2019.20.4.5 is out with Enhanced Summon

[WilliamG]

I do believe that Tesla will eventually get Enhanced Summon to work. But I wonder: would Enhanced Summon be easier to do with LIDAR?

No matter what Elon says, LIDAR would be another option to tap into if necessary. Just like rear RADAR. No reason to not have it except cost.

 

[Thp3]

I thought I read that Lidar uses a lot of watts. I read somewhere it was 40 watts. That would be a huge hit on Electric Vehicles. Is that another issue with Lidar ?

 

[P2AT]

I believe the biggest problem besides cost and electricity consumption is re-creating what they need for FSD through a neural net with LIDAR information.

The road to FSD will be focused on the amount of data. Billions of miles. This can be achieved with the cameras on 400k+ Teslas on the road. To retrofit all these with LIDAR is costly, and a logistical nightmare (as well as unsightly)... And THEN require Billions of miles recorded with these LIDAR sensors, which is redundant because they operate on the visual spectrum (photons)

Honestly I think Tesla's approach is correct. Perhaps more redundant cameras might be beneficial but cameras is the smart way to go, especially as a mass market option, when price is important. If they can achieve statistically safer driving with cameras alone compared to human drivers, LIDAR would be hard pressed for consumers to pay for that extra cost. I think the biggest takeaway is: we are all able to drive fine without shooting photons out in 360 degrees and building maps in our head before turning the steering wheel. And Tesla's already have more "eyes" (cameras) than the two we have now!

Then again, what do I know. I'm not an engineer

 

[Lunares]

Here’s how I think about Enhanced Summon. Under the Self Driving umbrella, there are 3 driving “modes”: the parking lot, city/neighborhood streets, and highways. What’s really exciting to me is that Tesla has basically solved (or is close to solving) the parking lot piece of the Self Driving puzzle. The data that Tesla will get from the groundwork they’re laying now will pay off enormous dividends in terms of future Enhanced Summon capability. (Remember the goal - the dream - is that some day we will be able to Summon our cars from across the country.)

And, Tesla has most of highway down. In fact as long as traffic is light to medium I would say NoA covers 95% of highway driving, with the remaining 5% being weird curves

So on the highway front Tesla needs to improve lane changes, especially in heavy traffic

 

[srs5694]

And, Tesla has most of highway down. In fact as long as traffic is light to medium I would say NoA covers 95% of highway driving, with the remaining 5% being weird curves

Not for true level 4 or 5 autonomy. Autopilot flakes out in too many common situations, like some highway forks (it was once driving down the middle of one of those for me, straight into a barrier, before I disengaged it), construction zones, avoidance of potholes (maybe not as common in San Diego, but here in the northeast, they're far too common), etc. NoA is aware of the traffic in its immediate vicinity (one lane over, one or two cars in front, and one car behind), but seems to be oblivious to everything else. As a result, it suggests some quite stupid maneuvers, like merging into a lane that's closed off ahead or that's faster right there but slowing ahead. It also fails to slow when approaching stopped traffic until the last minute, which is disconcerting at best. I suppose the car could drive through many of these problems, but not as well as a human, and some of these limitations could result in damage to the car or even injury to the passenger.

That said, NoA does reasonably well on most highway driving, in terms of miles -- probably 90%, and maybe a point or two more than that, which isn't numerically all that different from your estimate. The trouble is that it's just not good enough in the remaining conditions, and those are so varied, and so important from a safety perspective, that they'll take a lot longer to get right than the first 90%. This is actually pretty common in programming; you can get to 90% functionality pretty well, but it takes as long or longer to get that final 10%. (Of course, Tesla's neural net approach differs significantly from traditional programming, but I suspect a similar rule applies.)

All that said, Elon Musk seems pretty confident that Tesla will be able to deliver significant improvements in FSD in just a few months. He obviously knows more about what Tesla's cooking up than we do, and perhaps the new computer will enable a dramatic improvement, which could cut back on feature development time. OTOH, Musk has a history of offering, shall we say, optimistic timelines. Overall, I'm confident that Tesla's FSD features will improve, but whether my Model 3 will ever drive well enough for me to nap on a trip is an open question.

 

[SergeyUser]

Can someone on EAP suggest Tesla to auto-enable Standby Mode when turning air conditioning on from Tesla mobile app? Think about it - if you do that, it means you are about to drive your car, and possibly doing it a few mins in advance. So there is no harm to enable standby for a short period of time.

 

[JayNJ]

You don't need Enhanced Summon to do that.

No kidding. I was responding to Summon comment. Enhanced is not ready for prime time.

 

[JayNJ]

You apply wheel shine every week?

Pretty much, when I wash the car. It very easily comes off when driving in heavy rains.

 

[jjrandorin]

Different program. The referral prize at one point was for “priority access to software updates”. Just earlier access to public releases (which hasn’t really played out based on my experience). This is different than the Early Access Program which is a formal invite-only beta program requiring an NDA.

I was reading this thread and was going to post this, but @commasign beat me to it. The referral program award is "priority updates", not early access program. The "priority updates" from the referral program seem to do nothing for me (I earned that award from referrals). I do not get updates any sooner than anyone posting here, and in fact seem to get them a couple weeks after on average. I have wifi in garage, etc.

 

[mswlogo]

Not for true level 4 or 5 autonomy. Autopilot flakes out in too many common situations, like some highway forks (it was once driving down the middle of one of those for me, straight into a barrier, before I disengaged it), construction zones, avoidance of potholes (maybe not as common in San Diego, but here in the northeast, they're far too common), etc. NoA is aware of the traffic in its immediate vicinity (one lane over, one or two cars in front, and one car behind), but seems to be oblivious to everything else. As a result, it suggests some quite stupid maneuvers, like merging into a lane that's closed off ahead or that's faster right there but slowing ahead. It also fails to slow when approaching stopped traffic until the last minute, which is disconcerting at best. I suppose the car could drive through many of these problems, but not as well as a human, and some of these limitations could result in damage to the car or even injury to the passenger.

That said, NoA does reasonably well on most highway driving, in terms of miles -- probably 90%, and maybe a point or two more than that, which isn't numerically all that different from your estimate. The trouble is that it's just not good enough in the remaining conditions, and those are so varied, and so important from a safety perspective, that they'll take a lot longer to get right than the first 90%. This is actually pretty common in programming; you can get to 90% functionality pretty well, but it takes as long or longer to get that final 10%. (Of course, Tesla's neural net approach differs significantly from traditional programming, but I suspect a similar rule applies.)

It’s called the 80:20 Rule, not the 90:10 rule

https://www.quora.com/Does-the-80-20-rule-apply-to-programming

Keep in mind, Elon’s “feature complete” goal could easily be the 20% effort mark. Only 80% more effort after feature complete. We are, after all, have never even exited first Beta yet.

 

[bradhs]

I am part of the Early Access Program and have no referrals. I did take possession of my Model 3 Performance last August, and was invited to participate shortly afterwards.

I got my Model 3 Performance last August too but was not invited to the Early Access Program.

 

[BioSehnsucht]

No matter what Elon says, LIDAR would be another option to tap into if necessary. Just like rear RADAR. No reason to not have it except cost.

While rear RADAR would add functionality (seeing through fog/smoke/etc), LIDAR won't add anything that pure vision can't do, except as an expensive short term bandaid (depth from vision is relatively trivial even if it hasn't been implemented yet by Tesla outside of FSD development, even with a single camera, LIDAR only offers a shortcut that would be supplanted by vision anyways, and it has no other strengths like RADAR)

 

[Chewy3]

I use it every week in my driveway when cleaning my tires and apply wheel dressing. Apply dressing, move forward or reverse about 3 feet, apply dressing to the bottom part of the wheel you couldn't get to before.

....you know your not supposed to put any of those products on your tread, right? They are mostly silicone based and act as a lubricant = slippery treads.

 

[paulvinh]

Man early access is really keeping quiet this time around.

 

[pepperidge]

My force field is greater than 150 feet based on the measurement with Google map.

 

[BOBTHEJOCKEY]

Well, I guess we'll have to wait another 6 months before you like it then.

I'm glad to get Beta and watch it get better. That's more fun than just getting it (much later) and everything works. I would have been extremely happy if it went 0.5 mph.

"The more you experience, the more you see."
(me)[/QUOTE
BEARS BEWARE the autopilot MONSTER is a comin to get your money. If summons and autopilot keep getting better how can it not be just a matter of time before they have real value? Patience grasshopper.

 

[StellarRat]

I've used the summon feature once in the real world: at a CVS during a torrential downpour. It saved me from getting extra wet trying to get into the car in a tight spot. I see it occasionally useful in these rare situations.

This is exactly how I look at the usefulness of ES right now. I would add to that that would not use it in a busy parking lot no matter what the conditions. It's not smart about the direction of traffic flow and it's real slow.

 

[woodisgood]

....you know your not supposed to put any of those products on your tread, right? They are mostly silicone based and act as a lubricant = slippery treads.

Can’t speak to the person you quoted but Summon is helpful for wheel dressing that is applied as intended - on the sidewall.

 

[Alset Srotom]

My force field is greater than 150 feet based on the measurement with Google map.

How is v3 performing?

 

[Silicon Desert]

You must be hard to buy for at Christmas, if you don’t care about features like this....

ha ha, that's a good one I guess you can put me in the "hard to buy for" category. I wouldn't call it "advanced" summon. I consider it "improved" summon.