Separate names with a comma.
Discussion in 'Model S' started by PluggedINLife, May 21, 2017.
I'm sorry that video is so crappy. I was focusing on driving and not movie making
and I guess when I say it "irks me" it's that the users who have the homicidal cars make it sound like it's just assumed that it's a ubiquitous problem. In that everyone who has AP2 is experiencing these symptoms, and that AP2 is a complete failure from the first car made with it to the last car that just rolled off the assembly line. I don't think that's the case. I think there are some that have some calibration issue or something that is making them act erratically.
I drive to Des Plaines to visit my folks and take NW Hwy/Miner and onto further NW areas sometimes (there is a great bike shop in Mt Prospect). I actually drove this (in both directions) stretch this weekend and my car was fine. Try some roads elsewhere. Try Sheridan in the north shore. See if you live.
Its not necessarily hardware (or maybe it is but that isn't the entirety of the difference). I think the software isn't robust enough for most common situations. Sheridan rolls along the lake and has hills, windy stretches, and is only 35mph but my car's AS will repeatedly (casually!) make lazy curve turns and send me (at the tail end) into oncoming traffic. Its night and day from earlier FW but 17.17.x is still unsafe. If the sun is being filtered by trees, the vision becomes confused and my car will wobble like its drunk. If the road is wider than normal (like Green Bay Road), the car constantly hunts and looks drunk or hugs to the double yellow and close to oncoming traffic (making everyone else nervous too). I can list a bunch of local roads where I've had issues and maybe we can determine whether its hardware or just road choices. I find it works great on highway like local roads (Lake Cook, Mannheim, Half Day, etc.) but the less standard the road, the less successful AS is at keeping within the lines. There are issues and just because it works on some roads, doesn't mean it works well enough.
EDIT: I have 7800+ miles on my car so it shouldn't be calibration for the cameras since they automatically calibrate.
Very good. It works most of the time; not all of the time.
This "banter" on what qualifies as an AP failure is probably boring (non-hyperloop reference) those who were more interested in discussing the new update coming next month.
Hopfully the new smooth-as-silk update allows everything on the car to work better than before.
It's still a great car despite expectations having to be constantly readjusted.
The anticipation of what is to come is sometimes the best part of owning a Tesla.
I'll take that. It doesn't work in all circumstances, and I do have to take over. Again, my point is that I'm just annoyed at the posters who claim that since it only works most of the time then that means it fails all the time. I get that it is largely hyperbole, but after so many "AP sucks, it's useless, it tries to kill me" posts it just gets on my nerves. And I will just have to mature as a person and get past that, I get it.
I'd really like to get past this and move on to the discussion of what is meant by "next month." Does he mean calendar month of June, or in the next 31 days? Does it count if it rolls out to 1,000 cars on June 30? Is there going to be an update before that? It's a parsing nightmare.
It doesn't matter either way. It's obviously not yet ready for release, so release date is unknown. That combined with Mr Musk's track record says you should simply wait until there's actual evidence of arrival before you get excited. Or at the very least always take the upper end of the spread.
Thanks for your video, as with my AP2 I can see your car autosteer to the outwards of the lane but not crossing it. You did come terribly close the that guard rail though. And on the left hand turn, a driver in the adjacent lane would probably rather cut the inside of the bend, making it a close call between you. Not comfortable.
My theory is that the system yet don't compute the correct lane borders. An example is seen on the the "medium range camera" in Teslas original video. Hopefully, the next update does this correctly - and also is "silky smooth".
Alil bit of both.
December 2015: "We're going to end up with complete autonomy, and I think we will have complete autonomy in approximately two years." -
Elon Musk Says Tesla Vehicles Will Drive Themselves in Two Years
January 2016: "In ~2 years, summon should work anywhere connected by land & not blocked by borders, eg you're in LA and the car is in NY" -
Elon Musk on Twitter
June 2016: ""I really consider autonomous driving a solved problem," he said. "I think we are probably less than two years away." -
Two years until self-driving cars are on the road – is Elon Musk right?
The under-deliver is always on the timeline (with APv2 owners still holding the bag for APv1 features). And in at least 1 instance, the summon on private property blurb leading AP1 owners to believe that feature would happen on that hardware platform. So yes, I would say EM suffers from the over-promise, under-deliver problem. Does he do it intentionally? Without a doubt, have to keep that stock stabilized. Does he do it out of malice? Doubtful, I think he really believes the timelines will be met or at least that the timeline will motivate his employees to deliver on the unrealistic timeline. Does he straight up lie? You decide. He knows all the pieces to the puzzle and still makes statements that aren't fully truthful. Remember how non-AP cars couldn't be retrofitted to APv1? Well, they can, for about 10k (wholesale parts plus labor).
Well, if past is prologue, you can probably get excited about the functionality, just not necessarily the exact timeline. I think the most exact statement you could say from the timeline Elon gave is that it'll probably come sometime between June 1 and, oh, October 31.
but u did say FSD better than human will be available in 2018
I didn't say what time I personally believe when it will be available, just my interpretation of what Elon said he believed.
And ready technically and "available" is different. You can have all the tech ready and if the legal framework is not there to allow consumers to use it, then all that remains is a nice demo, and not actually "available".
In fact, looking through my comments, I specifically said to you directly that I found 2018 unlikely for regulatory reasons. Maybe Tesla can do it in one specific jurisdiction, but that certainly isn't summon "anywhere" (even if "anywhere" is limited only to the US).
2 years: Musk's new timeline for door-to-door full autonomy
Again, not sure why you keep calling me out for no reason.
Hate to say it but it kinda depends where you live and where you drive regularly. We've been on road trips to Florida and Atlanta where AP2 is 99% perfect. That was even the case driving back through North, South Carolina because of a traffic detour. North has unbelievably sharp S curves just as you describe and is literally out in the middle of nowhere. AP2 handled every sharp curve perfectly. But then we come home to Charleston and our local interstate, and AP2 "tries to kill me every time." If these reports cause you heartburn, the answer is simple. Don't read them. If they save some else's life, then I'm a happy camper.
That is a lot of 2 year promises. So the car is also going to pull up to a supercharger and charge itself? Can someone explain how that is going to happen?
If your girlfriend took $8,000 and gave you a receipt for loving you forever 40 years ago, I think you gave a good case for a very nice refund with 40 years of interest. At a reasonable 12% APR that would be a cool $744K.
I'll start getting excited when my HW2 car can do everything the HW1 car I test-drove could do. Until then each restored feature is just a consolation prize for having been conned.
What Tesla needs, I think, is a strong COO to actually run the company. Mr. Musk is the consummate genius inventor, but his enthusiasm for what's next gets in the way of making realistic commitments and then working diligently to meet them. Tesla is what Apple would have been with one Steve instead of two.
As the CEO, Musk should ultimately take responsibility for software/functionality deliveries that are late or missing functionality - which has been a pattern with Tesla the last few years.
However, the change Tesla probably needs is a stronger executive responsible for the software development - who would be held accountable for software releases, someone who would produce more realistic schedules and put the people and processes in place to ensure the software meets quality and functionality goals.
In any other company - it's unlikely the executive responsible for software development would have survived very long with the track record we've seen for Tesla's software releases...
I could have sworn you said that it will be ready in 2018 regardless of regulation.
You keep mentioning regulation but elon have sworn multiple times and have shown from prior practice that he will release fsd as a driver assistance. So why do you keep mentioning regulatory? Regulation has nothing to do with it. The performance of tesla fsd will be judged by how many disengagement/intervention per mile. That has nothing to do with regulation.
In 2018 when your tesla on the beta fsd software can go months without you intervening then you know that fsd is nigh.
But tesla will release fsd as driver assistabce so Please stop mentioning regulation.
Please quote that 2018 comment you made.
I just want to pin u down for early 2018 as you said. So when that comes and passes, i can hold you to it.
I just figured it out: "next month" means N+1. Now..... don't ask me what "N" means
Honestly the biggest feature I'm waiting for is Autopilot to Nav integration - that is by far the coolest and honestly not that hard to do (i think so at least)