2017.42 a88c8d5

[rush6410]

With the new update and several hundred miles with the update, I can report that this is by far the most significant improvement since I have had my X (Dec 2016). One major improvement which I have been complaining about for a very long time is how AP would not deal with double solid lines. I can now confirm that the problem has been resolved. In all areas where AP failed completely with double lines now works. It picks the inner most line and follows. Seemed like an easy fix. All I can say is it is about time. I have two 90 degree S curves with intersections mid curves leading to my neighborhood. AP always had trouble traversing the conditions. AP 1 wouldn't even attempt. AP 2 started even with AP 1 and has slowly improved overtime. The the latest update, completely traversed the S curves without slowing down. I would consider using AP through this part of the road as a very good test of capabilities. The two large left and right S curves has two intersections mid cure where lines drop completely through the intersections. In addition one of the S curves increases elevation peaking mid intersection and dropped lines. So, the fact the latest update traversed where all other versions failed miserably tells me we are seeing progress. BTW, through that section of the road the speed limit is 45. So, I am traveling at a good clip considering the challenges. Just wanted to share my experience.

 

[Electroman]

So, the fact the latest update traversed where all other versions failed miserably tells me we are seeing progress

Thats good to know and thanks for sharing. If you can post a video that would be great

 

[AnxietyRanger]

It isn't hard to see why people might have a bit "optimistic" views about Autopilot, Tesla obviously wants them to - if that page (Autopilot) is any guide. Reality be damned. That page really is an eye-opener.

But then, many of us got that eye-opener fater October, 2016 when the AP2 we were then pitched has looked and sounded like nothing the 10-11 months of AP2 have later brought us..

Keep up the good work, AR. I know you had a lot of buddies who would chime in along with you, but looks like you are getting out-numbered now after the last updates. But don't give up on your whine-fest. You may have to dig harder but I am sure you will find plenty to complain to eternity.

In all seriousness, do you guys think Tesla's Autopilot page is NOT misleading for someone who doesn't follow the forums for the nitty-gritty?

I mean read it and say if you feel an average reader will get an accurate description in your opinion?

Autopilot

As for .42, the AP indeed is improving in my opinion: How is .40/.42 AP2 doing for you?

.42 just has not added anything that would make the Autopilot page - that mentions things like automatic highway transfers and smart summon - any more accurate?

Or do you disagree and if so where?

 

[chillaban]

In all seriousness, do you guys think Tesla's Autopilot page is NOT misleading for someone who doesn't follow the forums for the nitty-gritty?

I mean read it and say if you feel an average reader will get an accurate description in your opinion?

Autopilot

As for .42, the AP indeed is improving in my opinion: How is .40/.42 AP2 doing for you?

.42 just has not added anything that would make the Autopilot page - that mentions things like automatic highway transfers and smart summon - any more accurate?

Or do you disagree and if so where?

Personally every car that I’ve bought has a grandiose web portal that promised things that simply were not the case or were contradicted by fine print.... Tesla seems to sell on an eventual vision rather than what is available today. But it is the continuous improvement aspect that redeems it in my mind.

Every other $60-120k car I’ve bought in the past has come with its share of dirty laundry that only became apparent after ownership, despite plenty of test driving. The thing that sets Tesla apart is, like you’ve seen first hand, it’s literally the only car where overnight a software update makes something function better.

Of course, that’s my opinion as a general fan of the car and what the company is trying to do. I do agree with you though that the marketing from tesla is more “misleading” in general in terms of the kind of future they want to sell. Though I am willing to bet that at least 75% of the claims would come true.

A few years ago I laughed and ridiculed Tesla when they announced AP1. 1 camera, one radar, it was a joke compared to any of the 2013-era German cars they came with multiple radars, many of which had multiple cameras, yet they implemented fairly awful ACC and almost no semblance of lane keeping. Of course I ended up being wrong about AP1 though it took a year or two for that to become apparent.

I remain optimistic but not hopelessly so. But yeah, I do wish Tesla would tone down the marketing a bit. The way they blur the present and the future detracts from the raw magnitude of what they’re accomplishing to date.

 

[AnxietyRanger]

The thing that most people here forget, when responding myself, is that I was never critical of AP1. I think it was a masterful achievement.

Ironically, the only thing about AP1 that I am critical of are the few instances they overpromised (incidentally, I support the notion that Tesla only announce software features as they ship, not when the hardware ships - they could just ship the hardware and not promise the software until it ships). These are the cases where traffic-light detection, meet you on the curb/find you on the premises types of things were promised but never delivered... Why Tesla felt the need to keep saying such things I'll never know. They could have just shipped software as it matured and let it speak for itself.

I think the Autopilot 2 suite and even some of its software features remain impressive-enough to market as they ship today, not as a future vision. There is no need to say things like "automatically change lanes without requiring driver input, transition from one freeway to another", yet for some reason Tesla has insisted for a year now to keep those on the site, while what we've gotten is something completely different. The page is so vague it makes it sound like these are the features already shipping and even more features will come in eventual updates... Not surprisingly, TMC has had new members come in with unrealistic expectations set by this marketing.

Look, Tesla has an impressive product already today. There is no need to market the future to buyers. They could just market the today. But as long as they market a future in a potentially misleading manner, IMO calling that out is perfectly fair.

 

[DaddyP713]

I've been on 2017.42 for the past week or two with my AP2.0 MX. Generally, I'm quite impressed with it. Still stellar on the interstate, maybe a little better than .40, particularly around on ramps.

I experience a few phantom mild breaking events, but they seem to be related to traffic in adjacent lanes in curves on the interstate.

The most significant improvements for me are on surface roads. It negotiates intersections very well, with and without a lead car to follow. Poor lane markings and tar strips don't send it off in crazy directions any more.

My biggest remaining gripes are that it can't read speed limit signs (the GPS database is 10mph below reality in my area), and that it loses the roadway at the top and bottom of steep hills. There are 2 roads near my house where autopilot is unusable because of this.

Just got an update notice, likely .44. I don't expect any autopilot improvements, but the driver entry/exit feature will be nice.

 

[rush6410]

Thats good to know and thanks for sharing. If you can post a video that would be great

Here is a link to a video. Sorry for the quality. Again, AP 1 would fail immediately. AP 2.0 up to now would fail through the intersections and construction shifting lanes.

 

[Electroman]

Looks good. Given that the lanes are very light and not clear and in some sections lot of shadows from the trees, I think it did a pretty job. But if i were you, I would never go on AP on the left lane with opposing traffic, even if the conditions are perfect. Agreed that you have double yellow lines with a large gap in-between, but still you have a pretty good right lane with a shoulder

 

[rush6410]

Looks good. Given that the lanes are very light and not clear and in some sections lot of shadows from the trees, I think it did a pretty job. But if i were you, I would never go on AP on the left lane with opposing traffic, even if the conditions are perfect. Agreed that you have double yellow lines with a large gap in-between, but still you have a pretty good right lane with a shoulder

Trust me, I have my hands on the wheel through that part of my drive. Even though I allowed AP to steer through the video, out of site was my hands on the bottom of the wheel ready to take over. I have been an early adopter for AP 2.0. In fact I am pretty sure I had one of the first deliveries with AP 2.0. So, I am very cautious testing AP 2.0 in new or challenging conditions. I was an engineering program manager with HP on design teams. I understand technology and the risks of cutting edge development. I am very familiar with the sensors and feedback the X is providing knowing when to be cautious in all scenarios. The way I approach AP is I acknowledge and understand that AP is only as good as the sensors are detecting the conditions. A driver receives feedback from AP through the instrument display. I spend every second observing the display and the road understanding which conditions challenges the sensors regardless of how perfect the lines are from my perspective. What is important is how the sensors detects and deals with the lines. I have never seen the X cross a line when the car detects that line and displays it as blue. If the line on the display drops or turns grey or black it may cross the line. I think one of the biggest mistakes drivers make is they judge AP based on how they would drive instead of how AP drives based on the sensors. The driver freaks out because the X didn't react the way they would. Most of the time it is unnecessary anxiety. What I have done is learned how AP drives based on displayed indications. I have learned over time which road conditions will give the sensors problems which clues me in on when I can relax or be more cautious.

 

[Electroman]

Trust me, I have my hands on the wheel through that part of my drive. Even though I allowed AP to steer through the video, out of site was my hands on the bottom of the wheel ready to take over. I have been an early adopter for AP 2.0. In fact I am pretty sure I had one of the first deliveries with AP 2.0. So, I am very cautious testing AP 2.0 in new or challenging conditions. I was an engineering program manager with HP on design teams. I understand technology and the risks of cutting edge development. I am very familiar with the sensors and feedback the X is providing knowing when to be cautious in all scenarios. The way I approach AP is I acknowledge and understand that AP is only as good as the sensors are detecting the conditions. A driver receives feedback from AP through the instrument display. I spend every second observing the display and the road understanding which conditions challenges the sensors regardless of how perfect the lines are from my perspective. What is important is how the sensors detects and deals with the lines. I have never seen the X cross a line when the car detects that line and displays it as blue. If the line on the display drops or turns grey or black it may cross the line. I think one of the biggest mistakes drivers make is they judge AP based on how they would drive instead of how AP drives based on the sensors. The driver freaks out because the X didn't react the way they would. Most of the time it is unnecessary anxiety. What I have done is learned how AP drives based on displayed indications. I have learned over time which road conditions will give the sensors problems which clues me in on when I can relax or be more cautious.

Love it. Exactly how I treat AP

 

[Electroman]

In all seriousness, do you guys think Tesla's Autopilot page is NOT misleading for someone who doesn't follow the forums for the nitty-gritty?

...but but you are an old rat here who follow this forum nitty-gritty everyday. So why are you complaining incessantly? Even the newbies aren't complaining. In fact they are very happy.

 

[AnxietyRanger]

...but but you are an old rat here who follow this forum nitty-gritty everyday. So why are you complaining incessantly? Even the newbies aren't complaining. In fact they are very happy.

No doubt my AP2 complaints have been numerous about the quality of EAP until recently (.36, .40, .42, .44 have been better as I have reported) as well the IMO misleading Design Studio EAP description in late 2016.

The Autopilot page you commented on came up recently and I think I commented about it in exactly two posts (one in that thread and another one in this thread). Rest is then response to others...

I live in San Diego and there are two places miles away and the other is 535. I was thinking that with autopilot on these roads that are most highway, I could go to sleep for a few hours while traveling. One would require a stop. Is it possible to nap or completely zone out with autopilot?

Trolling? I don’t know what you mean by that. I read what is says on the website, but it didn’t make sense to me and when I went to the dealer at UTC, they were wish washy on it. Attention to me means hands on the wheel all the time.

Well, truth be told guys...

Please, go read the Tesla website on Autopilot. Just look at the images and read the text fully.

Autopilot

It says some really wild things. It does NOT even mention that Enhanced Autopilot should be considered a driver's aid only!

That is SO easy to misunderstand. Also, it does not make it at all clear that there actually is no Enhanced Autopilot yet, no Smart Summon yet, no On-ramp, Off-ramp yet... even though the page implies it has all these things. There are vague notes and disclaimers about future updates, but if I were just reading that page to learn about Autopilot (not unreasonable, right, given that it is Tesla's page on Autopilot?), I too would get a completely wrong picture of what AP2 is...

Such as:

Enhanced Autopilot

Enhanced Autopilot adds these new capabilities to the Tesla Autopilot driving experience. Your Tesla will match speed to traffic conditions, keep within a lane, automatically change lanes without requiring driver input, transition from one freeway to another, exit the freeway when your destination is near, self-park when near a parking spot and be summoned to and from your garage.

Tesla’s Enhanced Autopilot software has begun rolling out and features will continue to be introduced as validation is completed, subject to regulatory approval.

On-ramp to Off-ramp

Once on the freeway, your Tesla will determine which lane you need to be in and when. In addition to ensuring you reach your intended exit, Autopilot will watch for opportunities to move to a faster lane when you're caught behind slower traffic. When you reach your exit, your Tesla will depart the freeway, slow down and transition control back to you.

Autosteer+

With the new Tesla Vision cameras, sensors and computing power, your Tesla will navigate tighter, more complex roads.

Smart Summon

With Smart Summon, your car will navigate more complex environments and parking spaces, maneuvering around objects as necessary to come find you.

From Home

All you will need to do is get in and tell your car where to go. If you don’t say anything, your car will look at your calendar and take you there as the assumed destination. Your Tesla will figure out the optimal route, navigating urban streets, complex intersections and freeways.

To your Destination

When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you.

Basically the only thing with a vague disclaimer is a separate box about Full self-driving that has the Design Studio text in it.