Elon: "Feature complete for full self driving this year"

[diplomat33]

But, they could train 20 cities at a time - and if it only takes 2 months per city, they can cover top 200 cities in a year. Afterall the differences between 2 US cities isn't that much (compared to say driving in Asia).

BTW, all the noise that Tesla is making in terms of FSD - is sure to put pressure on Waymo & Cruise to expand faster. They probably have better idea on exactly where Tesla is compared to publicly available information.

Yes, Waymo could certainly deploy in multiple cities at the same time. But right now, their deployment appears to be a bit slow as far as I can tell. Or maybe I am just not up to date on what they are doing.

I do agree that Tesla is probably putting pressure on Waymo precisely because of Tesla's software update capability. Waymo has to know that if Tesla makes that FSD breakthrough, that suddenly every Tesla car will have the FSD breakthrough. It is conceivable that Tesla could get to L4 in every city before Waymo finishes expanding to every city, if Tesla reaches L4 in their dev software and pushes it out to every Tesla in the US.

 

[Mojito]

Maybe they are leaving because they are getting offers by competitors and want to steel Tesla Ideas.

 

[daniel]

Well, just because someone who is not in the industry thinks 'x' means 'y' - that doesn't make 'y' the correct thing. ...

Which is why I pointed out that this all just obscures the fact that until recently Musk was indeed promising that our cars would go off without anyone in them to pick up the kids from school or soccer practice. And implying, if not explicitly stating, that this would happen soon enough that most of the people paying for it could actually use it before trading in their car. He was not talking about "feature complete" software that still needed another decade to work out the bugs. He was talking about what we would actually be doing with our cars. Now, if your interpretation is correct, he's saying that, no, people who paid the extra $5,000 (?) will not be able to send their cars off to pick up the kids, they were just donating money to Tesla to develop software. And paying for a few new features at Level 2 that EAP buyers won't get.

At the same time, definitely Musk needs a lot of improving his communication skills. But then, Musk sold a lot of Tesla's without ads and mostly by getting a lot of press with his personality. We have to take the good with the bad.

Which is why I keep repeating that I love my car and am amazed by what it can do, in spite of Musk's utterly unrealistic claims about what it will do within the next 2 or 3 years.

 

[Green Pete]

One thing that I haven’t seen mentioned. If they launch FSD to the fleet one bug they most likely have to work out not before next March is probably snow.

Meaning they will need drivers with hands on wheel in snowy conditions in the millions to iron out that scenario

 

[daniel]

One thing that I haven’t seen mentioned. If they launch FSD to the fleet one bug they most likely have to work out not before next March is probably snow.

Meaning they will need drivers with hands on wheel in snowy conditions in the millions to iron out that scenario

Oh, they'll get plenty of snow time in the eight to ten years before they really have a working FSD system ready to apply for regulatory approval.

 

[EVNow]

One thing that I haven’t seen mentioned. If they launch FSD to the fleet one bug they most likely have to work out not before next March is probably snow.

Meaning they will need drivers with hands on wheel in snowy conditions in the millions to iron out that scenario

Whatever happened to the shadow mode ?

 

[Green Pete]

Whatever happened to the shadow mode ?

I may be wrong but I think shadow mode would learn significantly faster if the baseline was software that could drive in the city, unlike now.

I could be mistaken tho

 

[EVNow]

I may be wrong but I think shadow mode would learn significantly faster if the baseline was software that could drive in the city, unlike now.

First thing to do is to be able to figure out the driving path (and lanes) and they should have enough images to do that already. Then they need to train the NN to recognize all kinds of objects (including cars) when covered in snow. They should have enough data for that already. Then, they need to adjust driving policy for which they would need the people to be using the actual feature.

 

[EVNow]

As of 7/12/19 this is what Tesla website lists as part of FSD. I've noted the feature number in the FC list with each feature.

Autopilot (Included)

• Enables your car to steer, accelerate and brake automatically for other vehicles and pedestrians within its lane. (2.1, 2.3)

Full Self-Driving Capability

• Navigate on Autopilot: automatic driving from highway on-ramp to off-ramp including interchanges and overtaking slower cars. (9)
• Auto Lane Change: automatic lane changes while driving on the highway. (3)
• Autopark: both parallel and perpendicular spaces. (10.1, 10.2)
• Summon: your parked car will come find you anywhere in a parking lot. Really. (1+)

Coming later this year:

• Recognize and respond to traffic lights and stop signs. (5, 4.1)
• Automatic driving on city streets. (2, 3, 6, 7) <-- This is the ambiguous one, which could mean just a few things or full set of features mentioned in the FC list.

 

[DanCar]

... Automatic driving on city streets. (2, 3, 6, 7) <-- This is ambiguous one, which could mean just a few things or full Level 5.

Very unrealistic to think it could mean Level 5. It means you have to pay attention and if there is a mistake it is your fault not Tesla's. This is Level 2.

 

[EVNow]

Very unrealistic to think it could mean Level 5. It means you have to pay attention and if there is a mistake it is your fault not Tesla's. This is Level 2.

Not talking about quality paying attention - but about what kinds of scenarios are actually handled. For eg. will it handle double parking or person manually controlling traffic or emergency vehicle etc. Also, why is traffic lights and stop lights mentioned separately ? Is it not part of normal city driving.

To remove the confusion, I've updated my post.

 

[diplomat33]

As of 7/12/19 this is what Tesla website lists as part of FSD. I've noted the feature number in the FC list with each feature.

Autopilot (Included)

• Enables your car to steer, accelerate and brake automatically for other vehicles and pedestrians within its lane. (2.1, 2.3)

Full Self-Driving Capability

• Navigate on Autopilot: automatic driving from highway on-ramp to off-ramp including interchanges and overtaking slower cars. (9)
• Auto Lane Change: automatic lane changes while driving on the highway. (3)
• Autopark: both parallel and perpendicular spaces. (10.1, 10.2)
• Summon: your parked car will come find you anywhere in a parking lot. Really. (1+)

Coming later this year:

• Recognize and respond to traffic lights and stop signs. (5, 4.1)
• Automatic driving on city streets. (2, 3, 6, 7) <-- This is the ambiguous one, which could mean just a few things or full set of features mentioned in the FC list.

Thanks for doing this.

Tesla has a somewhat more detailed description of FSD on the Autopilot page that might help you in figuring out what features match "automatic city driving":

"Your Tesla will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed."

 

[diplomat33]

In terms of the FSD features listed on the website, I think there are 2 possibilities:

1) Tesla does intend to finish all features for L5 and the website is just giving us the marketing highlights because listing everything would take up too much space.

2) The features on the website are actually everything that Tesla intends to do for "feature complete" and Tesla plans to add the remaining L5 features later after "feature complete" is done.

Personally, I lean towards #2. I think Tesla is focused on getting the features on the website done first. Those features are good enough for a foundation for FSD. And once they get that foundation done, Tesla can add additional features and increase the reliability to make it true FSD.

 

[rnortman]

It is very clear that "feature complete" in 2019 has nothing to do with L5. It is delivering on the website features, which are a complete redefinition of how FSD was originally sold and how Elon still markets it. FSD on the website right now is little more than EAP used to be.

The easiest way to get to feature complete when you were way too optimistic in your original promises is to vastly reduce the feature scope, which is what Tesla has done here.

I believe they will probably be feature complete for that list, which is still super vague in the one bullet that matters ("automatic driving on city streets"), by the end of 2019. Feature complete does not mean you will want to actually use any of it; it will be a gimmick and will be maddening to attempt to actually use. It will remain so at least through the end of 2020, if not forever (on existing hardware).

If it weren't for Elon nobody would even be talking about L5. Only Elon talks about L5; Tesla does not.

 

[EVNow]

Personally, I lean towards #2. I think Tesla is focused on getting the features on the website done first. Those features are good enough for a foundation for FSD. And once they get that foundation done, Tesla can add additional features and increase the reliability to make it true FSD.

Also, as a more practical matter, as remaining time declines and time pressure mounts - they will have to start looking to implement an "MVP" - minimum viable product. For city driving these basic things would work - incidentally all items Musk has talked or tweeted about at some point.
- summon
- navigate urban streets (even without lane markings)
- manage complex intersections with traffic lights
- stop signs
- roundabouts
- auto park

 

[diplomat33]

Also, as a more practical matter, as remaining time declines and time pressure mounts - they will have to start looking to implement an "MVP" - minimum viable product. For city driving these basic things would work - incidentally all items Musk has talked or tweeted about at some point.
- summon
- navigate urban streets (even without lane markings)
- manage complex intersections with traffic lights
- stop signs
- roundabouts
- auto park

Good point. We really should think of "Feature Complete" as a "Minimum Viable Product". That is a more appropriate name.

I agree with your feature list. I would add that if Tesla doe achieve this MVP, even with driver supervision, owners would have a very useful ADAS, able to handle a lot of their driving on highways and city streets. And if Tesla finishes the OEDR and gets the reliability high enough, it could become a good L4 prototype.

 

[daniel]

It is very clear that "feature complete" in 2019 has nothing to do with L5. It is delivering on the website features, which are a complete redefinition of how FSD was originally sold and how Elon still markets it. FSD on the website right now is little more than EAP used to be.

The easiest way to get to feature complete when you were way too optimistic in your original promises is to vastly reduce the feature scope, which is what Tesla has done here.

I believe they will probably be feature complete for that list, which is still super vague in the one bullet that matters ("automatic driving on city streets"), by the end of 2019. Feature complete does not mean you will want to actually use any of it; it will be a gimmick and will be maddening to attempt to actually use. It will remain so at least through the end of 2020, if not forever (on existing hardware).

If it weren't for Elon nobody would even be talking about L5. Only Elon talks about L5; Tesla does not.

I agree, with one quibble: If it works it will be a lot more than EAP. It will be Super-EAP if it can truly navigate city streets, with turns and stoplight/stop-sign recognition, and bicyclist avoidance, etc. But as you say, it will not be the system Elon promised.

At Level 2, I'm not sure I'd consider it worth the money. But I would pay for Level 3 of the features I presently have with EAP, even not including NoA.

 

[EVNow]

If it weren't for Elon nobody would even be talking about L5. Only Elon talks about L5; Tesla does not.

Actually, Elon doesn't talk about L5 either. He talks about robotaxis and FSD and no geofencing - which may or may not mean L5.

FSD on the website right now is little more than EAP used to be.

What was EAP earlier - I don't remember it ever saying anything about city driving.

 

[rnortman]

I agree, with one quibble: If it works it will be a lot more than EAP. It will be Super-EAP if it can truly navigate city streets, with turns and stoplight/stop-sign recognition, and bicyclist avoidance, etc. But as you say, it will not be the system Elon promised.

Recognizing stop lights and stop signs is not super valuable nor super useful on its own, and likely would fall under a safety feature and therefore would be part of both EAP and the base vehicle package. Ditto for bicyclist avoidance; that would need to be in effect any time Autosteer and/or TACC were in use, since that's a safety feature. (And EAP, for better or worse, allows you to enable Autosteer and TACC on local roads.) Navigating city streets (intersections) is the big thing added to EAP. But it will be L2, and will be a useless gimmick, and therefore I still stand by it being "barely more than EAP" in terms of actual practical value.

At Level 2, I'm not sure I'd consider it worth the money. But I would pay for Level 3 of the features I presently have with EAP, even not including NoA.

Holy cow, L3 on the highway even without NoA (so just stay in lane, don't take exits, alert driver with enough warning to take over if anything else is required) would be amazing if it actually worked. The original description of EAP implied L3 ("on ramp to off ramp with no intervention required by the driver" or something like that), which does admittedly give them plenty of wiggle room. But a lot of people who bought EAP read that as "L3 on the highway". In fact, on these forums, there were plenty of people arguing that EAP was L3 or even L4 on the highway when it was first announced.

What was EAP earlier - I don't remember it ever saying anything about city driving.

See above for a little more detail on what I mean by "barely more than EAP". I didn't say "nothing more than EAP". What I mean, really, is "little practical value beyond EAP", even if it checks off all those checkboxes. It is my assumption that city driving will be an L2 gimmick which is uncomfortable to ride in, surprising to other drivers, and requires constant intervention in most real city driving situations -- and therefore has very little practical value.

Wouldn't it be great for us all if I am wrong... but I'm not.

 

[EVNow]

See above for a little more detail on what I mean by "barely more than EAP".

There will always be people who won't like it and people who do. I for one will definitely use city NOA and will likely find it useful.

Currently with AP I have to disengage every time I'm the first at the red traffic light or at stop sign or when I have to turn. I'd consider much better than "barely more than EAP", if it handles just stop light & traffic light. Turns will make it even better. YMMV.

ps : For my current commute NOA is not very useful But I won't claim NOA is barely more than EAP. A lot of people use it and find it useful.