Anthony Levandowski beats Tesla's Elon Musk to first Auto Cross country (3k Miles, 0 disengagement)

[Bladerskb]

Here is elon musk prediction in comparison:

Elon Musk Confirms Cross-Country Autonomous Tesla Road Trips | Inverse

Looks like Delphi/Aptiv did 99% in 2015 and now Anthony Levandowski with his new company Pronto.AI has done 100% in 2018 using just cameras and regular navigation maps (no pre-mapping / hd map).

Wow. This seems to be a major blow to Tesla and Elon. It sorta shows you how easy it is to do a cross country. Then why is Tesla still struggling? This drive by Anthony shows you where the state of the art is by one guy with a very small team with no funding. Alot of people look at Tesla's release of AP and think Tesla is 3, 5, 10 years ahead because of NOA. But doesn't this show that the state of the art is well beyond what AP is?

Doesn't NOA reveal how behind Tesla is?

Completing the transcontinental voyage also took multiple attempts. The first try, in late September, ended on the Bonneville Salt Flats in Utah, when the system disengaged on a banked curve in high winds. On its second go, two weeks later, Levandowski says the Co-Pilot worked perfectly for 650 miles, again as far as Utah. But it was too perfect for one Nevada Highway Patrol officer, who pulled the Prius over after noticing it driving slightly below the speed limit in an area where most drivers were speeding.

“The team tried to tell me that it wasn’t a disengagement, but I said, I can’t touch the steering wheel, brake or gas otherwise everybody’s going to look for the gotcha. So we came back to San Francisco,” recalls Levandowski.

Pronto engineers adjusted the software so that the car would be allowed to travel faster on certain roads, and tried again. On his third trip, Levandowski said that he encountered rain in Nebraska and Illinois, high winds in Wyoming, and a rolled-over semi in Pennsylvania, but eventually made it to the George Washington Bridge without a disengagement.

“If true, a truck that used only cameras to steer, brake, and accelerate for 100% of any cross-country trip is impressive,” said Bryant Walker Smith, a law professor at the University of South Carolina and member of the US Department of Transportation’s Advisory Committee on Automation in Transportation. “Making a system work with cameras alone could be a major contribution, especially if this could be applied to higher levels of driving automation.”

What do you guys think? Has Tesla declined? When do you expect Tesla to do its own cross country drive to counter?

Controversial engineer: I travelled over 3,000 miles in a self-driving car

 

[croman]

Lol. Where can I buy? Oh wait.

Also dude is a thief and scumbag. At least Elon intends well.

 

[Bladerskb]

Lol. Where can I buy? Oh wait.

Also dude is a thief and scumbag. At least Elon intends well.

Idk. Some of the most accomplished people in the world are considered a thief and a scumbag

 

[BerTX]

Tesla can't afford multiple attempts and failures. The goal is to demonstrate that they can shoot and score every time, not that if they keep flinging the ball at the hoop it will eventually go in.

 

[Bladerskb]

Tesla can't afford multiple attempts and failures. The goal is to demonstrate that they can shoot and score every time, not that if they keep flinging the ball at the hoop it will eventually go in.

Shooting 3,099 times and it going in isn't just flinging at the hoop and hoping it goes in. The fact is, Tesla hasn't done it sucessfully once. In fact reports have come out that they tried multiple attempts and failed. Tesla AP today can barely go 100 without ANY disengagement.

 

[diplomat33]

Looks like Delphi/Aptiv did 99% in 2015 and now Anthony Levandowski with his new company Pronto.AI has done 100% in 2018 using just cameras and regular navigation maps (no pre-mapping / hd map).

So it proves that Tesla's camera only approach is the right approach and LIDAR is not needed? That's great validation for Tesla!!

Wow. This seems to be a major blow to Tesla and Elon. It sorta shows you how easy it is to do a cross country. Then why is Tesla still struggling? This drive by Anthony shows you where the state of the art is by one guy with a very small team with no funding. Alot of people look at Tesla's release of AP and think Tesla is 3, 5, 10 years ahead because of NOA. But doesn't this show that the state of the art is well beyond what AP is?

Doesn't NOA reveal how behind Tesla is?

Watching the video, it seemed mostly highway driving that NOA can already do. Tesla might be behind a little but this video is certainly not proof that they are "far behind".

Tesla AP today can barely go 100 without ANY disengagement.

I think that is an exaggeration. I've driven more than 100 miles repeatedly on highways with no disengagements from AP.

 

[Ande]

3k... k=kilo, so 3000... Miles ? Enough with that primitive mess already, I know USA have some lwarlear deficiency when it comes to metric system, but that mess above don't mix.

 

[Ande]

And... It proves nothing new just cruising on highway in good conditions, manual driving off, onto the highway.
Meh.
Also: until the system is in production, and available, it simply does not count, driver could influence needed lane changes and do corrections by a joystick or gestures.

 

[diplomat33]

Question: the video shows the car made a few pit stops for gas and for sleeping but the video is so fast, it is hard to see what happened during those pit stops. Did the car actually self-drive to the gas station and away from the gas station and to a hotel, self-park and then self-drive or did the driver disengage to do the pit stops and they are not counting those as real disengagements? I ask because if they did have to disengage for the pit stops then that is no different than AP now where AP can handle long stretches of highway but the driver disengages to pull up to a supercharger for example.

 

[Russell]

... The fact is, Tesla hasn't done it sucessfully once. In fact reports have come out that they tried multiple attempts and failed. Tesla AP today can barely go 100 without ANY disengagement.

Level 5 Full Self Driving is vastly different than a Level 2 Drive Assist.

But if you think Tesla's Level 2 AP is that good then that's a compliment for Tesla!


On many freeways I can drive my AP1 Tesla well over 100 miles without disengagement.


I see you are from Michigan. Which Tesla competitor do you work for?

 

[acoste]

Level 5 Full Self Driving is vastly different than a Level 2 Drive Assist.

But if you think Tesla's Level 2 AP is that good then that's a compliment for Tesla!


On many freeways I can drive my AP1 Tesla well over 100 miles without disengagement.

AP1 is MobilEye, right?

 

[DBeckwith]

So it proves that Tesla's camera only approach is the right approach and LIDAR is not needed?

That was my thought since Levandowski was selling Velodyne's 360 degree LIDAR back in the day. But I think he was moving towards passive vision also.

 

[Russell]

AP1 is MobilEye, right?

Right, hardware. Tesla constantly updates the software to make it perform better than when it was first released in late 2015.

 

[Tam]

...why is Tesla still struggling?...

That's because a Tesla car is supposed to pick up passengers in Los Angeles for the trip and goes to Times Square and drops them off then finds a parking space on its own to park.

 

[SeminoleFSU]

Here is elon musk prediction in comparison:

Elon Musk Confirms Cross-Country Autonomous Tesla Road Trips | Inverse

Looks like Delphi/Aptiv did 99% in 2015 and now Anthony Levandowski with his new company Pronto.AI has done 100% in 2018 using just cameras and regular navigation maps (no pre-mapping / hd map).

Wow. This seems to be a major blow to Tesla and Elon. It sorta shows you how easy it is to do a cross country. Then why is Tesla still struggling? This drive by Anthony shows you where the state of the art is by one guy with a very small team with no funding. Alot of people look at Tesla's release of AP and think Tesla is 3, 5, 10 years ahead because of NOA. But doesn't this show that the state of the art is well beyond what AP is?

Doesn't NOA reveal how behind Tesla is?



What do you guys think? Has Tesla declined? When do you expect Tesla to do its own cross country drive to counter?

Controversial engineer: I travelled over 3,000 miles in a self-driving car

Wow if this is legit it's very impressive. Also very interesting he was able to do this with just cameras an no radar (even in the front)

 

[BulldogsRus]

Here is elon musk prediction in comparison:

Elon Musk Confirms Cross-Country Autonomous Tesla Road Trips | Inverse

Looks like Delphi/Aptiv did 99% in 2015 and now Anthony Levandowski with his new company Pronto.AI has done 100% in 2018 using just cameras and regular navigation maps (no pre-mapping / hd map).

Wow. This seems to be a major blow to Tesla and Elon. It sorta shows you how easy it is to do a cross country. Then why is Tesla still struggling? This drive by Anthony shows you where the state of the art is by one guy with a very small team with no funding. Alot of people look at Tesla's release of AP and think Tesla is 3, 5, 10 years ahead because of NOA. But doesn't this show that the state of the art is well beyond what AP is?

Doesn't NOA reveal how behind Tesla is?



What do you guys think? Has Tesla declined? When do you expect Tesla to do its own cross country drive to counter?

Controversial engineer: I travelled over 3,000 miles in a self-driving car

Yeah...but does it have a fart app?

Exactly.

 

[ucmndd]

Question: the video shows the car made a few pit stops for gas and for sleeping but the video is so fast, it is hard to see what happened during those pit stops. Did the car actually self-drive to the gas station and away from the gas station and to a hotel, self-park and then self-drive or did the driver disengage to do the pit stops and they are not counting those as real disengagements? I ask because if they did have to disengage for the pit stops then that is no different than AP now where AP can handle long stretches of highway but the driver disengages to pull up to a supercharger for example.

Verge article indicates the fuel stops were manual “planned stops to rest and refuel” but doesn’t really go into detail. So yeah. Cool parlor trick but 3000 miles of lane keeping on a well marked interstate doesn’t seem like any sort of indication of revolutionary performance at this point.

 

[ucmndd]

I’m reasonably confident this exact test could be carried out today with Nav on Autopilot and manual “planned stops to rest and refuel”.

 

[SeminoleFSU]

Verge article indicates the fuel stops were manual “planned stops to rest and refuel” but doesn’t really go into detail. So yeah. Cool parlor trick but 3000 miles of lane keeping on a well marked interstate doesn’t seem like any sort of indication of revolutionary performance at this point.

It looked like it was flawlessly taking the connectors between interstates though too... which Nav on AP can do, but only if you "assist" it to get in the right lane. This dude looked like he was providing no input at all (so it was doing all the lane changes with no help).. If it was indeed doing all those lane changes with no human interaction, that is indeed way better than Tesla's current Nav on AP. I've found that Tesla's AP can change lanes decently well in light traffic, but once it gets heavy it quickly starts to fall short.. 0 disengagements in that many miles? I don't think Tesla would have made it across the 2nd state line without us having to intervene on its behalf (because of lane changing)... Just my $.02. Of course I've only driven a loaner for a week with Nav on AP. I'm currently on 48.1 but the maps didn't install and I haven't been able to get into the SvC to get them installed. Stupid this requires me to go in for it... c'mon Tesla just push me the maps on my damn WiFi. If they've frozen on install, restart the process!! jesh.

 

[ucmndd]

It looked like it was flawlessly taking the connectors between interstates though too... which Nav on AP can do, but only if you "assist" it to get in the right lane. This dude looked like he was providing no input at all (so it was doing all the lane changes with no help).. If it was indeed doing all those lane changes with no human interaction, that is indeed way better than Tesla's current Nav on AP. I've found that Tesla's AP can change lanes decently well in light traffic, but once it gets heavy it quickly starts to fall short.. 0 disengagements in that many miles? I don't think Tesla would have made it across the 2nd state line without us having to intervene on its behalf (because of lane changing)... Just my $.02. Of course I've only driven a loaner for a week with Nav on AP. I'm currently on 48.1 but the maps didn't install and I haven't been able to get into the SvC to get them installed. Stupid this requires me to go in for it... c'mon Tesla just push me the maps on my damn WiFi. If they've frozen on install, restart the process!! jesh.

Fair point about needing to confirm lane changes, though it sounds like that might be a short term issue (as with everything Tesla and AP, we’ll see).

Most of my highway to highway interchange tests so far have been successful. Sort of my point though - once you’re on interstate 80 pointed east, there aren’t a lot of interchanges to navigate. Just flat, straight roadway for as far as the eye can see. Toss out the complexity of fuel and rest stops, and that’s basically best case for autonomous driving.