. And three years in software dev isn't that long. You may see these are excuses, but it's just being realistic.

I'm not saying that Tesla doesn't have a long way to go with their software and AP system, but they're the only company with millions of real cars on the road running features in shadow mode, collecting real-world data, and providing mountains of feedback, so that's why I'm good putting my money there. If what Elon says is even close to true, we'll see pretty significant progress this year as they lean more and more on the neural net they've been constructing. But hey - only time will tell...so we'll see soon enough.

 

[Uncle Paul]

I see posts saying FSD will only be possible with Lidar.

All the Lidar equipped vehicles seem to have some large bubble on the roof or on a stalk. Would Tesla owners be willing to buy a car with one or more of these things sticking out of their car.

Can the Lidar be integrated into the vehicle so it would be acceptable to most of the buying public.
What would be the aero drag on these bubbles at 80 MPH?

Elon has made a good case that Lidar is not the long distance answer, but mostly only for tiny city cars with lower speeds.

 

[RayK]

I'm not saying that Tesla doesn't have a long way to go with their software and AP system, but they're the only company with millions of real cars on the road running features in shadow mode... <snip>.

Ah, reality/hyperbole check: I sure hope that Tesla does not have "millions of real cars" on the road as that would mean the federal tax credit would have been eaten up by now. Even accounting for the number of cars that were sold outside the US. Now if you meant "millions of real car miles", then okay.

 

[Dr. J]

Would Tesla owners be willing to buy a car with one or more of these things sticking out of their car.

 

[S4WRXTTCS]

I consider myself one of the main detractors of FSD, and in fact I think it was one of the three things that happened in 2016 that signified the end of reality itself.

But, when it comes time to ordering it's going to be hard not to get it.

I'll get it despite the fact that

1.) It doesn't have the sensors needed for FSD. In fact it doesn't even have the sensors required to safely backup. Heck a Subaru Outback with cross traffic radar can back up more safely. The camera's won't do any good if they're dirty, and the ultrasonics miss stuff all the time.

2.) We don't know what regulatory requirements will require for level 4 or level 5 self driving. You can't declare something as self-driving capable when you don't even know what the requirements will be.

3.) That I know full well that self driving cars have to be magnitudes of degrees safer than human drivers to be allowed.

I'll get it because I strongly believe that Tesla will throw FSD owners some kind of bone. Where there either going to do free upgrades to the drive computer. Or it will be capable of taking over in a lot more situations than EAP only. Like lets say I was coming to a 4 way busy stop, and I'm like ughh I don't want to deal with this and I turn it on to have the car deal with it. Even if i had to act as a second set of eyes it would still be worth it.

It would be like my little driving buddy.

 

[Jays200]

Actually, this is the sticking point right now.

NO ONE is responsible if a car in Level 4 or 5 (according to SAE Levels of Autonomy) crashes. It still needs to be sorted out.

The car just becomes the insured entity.

Obviously it has to be as if no one is driving and/or no one is in the car then the car, by necessity, must be insured.

My even be cheaper than insuring a human driver once Level 5 has some stats behind it.

 

[xav-]

Have you ever driven/ridden in a Google Car? Or even seen one for that matter? I haven't and don't know anyone who has (and I live about 20 mins from where they test them here in Washington). I'm sure they're doing good work, but they've been at it for over eight years now...

Remember that in those three years, they had a nasty breakup with MobilEye and had to basically recreate in a year or two what MobilEye had built over the course of more than 15 years. And they had a number of their main AI folks leave or get poached, so it's not like they haven't had huge obstacles and setbacks. And three years in software dev isn't that long. You may see these are excuses, but it's just being realistic.

I'm not saying that Tesla doesn't have a long way to go with their software and AP system, but they're the only company with millions of real cars on the road running features in shadow mode, collecting real-world data, and providing mountains of feedback, so that's why I'm good putting my money there. If what Elon says is even close to true, we'll see pretty significant progress this year as they lean more and more on the neural net they've been constructing. But hey - only time will tell...so we'll see soon enough.

Waymo 4 million miles self driven!

How many miles do you think a Tesla with AP20 would self drive itself without causing a major accident??

 

[xav-]

I see posts saying FSD will only be possible with Lidar.

All the Lidar equipped vehicles seem to have some large bubble on the roof or on a stalk. Would Tesla owners be willing to buy a car with one or more of these things sticking out of their car.

Can the Lidar be integrated into the vehicle so it would be acceptable to most of the buying public.
What would be the aero drag on these bubbles at 80 MPH?

Elon has made a good case that Lidar is not the long distance answer, but mostly only for tiny city cars with lower speeds.

It's a good question. I think the main reason why it's ugly is not because it's Lidar but more because it's 360 degrees. Maybe a forward Lidar only would look a lot better. I know Audi is planning to have a Lidar on their A8 this year and it doesn't look like that at all in the demos I have seen.

 

[xav-]

I'll get it because I strongly believe that Tesla will throw FSD owners some kind of bone. Where there either going to do free upgrades to the drive computer. Or it will be capable of taking over in a lot more situations than EAP only. Like lets say I was coming to a 4 way busy stop, and I'm like ughh I don't want to deal with this and I turn it on to have the car deal with it. Even if i had to act as a second set of eyes it would still be worth it.

It would be like my little driving buddy.

Isn't that what EAP supposed to be in the first place?