Tesla Autopilot 2.0 Update Imminent, But Software Update 8.1 Delayed



Tesla Autopilot 2.0 update should come out tomorrow

Tesla Autopilot 2.0 update should come out tomorrow

Autopilot software version 8.1 and the P100D Ludicrous "Easter Egg" speed upgrade may take another month

Autopilot software version 8.1 and the P100D Ludicrous “Easter Egg” speed upgrade may take another month for Autopilot 1.0

Tesla CEO Elon Musk took to Twitter to let customers know that the Tesla Autopilot 2.0 update may be ready to come out tomorrow. Tesla is doing everything it can to honor the promised timeline and get the update in place. However, software version 8.1 must be pushed back another month.

Musk explained that they are currently working on a fix that addresses an issue when the systems boot on “subzero cold-soak.” Unfortunately, the issue is also pushing back the much-anticipated Tesla P100D Ludicrous “Easter Egg” speed upgrade, but if the fix works, the upgrade will follow in the coming days.

Musk explained that the updates that are coming along in a day or so are part of the process of getting Autopilot 2.0 set up to reflect the features from the first version, that have thus far been missing.

As usual, some updates, like those to the user interface and media center, will affect all vehicles. In late January, all vehicles will get the software version 8.1 update, along with the Linux kernel upgrade. The Linus kernel upgrade should provide substantial speed and overall interface improvements.

via Electrek

Categories: Tesla

Tags: , , , , ,

Leave a Reply

25 Comments on "Tesla Autopilot 2.0 Update Imminent, But Software Update 8.1 Delayed"

newest oldest most voted

Sweet. New kernel FTW.

Sweet. AP would be nice on a road trip, but I just noticed that from Hobrook, AZ to Albuquerque, NM is 235 miles. That means that the Model 3 would have to have a battery that will at least provide a little over 235 miles or I’d be sitting on the side of the road playing with the AP controls while I wait for a charge??????

Or driving at 65 mph or slower.

There’s a super charger in Gallup.

Do you really want to trust your driving to something that was released because it had to be released today (in this quarter) instead of later?

Tesla muffed the timing of this swichover. I hope it doesn’t cause anyone to get hurt.

The faked “concern” of an anti-Tesla troll posting FUD is duly noted.

I swear, it’s like there is some sort of contest between anti-Tesla FUDsters to see who can be the most obvious, the most obnoxious, and the least likely to actually influence anyone’s opinion.

Huh. So when you like a company them squeezing safety-critical software out on a schedule instead of releasing it when it is done is a good thing?

And regardless of whether it comes out in the next few hours, how do you suggest that Tesla didn’t muff this switchover? They have shipped cars without their previous safety features for 3 months because they cut the hardware over before the software was ready. Even if the software comes out today and IS PERFECT they still muffed it because of that 3 month gap.

What is it with pro-Tesla people that they not only willfully overlook huge errors by the company but declare anyone who doesn’t anti-Tesla and a troll?

No one here believes in the spontaneous generation of complex, finished products that combine multiple hardware sensors, 2 generational platforms, including a new kernel with updated software libraries.

Reasonable people understand that good software takes time, and will cut them some slack. What Tesla is accomplishing, is no simple thing.

So your “concerns” and “hopes” come off as disingenuous negativity.

I’m sorry, what? Again, it’s not even just concern. If you release a car with a major reduction in features from the same model a day before you screwed up already. It doesn’t take me being negative about Tesla taking risks rushing software to turn this into a negative situation.

Both of you are being ridiculous.

The right way to handle this would have been for Tesla to not screw up their relationship with their current supplier before their new software was ready. Trying to sate that I’m being unreasonable and suggesting that Tesla be able to create new software in an instant is just distracting from who actually screwed up here.

Good software does take time. That’s why you don’t cut ties with existing suppliers and then put an arbitrary 3 month deadline on your safety-critical software. Tesla did it. That was an error. And all the attempts to try to indicate somehow I am the problem don’t change that at all.

Software is written and updated all the time. Tesla didn’t do anything wrong here. Please keep your negative comments to yourself. Their tech is abosolutly state of the art. Funny how you don’t see a bunch of videos and write ups on Takata airbags but 1-2 Tesla autopilot incidents and it’s headline news.

Tesla did something wrong here. These cars have had no autopilot, no automatic braking, etc. Tesla changed their cars months ago to have new hardware for these features but the software wasn’t ready yet.

This isn’t an “update” to those features for these cars, it is a new feature. Tesla shipped cars without these features and charged customers thousands of dollars each for them. This is bad business. It’s bad business even if you are a Tesla fan.

And it’s unclear which kind of authority you are referencing when you try to tell me what comments I am not allowed to make.

Unlucky, nobody here is fooled by your pretended concern. Its obvious, you aren’t a Tesla owner. (All the Tesla owners I know love their cars, even though they waited months to get one). Do you understand the concept of patience, when it comes to waiting for something great? No? Obviously, you never waited 9 months for a Tesla or a child to be born!

Tesla buyers are not a bunch of yokels who need to buy a car off a dealer lot. Tesla buyers who are willing to wait months for a car to be ready, don’t mind waiting another month for an autopilot version to be finished. They are willing to be patient, to get the most advanced piece of software in the world. They know its a work in progress and can still enjoy driving the car just like any other: manually, duh.

I am not a Tesla owner. Pretended concern? Is this what it’s down to now? No, I don’t understand the concept of being charged for something the car can’t do. And it wouldn’t matter if I did. The problem isn’t on my end. It is bad business to remove a feature from your car due to a snit with your vendor. And then to still sell the car with a promise of that feature later (and even increase the price) is worse. Tesla removed features, charged more for them and then promised them in December. Well, December 31. Well, that turned into a week from December 31. Oh yeah and you didn’t really get the feature promised anyway, the version which is coming for Jan 6 only auto steers up to 35mph! I can understand a customer who says “well, I’ll just wait on buying the car until the feature is back”. But I cannot understand a customer who would buy the car with a promise which may never be fulfilled. And Tesla shoving out software which is incomplete, underdelivers and only to 1,000 cars so they don’t have an accounting problem this quarter (you cannot legally book revenue for… Read more »

I think Tesla is going to need to add LIDAR and maybe a few more sensors before they really become FULLY autonomous.

You mean a LAWYER? 😉

LOL. No, they’ll just screw up things.

Well certainly Tesla needs to have active scanning in full circle, 360°, whether that’s lidar or radar.

I was very surprised indeed that Tesla did manage to pull off the “see two cars ahead with radar” thing, altho it remains to be seen how reliable that is. Anyway, it may be that Tesla can successfully use radar instead of lidar. But they can’t ignore the need for active scanning in all directions, not just to the front or just to the front and rear.

I keep comparing the Tesla Sensor Suites to humans– who only have two eyes and ears.

Seems the Tesla Sensor Solution has capabilities that seem like psychic powers, if a human could perform them. I certainly can’t look under and around vehicles that are not in my direct line of sight; in rain, dust, or snow… And I certainly don’t have 8 eyes like a spider, looking around in a 360 degree view of the world. I don’t have a built in GPS and I can’t pull external data into my head, from any other member of the fleet whenever I need updated position info.

Their Neural Net is really the last piece of the autonomous puzzle– and that seems to be going quite well. The cars are far better equipped now, than any human ever was, for making safe driving judgements. Just takes time to refine, test and train.

“The cars are far better equipped now, than any human ever was, for making safe driving judgements”

Regarding the sensor suite you are completely right.

Regarding the signal processing the cars still have quite a long way to go until they approach human capabilities (which vary widely 😉 )

Give it one more year… Then we are outdated.

Exactly. The autonomous systems win by far on sensors. But the humans still have a big advantage in recognizing, understanding, and navigating the objects in a 3D world.

It is very hard for computers to compete with a huge neural network that has received 16+ years of training of 3D navigation. They WILL eventually get there and be better than us, but right now they are not there.

IMHO, the comparison between human drivers and computer autonomous driving is this: Humans: Have an OK set of sensors (eyes, ears, touch, etc.) But have neural network that has been honed over millions of years to be excellent at navigating a 3D world. Each one of those neural networks is given a minimum of 16 years of 3D world training to learn how to navigate in the 3D world before it is allowed to drive. Then further training is provided. Autonomous driving: A great set of sensors that can sense with techniques unavailable to human like radar, lidar, infrared, 360 degree vision, GPS positioning, etc. Memory that can supply a full map of the area to be navigated. Processing ability that is much faster than humans in mathematical computations. But the software just is not as good as the human’s 16+ years of neural network training for understanding and navigating a 3D world. So to do the autonomous driving, the system relies upon better sensing data provided to its less capable software. The autonomous system will be able detect things that humans cannot and often detect & react faster. However, there is no fundamental strong 3D detecting, understanding, and navigation… Read more »

Cameras do actively scan and at much higher frequency than radar or lidar. And stereo cameras are capable of rendering depth/shape of objects. Just takes good software and lots of computing power.

Radar is really just needed for redundancy in the most critical direction.


Radar or lidar is also needed because the car is not allowed to actively shine light in random directions. If it wants to see a car behind it better it can illuminate it with UV, infrared or radar, but it can’t shine a white light on it, that’s not allowed by law.

I am skeptical that they will be able to reach full autonomy with only cameras. I think it will require active illumination, either lidar or radar. But then again I’m not capable of making an autonomous driving system so it wouldn’t be at all odd if I were proven to be wrong.

Does HW2 include a faster central display?