Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

Crash with needs to be smarter summons

This site may earn commission on affiliate links.
Why do you think that? It is pretty simple, every time the user releases and the car is not within a few meters from its target location then treat the summon as a failure and else as a success.

Making it to the destination doesn't imply success. It might have caused a traffic jam, acted illogically, or pist off others in the lot before it got to the destination.

Also, Tesla should have trained it in-house.

Furthermore, since most owners are either not using it at all or are cherry-picking only the safest use-cases, doesn't seem to me like this sort of "training" would be very useful (if Tesla is even collecting the data at all).

I'm really annoyed by Tesla owners justifying playing with their toy by calling it "testing" or "training." Let Tesla test and train its devices in a more methodical way, seeking out the data it needs using trained testers.
 
Making it to the destination doesn't imply success. It might have caused a traffic jam, acted illogically, or pist off others in the lot before it got to the destination.

Also, Tesla should have trained it in-house.

Furthermore, since most owners are either not using it at all or are cherry-picking only the safest use-cases, doesn't seem to me like this sort of "training" would be very useful (if Tesla is even collecting the data at all).

I'm really annoyed by Tesla owners justifying playing with their toy by calling it "testing" or "training." Let Tesla test and train its devices in a more methodical way, seeking out the data it needs using trained testers.

Well, say what you want but letting go of the kill switch will very likely state failure. Those results are most interestinf. Yes, reaching the destination could have false positives but not that interesting when looking at the totals (yet).

There is so much you can test with your own resources.

Looking at the recorded failures is pretty funny. Very stupid human errors by their owners so far by not testing the feature in a controlled environment.

To be honest, I expected way more videos but it seems the feature works better than anticipated.

Cars sent data back to the mothership all the time. My car almost always sends data after parking via WiFi as indicated by the telemetry of my router. That will definitely be happening with smart summon results too. I'm located in the Netherlands so the feature will not be activated for likely many years to come.
 
Tesla's Smart Summon Performance Doesn't Match Marketing Hype
Consumer reports agrees that summons, needs to be smarter.

Same article:
"CR's experience with the system shows that Smart Summon can exit a parking space, turn and start moving toward the vehicle owner, and negotiate around stationary objects. It also can detect and stop for pedestrians and slow down if it senses cross traffic."

Snark mode on:
It was the best of implementations, it was the worst of implementations.
-it's also the only implementation

Next time, CR should test in the rain...

Lame joke: What do Smart Summon and the CR review of it have in common?
They're both all over the place.
 
Tesla's Smart Summon Performance Doesn't Match Marketing Hype
Consumer reports agrees that summons, needs to be smarter.
Can not disagree with the actual performance and use of Smart Summon as described in the article. The potshots and snakiness are on-par for Consumer Reports writers...but that doesn't override the fact that Smart Summon does not do well in most circumstances. I have only had it work 1 in 6 tries without having to intervene due to traffic or just heading in the wrong direction. ¯\_(ツ)_/¯
 
  • Like
Reactions: DanCar
Lame joke: What do Smart Summon and the CR review of it have in common?
They're both all over the place.

Realistically: Wouldn’t that where a good Smart Summon review then would be? All over the place?

Smart Summon is a controversial feature. On one hand it is a bold move and a unique offering in the consumer space, and also it is quite bad in other ways.

Isn’t this true for the entire Autopilot 2+ experience actually?
 
Realistically: Wouldn’t that where a good Smart Summon review then would be? All over the place?

Smart Summon is a controversial feature. On one hand it is a bold move and a unique offering in the consumer space, and also it is quite bad in other ways.

Isn’t this true for the entire Autopilot 2+ experience actually?
Yah, I'm all for balanced reporting, but this was too bipolar for me.
  • Says it is not useful but then says when it is useful
  • Says it has warnings about being not perfect, then complains about not being perfect
  • Relates 'data' implying it is unsafe, yet complains is that it is slow and over cautious
Of the 21 paragraphs, there are 5 paragraphs with actual test data reporting and 3 paragraphs regarding NHTSA investigations, one of which has nothing to do with Smart Summon.
 
  • Like
Reactions: DanCar
So I just got back from shopping. While taking the cart back in a light drizzle I noticed a Model 3 start backing into the lane. Then it pulled forward across 4 spaces and proceeded to cross two more rows of spots on a diagonal trajectory toward the dead end corner of the parking lot.

Overall impression was that the driver might have been better off using Summon.
 
  • Informative
Reactions: electronblue
I failed at communicating, should have said Smart Summon at the end as this car was manually driven.

So I just got back from shopping. While taking the cart back in a light drizzle I noticed a Model 3 start backing into the lane. Then it pulled forward across 4 spaces and proceeded to cross two more rows of spots on a diagonal trajectory toward the dead end corner of the parking lot.

Overall impression was that the driver might have been better off using Summon.
 
  • Informative
Reactions: electronblue