2017.46 3387a54 is out

[calisnow]

driving 85 on left lane with car pool barricades to your left? you should be crazy.

If @buttershrimp can't stop the madness boy are we all in trouble...

 

[BlueMX75D]

Which firmware are you on? Touch the big "T" at the top center of your screen. The screen that then pops up will have your firmware written at the bottom.

FW version: 2017.44 0fdc86

 

[Electroman]

It was actually nerve wracking just watching that video. I wouldn't do 85 even if there is no traffic on either sides and empty lanes on a straight stretch.

 

[buttershrimp]

driving 85 on left lane with car pool barricades to your left? you should be crazy.

Also what happened at 1:35? it automatically shifted to car pool lane ?

Two things happened: 1) the lyrics of the song previously indicated that the rap artist caused this listener to crap his jeans 2) I autolane changed into the toll lane. At the end I tried to change lanes through the plastic dividers but failed as it shows on the camera.

 

[buttershrimp]

It was actually nerve wracking just watching that video. I wouldn't do 85 even if there is no traffic on either sides and empty lanes on a straight stretch.

My friend, the camera ads 10 lbs and 10 mph. Also, I know the route, i tightly control the variables and I stay frosty on the wheel. No messing around, just hard shrimpin’ and beautiful AP2

 

[calisnow]

It was actually nerve wracking just watching that video. I wouldn't do 85 even if there is no traffic on either sides and empty lanes on a straight stretch.

You are one very cautious man Nothing wrong with that of course.

 

[Ken7]

Ahhhhhhhhhhhhhhhhhhhhhhhhhhhh FREAK OUT!

New Autopilot Challenge Video with Your Friendly Buttershrmip.

The video is fun to watch, but I find myself praying for a total failure of my ipad’s audio circuits while watching. Ah, the volume control, the volume control!

 

[croman]

What do you mean? Like it veered for "left turn only" lanes it hadn't before?

Yes exactly. Even ones with dashed lines which was odd.

 

[jjh1234]

Ahhhhhhhhhhhhhhhhhhhhhhhhhhhh FREAK OUT!

New Autopilot Challenge Video with Your Friendly Buttershrmip.

Pimpin' and shrimpin', I love it.

 

[_jal_]

My friend, the camera ads 10 lbs and 10 mph. Also, I know the route, i tightly control the variables and I stay frosty on the wheel. No messing around, just hard shrimpin’ and beautiful AP2

You really upped your game. I *did* notice the shrimp logo in the corner. Seems like eons since you posted your first portrait mode movie.

 

[_jal_]

Yes exactly. Even ones with dashed lines which was odd.

It seems like that has come and gone with releases. It's really annoying, but it has usually been ok w dashed lines. Oh well, we can wait for 2017.48

 

[croman]

It seems like that has come and gone with releases. It's really annoying, but it has usually been ok w dashed lines. Oh well, we can wait for 2017.48

This one is a step backwards in this aspect. It also seems to dart in and out rather than fully commit. I don't like that because it is jerky and sudden.

 

[d21mike]

Ahhhhhhhhhhhhhhhhhhhhhhhhhhhh FREAK OUT!

New Autopilot Challenge Video with Your Friendly Buttershrmip.

Looks like AP2 is working very well indeed.

 

[BigD0g]

Ok, we need to chimp in together and get @buttershrimp an internal tripod....

 

[Hey_Hey]

I just drove about 40 miles on the interstate after receiving the update last night. It seemed to me that the changing lanes has become smoother than it was previously. I thought the transition was fairly jerky before, but today it seemed smoother. Of course, maybe this was just placebo effect on my part from the update last night.

 

[_jal_]

This one is a step backwards in this aspect. It also seems to dart in and out rather than fully commit. I don't like that because it is jerky and sudden.

What about highway exits? Any experience w that? I've got a 700 mile trip to PA tomorrow that would really stink if I'm wrestling my car at every exit.

 

[BigD0g]

Well I gotta do something to generate some suspense for myself. Also I'm basing this on on a few trends with Tesla and Musk: 1 - Musk likes to at least make symbolic deadlines (delivering a few hand built cars by a stated deadline). 2 - Dec 31 will be a full year late on the FSD claim - seems like squeaking in at least one feature by EOY will help him symbolically. 3 - I don't make any claim as to how well this feature will work. 4 - it may come with a bunch of disclaimers saying "You better watch your ass while it makes these auto lane changes because they are still very glitchy." 5 - The recent speech by whats-his-face at the owner event in Europe said we'd be pleased with AP progress in the coming weeks and I believe he even mentioned the current quarter. So hopefully that was a tease of a feature to drop in 2017.

Can anyone get in on this bet? I'm betting even this walk back won't happen by Dec 31 this year. The "only" way they can possibly ship this feature is if 100% lane detection on highways works even with HD Maps 100% of the time, and that's just not the case, so they might solve that by Dec 31, which would cause your feature to slip to next year.

No the big bang for Dec will be rain sensing wipers (doubtful, but possible) , vector maps and then @boonedocks will probably have Elon's baby in joy.

I suspect they walked back the NN for rain sensing wipers because it slipped into the release inadvertently and not because it was crap code, but honestly who the heck knows anymore.

 

[Joe F]

Ok, we need to chimp in together and get @buttershrimp an internal tripod....

You don't monkey around, do you?

 

[Joe F]

What about highway exits? Any experience w that? I've got a 700 mile trip to PA tomorrow that would really stink if I'm wrestling my car at every exit.

I don't have .46 yet, but .44 is okay as long as there are dashed lines at exits and entrances. Crap shoot without them though, but from those I've seen, it's better than older releases.

 

[BinaryField]

As we say in aviation - what's your mission, if I may ask? And how long do you plan to keep the car? Personally I think if you look at the progress of AP2 in year one it has been spectacularly fast and should encourage you not discourage you.

Stop focusing on Elon's imaginary timeframes. Stop listening to the hater brigade for a minute and start paying attention to people like @verygreen who bring real knowledge, and look at the actual progress rate of AP2/2.5. Do some reading about neural nets and reinforcement learning and what is happening on the cutting edge. Remember that learning involves lots of struggle with no visible progress and then breakthroughs in performance happen. Remember that the brain of AP 2/2.5 is a swappable board. Finally, remember that Tesla is still, one year later, betting the company on vision and selling the FSD option. They are ramping up the Model 3 with a vision based autonomy system. This says something. Also remember a lot of progress is likely happening in traffic logic simulation that we know nothing about.

The simpletons here on TMC believe that "No me car turn by itself! Been one year! Me angry! Me car no workie! Me angry! Elon dumb! This impossibowl! See Waymo! See big spinning dome! Dat key! Me want spinning dome too! Why Elon dum?"

For me, my mission is driving long distances on highways for business. AP 2 suits this mission perfectly in current 0.42 form.

To believe the software will NOT improve drastically from here you have to buy into several ridiculous claims:

1 - The autopilot team suddenly stops making progress.
2 - Tesla mysteriously becomes the only self driving player unable to teach traffic logic to its cars, despite having a world class AI team and having the largest fleet of vehicles in the world by orders of magnitude, gathering image data for learning.
3 - Despite the fact that vision plus radar is already reliable for lane keeping and cruise control, for no apparent reason Tesla is unable to make the cars turn left and right without hitting things.

At this point vision is great and getting better. The real problem is traffic logic - that's where the hard part is and that has nothing to do with sensor tech. That's stuff like "If I cut one lane over what will the guy two lanes over do?"

George Hotz gives a great talk on this recently where hen points out that sensing is the easy part - the hard part is decision making and that has zero to do with sensors.

Also remember that reinforcement learning is bleeding edge tech and Karpathy is an expert at it.

For other stunning reinforcement achievements see Alpha Go Zero's recent crushing of the original Alpha Go. @verygreen gives some good comments elsewhere here recently on the challenges of reinforcement learning for driving.

But if you want guarantees that the current setup will take us to Level 5 - I think you're asking too much. Very robust L3/4 I think is the likely outcome. Lack of rear radar or a camera high off the ground is my concern.

However even during a recent rainstorm I was able to discern incoming vehicles from behind on the rear camera despite it being covered in vision distorting water drops.

I can see us getting to a system that alerts you to take over if it loses confidence in its rear vision. So a Level 4 that requires frequent cleaning of the rear camera (maybe a quick wipe before each departure and then a touch up wipe every few hrs).

I'd say that my primary use case for AP would be to take care the majority of a daily 90 mile round trip commute to work. Secondary would be to have the ability to take road trips more often. It would definitely take the fatigue and frustration out of stop and go traffic or prolonged time in the car. I rented an AP1 vehicle last month and was extremely impressed with its performance and convenience.

As for life span, I plan to have my next car last for five or more years. I don't really care to change my car more frequently than that, so I'd like something that will be functional and can somewhat scale into the future. I realise that this is the Tesla game plan, and you do make a compelling argument. I just question the timeline on which this will be done and what features they intend to deliver in their mature build. If development is going to drag on or they go back on features, I'm better off letting the competition enter the market so that I can have options and get the best deal (be it Tesla or some other company). You can't fault me for being a cautious and value-oriented consumer.