Received AutoPilot trial this morning

[Nocturnal]

I love it for daily commute. Like a second driver who doesn’t audibly gasp at every car. It takes a little practice to understand its limitations and have confidence in it. I like using it now and adding features as I go rather than waiting a year and then trying to take in whatever system is in place, if that makes any sense.

After losing it today I tried using regular cruise control. I had to make myself grab the wheel and steer. lol.

I do miss it, just hard to justify another 5k right now.

 

[ranss12]

I just purchased EAP after the 2 week trial. I like it on the interstates, but not so much on 2 lane roads. What I really don't like are the 'false positives' where it brakes suddenly (usually for only about 1/2 to 1/3 of a second) for no apparent reason. I also don't like when it brakes because a car at an intersection crosses the road 100 yards ahead. But by purchasing it, I get a front row seat to watch it while it improves over time, and I get to use it where it is really valuable, on the interstates.

 

[KarlaM3]

So, the trial opened up for me as well and I tried it in commute traffic on an expressway, which is supposed to be a scenario where EAP shines. I have to say this was more scary than relaxing. There were several situations where my lane was clear but I could see stopped traffic maybe 150 ft ahead, and the autopilot just continued driving at full speed up to a point where I had to brake pretty hard to avoid crashing into the stopped vehicles. In all cases I took control because what it was doing seemed way too risky. It's as if the car can only see a very short distance ahead even if the lane is clear. There was also a situation where the car in front of me that the autopilot was following changed to an exit lane and my car for some reason followed, even though there were clear markings on both sides of the lane I was in.

I'll try some highway driving on a trip over the weekend, but so far my impressions aren't good.

I just finished my trial period and I am glad I didn't buy this option. I was driving 70 down a highway on a very sunny day and the shadow from an over pass caused the car to slam on it's brakes. Glad no one was following me closely or it could have been a problem. I responded quickly and got back up to speed again. That was enough to make me not want to use auto steer.

 

[SOULPEDL]

There was also a situation where the car in front of me that the autopilot was following changed to an exit lane and my car for some reason followed, even though there were clear markings on both sides of the lane I was in.

I don't see how it is possible that your M3 changed lanes to follow another car - they don't follow cars and I've never heard of this situation... ever on this forum. I suspect you used your turn-signal without knowledge.

Most feedback (I've read all of them) is that people are uncomfortable with the pace of stopping. First, the car sees way further than you think given it's radar, so don't think it doesn't see the changes far ahead. One fellow (on another thread) wanted it to drive more fuel conscious by coasting more and trying to achieve a constant speed (his goal as to reduce the jam). I leave that to future improvements; safety then usability is priority for now... cars will eventually talk to one another but not for a long time.

Then there's the variable of traffic behaviors in your city (and there's a difference) and how people respond. I've heard late braking as a concern, but to the next person it's also too timid, enough so that others squeeze in. Case and point: During a Stockholder meeting a few months back, a young lady walked up to the mic and asked Elon if there was a way to make the car tone down the Emergency Alerts when her mom drives (paraphrasing here). Obviously mom's a tailgater and brakes last minute, it's just her style. My opinion is that eventually, it will learn your preference (in your area) and adjust accordingly.

What many don't realize is that this technology is evolving over time. Myself, I initially drove in AP as though it could do anything on it's own - extremely cautious. In fact, on freeway exchanges it could hit a bump that caused it to disengage (while I was trying to also steer the car in the turn). There was a reason for all that as explained on my other posts. Today the car slows to about 65 mph automatically on these same turns and I have learned to trust the technology. That's an improvement that occured right before my eyes.

As the saying goes, the most experienced pilots crash planes. If you approach this technology as if it's experienced, you are taking risks and those are likely the people not paying attention while in AP. The same will likely occur on FSD.

What I find is that I cannot possibly drive straighter in the lane than this car, so we're a great team. (There is an exception when the lane lines widen and it drifts for a moment always trying to center itself). So I supervise, but it does the driving and I know when I can trust it 100% and when I should keep an eye out. This will change over time to where I cannot possibly see in all directions simultaneously and respond accordingly, and so I'd be a fool to think I could do better.

One more thing. To the occasional braking on, say a bridge shadow or car in the next lane, for seemingly no reason, it's a delicate balance between overcautious and unsafe. The Uber car in Tempe saw the pedestrian yet determined that it was reasonably safe to ignore. As a shareholder and M3 owner, I expect they err on the side of safety. The car is learning along with a fleet of cars in your area (as Tesla explained to me), very location specific learning. Once a few cars pass under that bridge successfully, that hesitation will eventually disappear and that's the learning process. This goes right down to the reflective color/temperature of the line paint on the road where DOT mixed a color out of spec and the cars needed to make exception for their error. These adaptations are happening world-wide, non-stop, so please be patient, enjoy your safer ride, and be amazed with every software update!

 

[Eno Deb]

I don't see how it is possible that your M3 changed lanes to follow another car - they don't follow cars and I've never heard of this situation... ever on this forum. I suspect you used your turn-signal without knowledge.

Uhm, no.

Most feedback (I've read all of them) is that people are uncomfortable with the pace of stopping. First, the car sees way further than you think given it's radar, so don't think it doesn't see the changes far ahead.

It likely doesn't use the radar to detect stationary objects, since radar is not precise enough to be able to tell relevant objects (such as stopped traffic ahead) from things like road signs or overpasses.

What many don't realize is that this technology is evolving over time.

Oh, I think most people are acutely aware of that. The thing is that I don't necessarily want to trust my (and others') life to "evolving" beta software.

 

[theschnell]

I’ve had my trial for like 3 days now. I enjoyed it on 2 lane highways, but today was my first day of interstate driving, and it was super cool! Lane changes very smooth. Spacing in traffic very good. I also really like how the bags are visual and I found them very easy to clear. I went through several situations where I expected it to fail and have me take over, but it worked fine.

If I had an interstate commute I would be extremely tempted. But I don’t and so cannot justify $6k. But I am really enjoying the trial to be able to experience autopilot and it has exceeded my expectations, especially on the interstate.

 

[Matt L]

I’ve had my trial for like 3 days now. I enjoyed it on 2 lane highways, but today was my first day of interstate driving, and it was super cool! Lane changes very smooth. Spacing in traffic very good. I also really like how the bags are visual and I found them very easy to clear. I went through several situations where I expected it to fail and have me take over, but it worked fine.

If I had an interstate commute I would be extremely tempted. But I don’t and so cannot justify $6k. But I am really enjoying the trial to be able to experience autopilot and it has exceeded my expectations, especially on the interstate.

What about $5k?

 

[theschnell]

What about $5k?

No, even at $500 I would have to think about it because I just wouldn’t use it very much.

Again if I had an interstate commute I would be tempted at $5k or even $6k, but I don’t so i’m just not interested right now

 

[SOULPEDL]

The thing is that I don't necessarily want to trust my (and others') life to "evolving" beta software.

I respect your decision and I can see how quirky things we don't understand can be unnerving. But you also realize you're driving on 4 rubber balloons at speeds that would have people's heads spin in the 30's. I had to add seat belts to my '63 buick because people didn't think they were necessary... for decades, while it was an OK family thing to go for a drive that way at freeway speeds on 2 lanes. I think the car is safer than you think - outstanding NTSB safety tests last week with 5 stars. And I'll make the case it's safer, EAP + You > You.

On your unexpected lane change "following the car in front" concern, it could be that the lane lines changed on the road and it "appeared" to follow the car in front of you when in reality the car in front had nothing to do with it. I've seen my M3 move over when a lane line ends like when a freeway on/off ramp merges. It will find the center between the lines which could have made you think it was following. There is no reason to follow a car, that algorithm does not exist. That's like "If Joe jumped off the bridge, would you jump off the bridge" logic.

I study hard to understand what this car does and how it thinks. I've repeatedly recommended Tesla educate owners on the basic automation concepts because knowledge always reduces fear. But if you expect that you can nap or work and drive, that would be unrealistic. It doesn't like tight turns, lines that drop off the horizon on L/R side, but I'll tell you...

I was driving back from California on I-10 eastbound Wed, and there was a rain storm I haven't seen in years. I could barely see, but the car still drove straight even in extreme crosswinds. It's times like this where the car has superior skills. That's why I take the team approach because when we both drive (it drives, I supervise) I honestly believe it's safer than not using it. I simply can't drive that straight no matter how straight I am. And it just takes one distraction for people to hit a bicycle or something. You realize that's why he's doing this right? It's to save a life, your life.

But I'd also go on the Moon fly-by with Yusaku Maezawa. He's not crazy, he realizes the risks and they will be informed ones. With the M3, I think it's a risk if you don't use EAP, unless you actually are perfect.