Full Self-Driving - feels like a long way off to me...

[Matias]

I'd buy buckets of food from Jim Bakker before I'd buy FSD from Musk. At least you are receiving a product as-advertised on-time from the ex-con.

TMC (and Google) has taught me many small details about U.S. culture. I add this to them

 

[Matias]

Part of the issue too is that people have different expectations for what constitutes true FSD. For many, they will only accept FSD when it is perfect L5 autonomy that handles every single road in the entire world and every single driving situation always perfectly. If that is your standard, then yeah, you are going to find a lot of reasons to be skeptical. Personally, I define FSD as L4 autonomy.

Musk has defined Tesla’s FSD. You can call your car across the continent or use it as autonomous taxi.

 

[OilSucks]

All this concern about what's not even here yet. People lying awake nights worrying about whether or not FSD will be able to take the off ramp, whether or not they will be able to get to work without touching the steering wheel. Wondering how long before all cars have FSD (That will be a few days...) so they can work together.

What's the matter here? Have people forgotten how to drive?? Do they think they can't make it to work if there's a complicated corner? WHY? You know, you can take over, help the car, or even put both hands on the wheel and do it yourself. Let the car do boring part of the drive, the stop and go. There is no reason to think that you MUST let the car do all the driving. It's STEERING ASSIST. If it ever really becomes Full Self Driving, I probably won't use it, as I don't need it. I learned to drive and still have my license and my full attention span. Maybe 5-10% of the time I can help the car over the rough spots, if I'm not unconscious.

Oh calm down jeez, the guy is just asking a question and inspiring a conversation. Relax a bit. My goodness.

 

[diplomat33]

Musk has defined Tesla’s FSD. You can call your car across the continent or use it as autonomous taxi.

Yes, that is what Tesla aims to make FSD when it is finished. But, I doubt that is what FSD will be when it is first released.

 

[jrad6515]

I pretty much agree with the OP. One thing I thought about the other day is that in this age of declining infrastructure will FSD be able to avoid the many dangerous potholes in the road? If not then I would not use it even if it can get me from point A to point B on its own.

 

[Lasairfion]

Well the more smart vehicles on the road doing a chunk of thinking for their drivers, the easier the driving experience is going to be for the other cars around them. So I think that the more we see semi-autonomy and localised full autonomy, the more likely it is that your Tesla's full-autonomy is going to launched earlier.

Renault recently showcased their EZ-Go fully autonomous transportation pod which they are hoping will be on the road by 2022 and Zoox has been working on theirs and are looking at a 2020 time-line for release.

EZ-Go:

Zoox:

https://www.bloomberg.com/news/feat...startup-zoox-has-800-million-and-a-wild-pitch

 

[diplomat33]

I think it is a mistake to use EAP to judge what FSD will or won't be able to do. EAP is not the FSD software. EAP is just the driver assist package that Tesla created to duplicate the Autopilot software that MobilEye had made. It is not designed to do any self-driving stuff. We should wait until Tesla does release the FSD features and then judge if they are capable of doing FSD or not.

 

[Theflash95]

I think it is a mistake to use EAP to judge what FSD will or won't be able to do. EAP is not the FSD software. EAP is just the driver assist package that Tesla created to duplicate the Autopilot software that MobilEye had made. It is not designed to do any self-driving stuff. We should wait until Tesla does release the FSD features and then judge if they are capable of doing FSD or not.

My point is to look where EAP is now. To go from that to FSD is quite a leap. I feel like while EAP is useful and I use it daily... even that isn’t completely reliable right now. To think that we’re going to go from that to FSD anytime soon is, in my mind, not going to happen. Remember, even with EAP where it is right now it’s not technically reliable enough (even though people do it all the time) to allow us to be completely hands off the steering wheel.

 

[Barklikeadog]

TMC (and Google) has taught me many small details about U.S. culture. I add this to them

The man was head of the PTL Club, that gave us one of the more notorious accounting scandals. Idiots would send in millions of dollars to his 'ministry' so that he could fly around on a private jet, because God wants him to. The PTL built these Jebus theme parks and sell lifetime 'memberships' for loads of money. Well... soon they began selling far more 'memberships' than they had capacity. The board of PTL secretly spent millions on Jim and Tammy Faye. Auditors didn't do their job, people were scammed nationwide, Jim went to jail... and now he is on TV telling you that God wants you to buy his food and poop buckets. Use the buckets as furniture in your house to prepare for the apocalypse!! It's unreal that anyone would buy this garbage from this person, but I'm sure he's raking it in again.

I also hear he is getting into the EV business.

 

[brianman]

I pretty much agree with the OP. One thing I thought about the other day is that in this age of declining infrastructure will FSD be able to avoid the many dangerous potholes in the road? If not then I would not use it even if it can get me from point A to point B on its own.

Another consideration...

As infrastructure decays it needs to be updated and/or replaced. As self-driving becomes more viable, places with significant crowding (China comes to mind) will be motivated to make adjustments to infrastructure to allow more throughput -- autonomous driving is easier if you have government bodies actively wanting to make it easier.

 

[Kant.Ing]

AP2 is still thrashing into other lanes in city intersections, how the missed the mark on just following the normal / current trajectory is beyond me, why would you all of a sudden swerve. 2018.26. Somehow AP1 was able to get this right years ago. It feels like it's regressed a little (AP2)

Hopefully firmware 9 has some kind of major code rewrite only then do I see FSD on the horizon.

City streets and highways need some standardization as well, lines can't just disappear and exits should have dotted lines for the regular lane as they do in some states (I think AZ is one that does).

I hope they will switch back to use the cameras' vision as the primary or even the main source for FSD, and use radar only for TACC. Otherwise we will still experience a lot of phantom collision warnings, phantom braking/AEB, phantom wipers, phantom nags. Things can then improve from there, unless vision turns out not yet a solved problem in Tesla. Cheers!

 

[Electroman]

You have a fantastic product, just sell it, you don't need the extra stuff. Take FSD completely off the table and you will still have 400,000 Model 3 reservations. So why do this? It is only going to lead to lawsuits and bad press. How about focusing on things that really matter to every owner...

100% agreed. All this FSD talk only makes him easy fodder for the haters. He can skip the FSD talk altogether.. or maybe it is too late.

 

[Electroman]

AP2 is still thrashing into other lanes in city intersections

Nope. I have done many intersections and even curved ones too and it goes through like a champ.

 

[d21mike]

I think it is a mistake to use EAP to judge what FSD will or won't be able to do. EAP is not the FSD software. EAP is just the driver assist package that Tesla created to duplicate the Autopilot software that MobilEye had made. It is not designed to do any self-driving stuff. We should wait until Tesla does release the FSD features and then judge if they are capable of doing FSD or not.

Normally I like your optimism and somewhat agree with it. But I do not agree with this. Sounds like you are saying that EAP Build 2018.26 will continue with it's own release numbers and if Elon is correct we will start to see a complete separate track of builds only for cars with FSD? I believe that FSD is additional features on top of the EAP Base AP development. Example: I believe On-Ramp to Off-Ramp is pure EAP and that is what FSD will be on Freeways. And that will improve for both EAP Cars and therefore improve for FSD Cars. But for surface streets all of the improvements will be FSD with very little improvement for EAP Cars. Of course we will see soon if we can believe Elon.

 

[d21mike]

My point is to look where EAP is now. To go from that to FSD is quite a leap. I feel like while EAP is useful and I use it daily... even that isn’t completely reliable right now. To think that we’re going to go from that to FSD anytime soon is, in my mind, not going to happen. Remember, even with EAP where it is right now it’s not technically reliable enough (even though people do it all the time) to allow us to be completely hands off the steering wheel.

When it comes to freeway driving many many (and maybe you) use it for 95% of the time without taking it out of AP Mode. It is that way for me. I could go hands free except for the nags and my own personal wanting to be extra safe. But I believe EAP is primary a freeway driver assist (or full assist 95% of the time) and we need much more for surface street driving and that is where I believe FSD comes in. Remember EAP means On-Ramp to Off-Ramp without driver input which Elon actually said he used the night before to get home when talking about v9 in August. Elon seems to believe EAP for that purpose is close.

 

[chillaban]

FWIW, current EAP behavior is pretty much described by the simple rules of centering between well-defined lane markers and regulating distance to a single lead car in front.

That's probably not what FSD will do, so it's likely misleading to translate current EAP behavior into future FSD behavior.

More concretely, we've seen many updates where braking smoothness got better or collision alarms start happening for new situations but there's no changes to the neural nets. It may be possible that the neural net is already capable of recognizing situations beyond what EAP demonstrates to us, just there's either no software to act upon it, or that software is hidden behind a switch or different codebase.

Path planning for FSD is likely very different in philosophy compared to EAP.

 

[d21mike]

FWIW, current EAP behavior is pretty much described by the simple rules of centering between well-defined lane markers and regulating distance to a single lead car in front.

That's probably not what FSD will do, so it's likely misleading to translate current EAP behavior into future FSD behavior.

More concretely, we've seen many updates where braking smoothness got better or collision alarms start happening for new situations but there's no changes to the neural nets. It may be possible that the neural net is already capable of recognizing situations beyond what EAP demonstrates to us, just there's either no software to act upon it, or that software is hidden behind a switch or different codebase.

Path planning for FSD is likely very different in philosophy compared to EAP.

Could you clarify? Seems like you and @diplomat33 believe that EAP will have some changes going forward but there will be pretty much New Software loaded on cars that have purchased FSD that will replace (not added too) the current software being used for EAP? Like the freeway driving on AP will have EAP Firmware for the cars without FSD and FSD Firmware for the cars that have FSD and the software is not the same code base for AP? As I said I personally do not think this will be the case but of course I could be wrong.

 

[WannabeOwner]

Look where the current autopilot is now... do you really think within 5, 10, 15 years we’re going to have cars with no human controls?

Yup. I expect AI to delivery an exponential rate of improvement. Humans are typically very bad at visualising exponential growth.

For example:

At some point we will have used up 50% of the oil reserves. Let's assume that growth in oil consumption is 10% p.a.

Half the oil is still left in the ground (forget about the fact that it might be harder to get that out than the first half). So the amount of oil left in the ground is equal to all the oil we have ever used in the whole of automotive history. How long have we got left?

Answer = 7 years. Just 7 years.

We find some more oil. Actual we find a deposit equal to the total oil we ever knew about before. How long an extension will that give us?

A = 7 years

You want another 7 years? now you are looking for a reserve that is 4x as big as the original total deposit.

 

[chillaban]

Could you clarify? Seems like you and @diplomat33 believe that EAP will have some changes going forward but there will be pretty much New Software loaded on cars that have purchased FSD that will replace (not added too) the current software being used for EAP? Like the freeway driving on AP will have EAP Firmware for the cars without FSD and FSD Firmware for the cars that have FSD and the software is not the same code base for AP? As I said I personally do not think this will be the case but of course I could be wrong.

I'm not saying I expect the entire codebase to be different. I'm saying I would expect different features to be activated/inactivated depending on your paid features and other conditions.

For example, EAP cars may forever just have EAP mode, where they have a software engine that just gets lane lines and lead car information from the Vision system and centers the car in the lane. While FSD might be a separate engine that takes into account the relative positions of other cars, which lane you need to be in to get to your destination, etc etc etc. And depending on what you paid for, it might activate different things.

All I was saying is that what EAP does *right now* can't be turned around to be evidence that it can't do certain things. Like remember back almost a year ago, when they suddenly pushed a new trigger for "construction zone" snapshots with a magic 4KB blob identifier and all of a sudden AP2 cars were recording video clips of construction cones? Before that day we'd all say "AP can't recognize construction zones" but all of a sudden, they revealed this ability of the neural net (the same one we've been analyzing) and that was a revelation for us.


For all we know, there could be more functionality like this under way that we just aren't finding until Tesla turns it on.

 

[cybergates]

Nope. I have done many intersections and even curved ones too and it goes through like a champ.

does ur city or freeway exits have dotted lines across? Not all cities have this and this is where problems start - there's no standardization to street lines.