Tesla Autopilot 2.0 Feature Parity With 1.0 Still Months Away

Tesla Autopilot

NOV 4 2017 BY STEVEN LOVEDAY 21

Tesla Autopilot

Tesla Model 3 Autopilot Engaged

While Tesla is potentially making great strides with its proprietary Autopilot 2.0 system, it will still be awhile before parity with the original system is achieved.

Most Tesla owners aren’t too thrilled that the automaker has taken over a year to bring its second-generation Autopilot software to parity with the Mobileye’s Autopilot 1.0, and rightly so. Especially since the automaker claimed to have completed its own system in just six months (six months ago), and has been saying that it would be up to parity soon.

Though the automaker has built its own system from the ground up, following the Mobileye split, it’s still missing key features, and many say that it just doesn’t work quite like the original. Tesla assures that the system will continue to receive incremental over-the-air software updates, and the final product should blow us away.

Tesla Model 3

It comes as no surprise that McNeill has earned his current role. He’s definitely a “people person” by all definitions.

We recently reported about Tesla’s President of Sales and Service, Jon McNeill, speaking in Amsterdam about the upcoming Model S and X-based Mobile Service vehicles. According to Electrek, he answered a few more important questions, and some directly related to Autopilot 2.0 parity concerns.

McNeill admitted that getting the software up to par has been a challenge, but he also spoke about experiencing a new iteration of the system firsthand, and he’s beyond impressed.

Musk has made similar comments in the past about the level at which the updated system is expected to function. Apparently, it’s just a matter of excessive patience at this point. Hopefully, it will be well worth the wait. When asked when Autopilot 2.0 will have feature parity with 1.0, McNeill seemed to suggest that it will be weeks or months. He responded (via Electrek):

“Writing software for the new [Autopilot 2.0] hardware stack has been challenging, but we have some really great progress to reveal to you over the current quarter.”

His main point was that the automaker is exercising extreme caution to assure that everything is working correctly. He also noted that new features are already up and running in shadow mode. Once a feature is working in shadow mode, Tesla puts it through millions of miles of real-world testing prior to fleet-wide activation.

A Tesla spokesperson elaborated/clarified McNeill’s statements:

“We’re working super hard on new features, and rapid progress is definitely being made. We are continuing to refine our existing features and will be rolling out new features as soon as they’re validated. We’re grateful for everyone’s patience.”

Due to over-the-air updates, and the automaker’s online access to every vehicle in its fleet, it can log millions of miles of data in a matter of days. Despite the fact that Tesla owners have to deal with these type of waiting periods, unlike legacy automakers, the cars can be updated constantly. No service center visits or recalls are needed for such updates, and features can be immediately added in shadow mode, tested on public roads through the connected fleet, and then universally implemented … for the life of every Tesla vehicle.

Sure, it’s a huge hassle and it should’ve been completed by now, but when you’re buying into the future (new technology and non-traditional practices), it’s something that you’ll probably be dealing with indefinitely and inevitably as a Tesla owner.

Source: Electrek

Categories: Tesla

Tags: , ,

Leave a Reply

21 Comments on "Tesla Autopilot 2.0 Feature Parity With 1.0 Still Months Away"

newest oldest most voted

“His main point was that the automaker is exercising extreme caution….”

Really? That’s what you got out that comment?

It should also be pointed out that while there are features not at parity, there is also a long list of new features that have been enabled in the same time:

New Features:
Keyfob Profiles
Supercharger Ammenities
Detailed Equalizer Settings
More detailed Fan Speed Options
Model X Headrest Control
“Favorites” Radio Station
Custom Voice Navigation Destinations
Adjusted range estimation
Supercharger Power Availability
Homelink Distance Control
True Camper Mode
Manual Narrow Streets Mirror Fold

All good stuff, but this article is talking about Autopilot features.

Yes, that’s my point. AP isn’t the only feature on the Model S/X. Just adding perspective.

There’s such a thing as a warped perspective, and adding to it isn’t generally a good idea.

Said it before, repeating the FACT. “Autopilot” it is NOT. It will not drive your car automatically and yes, I know all the caveats Elon has put in place but it is in actual FACT, “Driver Assist”.

Yes, it is less fallible than the average motorist ‘cos it doesn’t drive drunk, on drugs, tired, aggressively, or whilst texting or reading a paper map… but it’s still fallible.

It amazes me how people seem so willing to show in public like a badge of pride, that they completely misunderstand what autopilot does in planes, and don’t actually know what the term autopilot actually refers to.

I agree!! Similarly, I was really pissed when I tried Apple’s “Air drop” only to find out that my photos were not, in fact, dropped from an aircraft over their destination, as the name implies. What is it with these tech companies and their cute feature names anyway???

So a year later and they still don’t have feature parity with version 1.0, while

* Tesla still has to significantly improve computer vision, because you know.. no LIDAR (not requirement, but handy).
* MobilEye is not standing still since the split year ago and since they have trouble with feature parity, I don’t think they are as lame as many say
* GM will have Supercruise on Cadillac CT6 2018 (to be released in fall 2017)
* Audi A8 Traffic Jam Assist will have Level 3 (limited to 60km/h), but at least in Germany, they don’t have to pay attention every second (Reuters: “Germany adopts self-driving vehicles law”). To be released this year/early next IIRC.
* There is Nissan with ProPilot (to be released 2018)
* Waymo has moved so far, that real-world miles are of very limited use (they are mostly boring miles with no added value) and uses simulation for nearly everything.
* Tesla is quite mum on their promise of coast to coast before end of 2017.
* I was far more impressed with Nissan autonomous demo in London than with Tesla’t one. Because Nissan’s involved journalists and urban traffic.

Somehow, I don’t believe that they are the leader of the autonomous pack.

At this time, Tesla still remains at the top.
And more importantly,I think that with their update approach, that they will be able to stay ahead of the rest.

The only thing that Tesla autopilot is ahead in is the number of deaths it has caused.

The only thing true about your post is you are a troll and a new username here.

Hey Fox is showing Trump via Facebook, don’t missed.

“Tesla is quite mum on their promise of coast to coast before end of 2017.”

Actually, Tesla mentioned that in their shareholder letter for Q3 just a couple of days ago. It was buried in all the other news, but they mention it here:

“we expect a rapid rollout of additional functionality over the next several months and are progressing rapidly towards our goal of a coast-to-coast drive with no one touching the controls.”

They aren’t putting a specific date on it yet, but do you know of any company who has beat them to it? Or that is even talking about doing it before Tesla?

A year later, and MobilEye isn’t in any cars that have feature parity with AP1 either.

In fact, Supercruise, Traffic Jam Assist, and ProPilot won’t be at parity with AP1 when they are released either. Waymo and Nissan don’t have a system with AP1 parity headed to market either.

To meet AP1 parity that Tesla is approaching, they need all of these features. Keep in mind that the current AP2 cars have about 3/4 of these features fully enabled, and of the rest of the features all but 2 are at least partially enabled:

Traffic Aware Cruise Control
Automated Headlights
Automated Highbeams
Automated Windshield Wipers (DISABLED)
Autosteer (Highways)
Speed Limit Detection/Warning
Autosteer (Local Roads)
Basic Summon
Parallel Parking
Perpendicular Parking
Side Collision Warning
Side Collision Avoidance
Forward Collision Warning
Automated Emergency Braking
Automated Highway Lane Change
Automated Local Road Lane Change (DISABLED?)
Lane Departure Warning
Engage AP/TACC at 0mph
Different Display Icons for Different Vehicle Types
Display Vehicles in Adjacent Lanes
Automatic Display Brightness

“…I don’t believe that they [Tesla] are the leader of the autonomous pack.”

I agree that it appears Waymo is way out in front, regarding the capabilities of their prototypes. What Tesla leads in is actually putting semi-autonomous driving systems into actual production vehicles. In that, Tesla is so far ahead of the pack that it has no real competition.

Nice, when all the companies released what you say, we may compare, now is as truth as the WV twenty EVs Tesla killers that coming up soon.

It’s good the Autopilot part of Autopilot 2 is about identical to Autopilot 1. The car adaptive cruises and autosteers fine. While missed, features like auto-wipers and other much less immediate aids are drowning out the progress of the underlying system.

I’d be bummed, not to have my auto-wipers, but would think of that more as a “missing feature”, than not being on “parity”.

To the people talking about other automated driving systems out there, there have been a few side-by-side tests of different automakers’ solutions in production cars. Tesla has won all of them for having the smoothest and most fully-featured automated system thus far.

I don’t doubt that GM, Waymo, Nissan, and others have prototypes that are far more capable. But for production cars in actual driving conditions, Tesla is the current leader.

Yes, and Tesla has better stuff in development than what they have released too. In fact, some of it has been being run in shadow mode in real world driving by a whole fleet of drivers collecting data.

That is where Tesla has a systemic advantage over all those car makers who don’t have OTA. Tesla is developing their product with so much more real world data, and so many more miles run in shadow that it is going to be hard for other companies to catch up.

Tesla gets results back from running in shadow, includes some code updates in the next OTA release, and then they get millions of test miles and shadow results.

I’m puzzled to still see articles on Tesla developing Autopilot HW 2.0. Tesla is now putting HW 2.5 into production cars, and it is reportedly working on HW 3.0.

https://teslamotorsclub.com/tmc/threads/hw2-5-capabilities.95278/page-59?utm_source=threadloom&utm_medium=email&utm_campaign=ed9&utm_content=iss16#post-2372408