Autopilot Brake checking

[john5520]

Other than where they've been saying they had to totally re-write the software to be capable of reaching their long-stated goal of city driving and have been doing so since last year and hope to be able to release it by end of this year?

It's frankly baffling you keep misunderstanding basic facts this badly.

There are multiple videos showing the pre-rewrite software driving on city streets, one of which you chose to ignore and claim as false. That's pretty baffling if you ask me.

I don't believe Tesla ever claimed it was done in 1 take- so they didn't technically lie. They just let folks like you assume.

Here's your statement:

That's flatly false.

Again they had to run that short route dozens and dozens of times to edit together enough short snippets to make it look like it did that all without touching the wheel.

They touched the wheel a ton in reality.

Tesla had 180 disengagements through only 550 total miles driven to be able to mock up that fake demo video.

That means the software failed and needed a human to take over once every ~3 miles on average. That's garbage

Stop citing fake news as a source of facts please.

You were claiming they deliberately edited the clip to make it look like it did all that without touching the wheel. That's false advertising. Maybe you should re-read what you write.

Dude.

Again- the current version of the software is listed on PAGE ONE OF THE MANUAL.
Where AP is intended to be used is listed RIGHT UNDER THE DESCRIPTION OF THE FEATURE in the manual.

it was already quoted by several folks throughout the thread.

Why do you keep repeating things you've already been given the answer to multiple times?

A quick copy and paste will suffice. I know what's in the manual.

And they are wrong.

Hence why I bothered explaining what actually IS the problem so they might understand what's actually wrong instead of continuing to complain about something that's not even broken.

If they are uninterested in understanding the actual problem- that's on them.

Joe Schmoe bought a Tesla, turned on AP, took it out in the highway, it phantom braked. Do you think he'll care about your speculation on what the problem is? No, he'll just want it fixed.

Go tell the soccer mom who almost got rear-ended with her kids in the car that the map data was wrong. What do you think her answer is going to be?

It can't.

Thanks.

Nor did you claim it did.

In fact you specifically claimed the speed was different in the far left lane and far right lane.

Which- and lemme know if you need me to go into this in more depth- is not the same place

I've had the car since March. I drive the same route to work everyday. I was commenting on ONE specific occasion to make a point. There have been days when it said the correct speed limit (55mph) in the same places it said 45. Today, it said 45mph throughout the whole route till the onramp to 95 south. There have been other days when it dropped to 35 just before the on ramp to I95. I'll see what it does tomorrow.

No, but that's not the "years old" one you were referring to... which as explained to you is largely special effects not a real demo.

I know it's not, hence the reason I asked. So your answer is no?

We do know none of the demo rides given to guests present that same day were done using the same system (all the demo rides were level 2 human supervised- rather than the one-time-only level 3 demo used to shoot the video).

Which is prudent. Why risk the well being of your guests with unproven software?

So whatever they used- it ain't what's on your car.

The point is, the car was able to do those things back then.

My last post. If you're going to claim their video(s) were possibly faked or manipulated, it's time to halt.

 

[Knightshade]

There are multiple videos showing the pre-rewrite software driving on city streets, one of which you chose to ignore and claim as false. That's pretty baffling if you ask me.

More baffling is why you're lying about this.

Not only did I not "ignore" the 2016 video- I explicitly debunked the idea it shows what you think it does- even citing a government source showing Tesla experienced 180 disengagements while trying to get enough seconds-at-a-time footage they could edit together from dozens of attempts to produce one video about as "real" as the last Avengers movie.


Apart from that special effects reel, Tesla has shown one other video of the car driving itself on city streets ever.

The autonomy day one.

Which again wasn't running production software, and nobody physically there for the demos got to actually ride in the car running that software.

They did one run with custom L3 software, and that was it.

So for all we know it was written just to drive that specific route.

We DO know it's over a year later and nothing has come to the public that does the stuff in that video though.

Here's your statement:

You were claiming they deliberately edited the clip to make it look like it did all that without touching the wheel. That's false advertising. Maybe you should re-read what you write.

So you claim I "ignored" the video- then quote me directly addressing why the video is special effects that don't actually prove anything?


Man- even you don't agree with you.

Also you don't appear to understand what "false advertising" means either.

They make no claims of any kind in the video other than it's a self-driving demo.

Which it is.

It's a mostly fictional one of course since to GET that video they had to splice together tiny slivers of dozens of runs of the same route, with over 180 disengagements that would require touching the wheel

So they only took the seconds or minutes at a time between disengagements and spliced them all together into what looks like but factually is not a single continuous run.


It's incredibly deceptive.

And, of course, you fell for it hook line and sinker.


Even now- presented with evidence from the State of CA that that's what happened- you seem to want to deny it for some reason.

But since they didn't actually make any CLAIMS in the video that it was a single take- nor that the current system you bought today could DO what is in the video- it's not "false advertising" because that term has a specific legal meaning you apparently need to look up.

I know what's in the manual.

This, too, appears not to be true.

Since you've been told multiple times and keep asking someone else to prove it by quoting the thing you claim you already know.

So it seems you don't know what is in there at all.

Since both the software version and the "intended for use on divided highways with limited access" are in there and you keep denying that too.


Denying facts seems to be a thing with you.

Joe Schmoe bought a Tesla, turned on AP, took it out in the highway, it phantom braked. Do you think he'll care about your speculation on what the problem is? No, he'll just want it fixed.

If he doesn't understand what's actually wrong then he's welcome to bitch about the wrong things all he wants, but it'll just be him wasting his time.

Which increasingly looks like why trying to discuss this stuff with you is too in fairness.

Go tell the soccer mom who almost got rear-ended with her kids in the car that the map data was wrong. What do you think her answer is going to be?

Probably that they should fix the map data.

Which- again- has nothing at all to do with the AP code.

Since, hopefully, the soccer mom is more rational about this than you seem to be



Thanks.

I've had the car since March. I drive the same route to work everyday. I was commenting on ONE specific occasion to make a point. There have been days when it said the correct speed limit (55mph) in the same places it said 45. Today, it said 45mph throughout the whole route till the onramp to 95 south. There have been other days when it dropped to 35 just before the on ramp to I95. I'll see what it does tomorrow.

I'd suggest you get some video of this happening- ideally showing the cars displayed limit and the actual road you're on (including the lane)

Show us some video of the same spot in the same lane showing a different speed limit in the car from day to day.


Because so far your story isn't really adding up- since without a map update this would be impossible for it to change the speed limit for the SAME physical spot.

The point is, the car was able to do those things back then.

Using custom software specifically written to do it? Sure.

That's utterly worthless info though since not only did nobody else get the SW- they didn't even trust it enough to do demos on the SAME ROUTE IT WAS WRITTEN FOR.


It was good enough to get one short demo video shot- and not good enough for anything else.

So that doesn't appear to be of any value at all to the discussion.

My last post.

I seriously doubt that

If you're going to claim their video(s) were possibly faked or manipulated, it's time to halt.

The 2016 video was proven as a fact to be manipulated.

This is well known.

They needed 550 miles of driving the same short route over and over- with 180 disengagements in those 550 miles- to get enough tiny clips of it "working" to edit together into ONE short demo video that tricked folks like you into thinking they did all that in 1 take.

Again the 2016 CA autonomous driving report confirms the numbers I just cited in public record.

Denial ain't just a river in egypt man

 

[1.21GW]

You both get internet points. +50

Can we stop this argument now?

 

[Knightshade]

It's less an argument and more me presenting facts and him refusing to accept them

BTW, this morning I did some driving- (HW3, current firmware and maps, emergency braking set to late)- about 35 miles one way with all but a couple miles on either end being interstate- 0 phantom braking events... then another ~15 mile trip one way with again all but the last mile or so on either end being interstate- 0 phantom braking events doing that trip twice (once each direction).

Only time from on-ramp to off-ramp during any leg of it I had to manually do anything was one interchange between two interstates was unsupported (it's one of those where the on ramp lane turns into the off-ramp lane with very very little distance from one to the other- something current NoA seems to kinda throw its hands up at).


Only difference (besides visualizations) I noticed having gone to HW3 is the car-initiated NoA lane changes to pass slower cars or follow route are maybe very slightly quicker/more aggressive than they were with 2.5...(though that could entirely be placebo) and they're still noticeably less so than driver-initiated AP lane changes.

 

[john5520]

Moderator: I'd like to request that my posts from #31 onward get moved to another thread or deleted. My counter arguments are quoted so this is simply for the sake of others who are reading. I just wade through the thread and it is a chore.

Today, the reported speed was 45 mph on the freeway, and just before the final bridge to the on ramp south, it switched to the correct speed: 55mph. I'll see how repeatable this is. Earlier along the way, It phantom braked fairly hard while approaching a large vehicle from behind which is rare (it usually happens when the vehicle is on the adjacent lane). What's also strange is I wasn't that close to it (maybe 8 car lengths).