Would you trust AP2.0 with your life?

[DJ 240V]

I'm looking for input from experienced AP1 owners . Autopilot sure does lull you into a false sense of security. Being a techie I know software will have bugs and hardware will fail. Its a matter of "when" not "if".

I've just placed an order and excited about getting the new hardware. However I'll be paranoid to use it or rely on it! (Its kind of like how doctors are the worst patients, I'm a programmer ). During the test drive the AP just increased my stress levels , I'm a pretty bad backseat driver anyway!

A few years back, in an interview with NYT Elon must said he would never put his kids in a self driving car and said they tec probably needs another 10 years to mature. We all now AP2.0 and the self driving feature will convince us 99% of the time that its safe and increasingly have you less focused on being the ultimate fail safe machine ( the driver). Its the 1% use cases that could be fatal. So the question is what tips / thoughts do you have to use AP2.0 prudently?

Personally, I think it will just be a toy to show my friends and family, until we have enough data and experience with the system.

 

[Jtrader]

I just hope the car pulling up behind me at the traffic light with the texting teen has it.

 

[taurusking]

I just hope the car pulling up behind me at the traffic light with the texting teen has it.

Not at this time...May be in the future when all the cars on the road ( different manufacturers) have self driving capability

 

[Drewflux]

The phrase trust but verify comes on mind on this subject. AP2 will catch most things but at the end of the day its still a driving aid.

 

[Camera-Cruiser]

Hmmmm. AP2.0 vs unknown taxi/uber driver with my kids.

The decisions we will have. I'll go with AP.

 

[DJ 240V]

Hmmmm. AP2.0 vs unknown taxi/uber driver with my kids.

The decisions we will have. I'll go with AP.

Kinda sad that humanity has come to this...

 

[Petra]

The phrase trust but verify comes on mind on this subject. AP2 will catch most things but at the end of the day its still a driving aid.

Trust but verify is a good way to look at Autopilot 1.0 and will be exactly the right approach for AP2.0 and all future versions until true level 5 autonomy is actually reached (which, by definition, doesn't require operator oversight or intervention).

Personally, I don't really like the word trust when it comes to current levels of driver assistance but that's an entirely different discussion.

 

[ecarfan]

Kinda sad that humanity has come to this...

Why is it "sad"?

 

[DJ 240V]

Why is it "sad"?

I understand its practical and safe to trust a machine more than another human being ..but then..we are not in the matrix...yet ( although Musk is sure we are)

 

[Camera-Cruiser]

There is no spoon. Only the matrix.

 

[essmd]

It may not be a question of the owner trusting AP 2.0 and a fully autonomous vehicle, but what had not yet been discussed is whether or not the municipalities will trust it.

No doubt Tesla can produce the tech, especially in a "vacant" world, but in reality there are other vehicles, pedestrians, tree-hidden stop signs, human drivers, etc.

I forsee cities, counties, and even states potentially regulating autonomous vehicles on the road. It is unlikely it will be a free reign for unattended vehicles on every public and private road and parking lot.

Dreams are wonderful...until reality kicks in....

 

[S4WRXTTCS]

I said I wouldn't trust it with MY life.

But, my main use for it will be to drive a robot to the grocery store so the robot can get my groceries, and then drive the robot back home.

I don't want the robot to have to walk all the way to the grocery store.

Of course I'm joking because I can't possibly take a self-driving car seriously at this time. I do trust my life to a machine all the time (we all do without really realizing the extent we do), but it's going to be at least a 2-3 years before I'll let go of the steering wheel from on-ramp to off-ramp. I'm pretty sure I'll trust AP 2.0 in 2019 for level 3 driving. But, level 5? NO!!!

Plus notice how Tesla says 2X better than the average driver. Have you seen the average Seattle driver? *shudders*

 

[essmd]

I will say that AP 1.0 makes me a safer driver as it provides incredible assistance to the task with additional sensors to provide input as well as control of steering, acceleration, braking, etc.

 

[davidc18]

I was expecting radar at the rear corners. It will be interesting to see if they can pull this off with the announced sensor suite alone.

 

[WhiteCap]

If Elon/Tesla demonstrates a successful, LA to Times Square, fully-autmated run, as intended next year, that will go a long way towards building my confidence in the system. If they hybridize the route to include significant non-interstate segments that go through both smaller towns and larger urban cities along the route - all the better. After such a proof-of-concept demonstration, I'd feel much more comfortable trusting automation - assuming it's a fair demonstration vs a media stunt. (I'm assuming no cheats: such as putting the car inside a caravan of manually driven cars, so all it would really be doing was chasing a lead-car that was running clearance.)

 

[Chopr147]

Right now not enough info.
We need to see 2.0 in action after it's learning phase. Maybe 2018, but that's even too soon.
I love AP 1.0 and it does a great job in stop/go and highway driving but the driver always has to be engaged.
It will be interesting to watch the next few years with Tesla and all manufacturers. The big auto-makers have showed their EV plans. Now Tesla may force their hand on the AP front

 

[jbcarioca]

...
A few years back, in an interview with NYT Elon must said he would never put his kids in a self driving car and said they tec probably needs another 10 years to mature. We all now AP2.0 and the self driving feature will convince us 99% of the time that its safe and increasingly have you less focused on being the ultimate fail safe machine ( the driver). Its the 1% use cases that could be fatal. So the question is what tips / thoughts do you have to use AP2.0 prudently?

Personally, I think it will just be a toy to show my friends and family, until we have enough data and experience with the system.

99% will not even quite equal the present level two Tesla capability. That is a pitifully low standard. We are looking for something on the order of 99.999% or better for level Five. Remember taht there will be gradual release of functions and feature as data develops and technical refinements continue. Even the present Level 2 capability is very useful, but all users MUST understand and carefully monitor operation. That is the issue. Further, we are in dire need of serious driver education now and in the future. Our Tesla vehicles are amazing already, but every user needs to know and understand soem about teh technologies involved and their limitations. The days of jump in drive and forget are gone, if indeed they ever were.

 

[Tam]

To trust it, I need to know how it works and how different it is from the classic Autopilot.

I assume, this time, the camera is full color not like the AP1 monochrome camera that couldn't figure out there's a different color between a white tractor-trailer and the brightly lit sky background.

There's need for a test of the Florida case whether it can stop hitting tall/high obstacle while traveling at 74 mph.

There's need for a test of the China case to see whether it would avoid hitting a slow street sweeper that's hugging all the way to the left side of the road.

We need a test to see it would still crawl under flat bed trailer during an unsupervised summon.

We need a test to see it still whack a tall/high pickup truck while doing a parallel parking.... and so on.

Ok! So who's among us going to do those tests?

 

[Clomer]

I'll trust it when Tesla is willing to accept full liability for any accidents it causes while it is enabled. It's not truly L5 until this is the case.

 

[Tam]

I'll trust it when Tesla is willing to accept full liability for any accidents it causes while it is enabled. It's not truly L5 until this is the case.

Looks like Tesla will not accept liability blindly. Elon said that you work with your insurance first. He cited an example about everyone has an insurance policy on their elevators in case some thing happen with them. But sure Tesla would take responsibility if it's Tesla's fault.

There goes how are you going to prove whose fault it is, regardless whether you applied the brake or not:

"Arianna Simpson, a venture capitalist in San Francisco, said the Autopilot in her Model S “did absolutely nothing” when the car she was following on Interstate 5 near Los Angeles changed lanes, revealing another car parked on the highway.

Her Model S rammed into that car, she said, damaging both vehicles but causing no major injuries.

Tesla responded that the April crash was her fault because she hit the brakes right before the collision, disengaging Autopilot. Before that, the car sounded a collision warning as it should have, the car’s data show.

“So if you don’t brake, it’s your fault because you weren’t paying attention,” said Ms. Simpson, 25. “And if you do brake, it’s your fault because you were driving.”

She doesn’t expect to use Autopilot much once her Model S is repaired, partly because she thinks she would constantly second-guess the automated-driving system."