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

2020.24.6

This site may earn commission on affiliate links.
I'm guessing that there is a lot the cameras see but is not displayed on the screen that we are allowed to see.
Such shapes such as cones and wheelie bins are displayed but the massive majority are not. I saw a stop sign displayed this week.
It was a fairly unusual hexagon shaped one whereas we more often have the give way signs that it ignores at the moment.
Many passengers mention that passing cars are shown after they have passed. I rather hope that is simply a lag in display and not in detection!
 
Nope. There’s a set of lights near me where if your first in the queue, the lights appear on screen with the appropriate colour as you draw near. But at the line, they’re too high for the camera to see them so all colours are grey and changes aren’t seen.
Had a similar thing yesterday. Big junction, loads of lights, all seen but nearest set went grey when I got to the line.

Yep, have noticed the same problem quite frequently.

The only possible solution to this (in Europe) is that, if it is first up, FSD must stop the car up to 5m before reaching the line, such that all lights pertinent to that lane remain in view of the cameras.
 
Sounds like a good plan from the car's viewpoint but not necessarily for the traffic lights.
Seen the problem with a lot of new drivers who don't move up to the stop line and then wonder why the lights aren't changing for them.
In order for the lights to correctly detect the vehicle you have to cover the sense wires.
With all too many traffic light junctions with no repeater light at the other side of the junction it is a toss-up between stopping short to still see the light and hope the lights have "seen you" and will change for you.
 
Can't see how that would work with sunglasses, tinted glasses, maybe peaked caps etc

There is also face tracking and expressions that add useful information too. So eyes only play a part. This has been solved for quite some time by others. We 'just' need Tesla to solve it with the view that they have to work with. Whilst not perfect, I think they have enough to work with. They perhaps do not need to differentiate between different types of distraction that these systems are capable of, and 'just' need to answer the question of whether the driver is concentrating on the road and able to take over control of the vehicle.

I was in a US University* live video lecture covering this 3 or so years back, and from what I remembered, everyone seemed to be on top of the problem and it was more of an aside about what techniques were being used to determine levels of driver distraction.

* can't remember if it was Stanford or MIT and unfortunately I lost access to this insider research resource when I left the research company I was working for.
 
3am - still only a few dozen cars though taking 2020.24.x so still early days. 2020.24.6.1 seems to be update only for cars previously on 2020.24.6, so quite possibly an 'emergency' update.

Software version 2020.24.6.1 was just detected on a Tesla Model 3 Long Range in Texas United States. The highest previous version was 2020.24.6. Follow the 2020.24.6.1 release at TeslaFi.com Firmware Tracker .
 
Last edited:
Yes very interesting. I feel like a randomly selected guinea pig.

Anyway clearly they didn't like what they saw from my totally non driving yesterday because this morning I have woken up to

2020.24.6.1 !

As someone who runs a software company i know what an emergency release looks like!

I havn't got time to apply it before i drop my daughter to school so hopefully its not totally unsafe. I will let you know if the car explodes.
 
We may have a wait for traffic light functionality

https://twitter.com/elonmusk/status/1275762954860822528?s=21


0A17700D-28CD-48DA-9076-47903CE5C3DE.jpeg
 
I wouldn't say that we are paying the same price for less. The US had a $1k price rise back last iirc November that never filtered down to Europe or I think much of the rest of the world - that coincided with introduction of smart summon in US. So if we don't get the price rise on July 1st, then we will be two price rises behind. Looking forward to see what comes along to justify the July 1st rise, even if we don't benefit for some time. Does anyone actually use the smart summon functionality to know if its improved much since its introduction - restrictions excepted.

As for stop light functionality, iirc that it was originally mooted for Q3/Q4. I think AUS getting it was a bit of an early surprise. Good that EM recognises the more diverse signs, lights and rules that we have over here. And being delayed, hopefully we benefit from some of the early refinements. Sometimes I get the impression that things are a little too simplified/lack of thought for our needs.
 
FSD has a way to go. I swear it’s not clear that left is a default lane.
I was on a section of the A47 where it went from single lane to dual carriageway. I was hugging the left as my turnoff was a few hundred yards away.
As we I got to a point just prior to two lanes but widening, I got pulled into a course toward the right hand lane and had to snatch it back.
The edge of carriageway marking was clearly marked but the car obviously wanted to hog the middle of the widened single lane.

if a police car had been following I would have expected to be pulled over for that manoeuvre.
 
  • Like
Reactions: Roy W.
Probably my biggest FSD gripe is its inability to handle a flare (road split into more lanes). On one occasion, M3 south exit J5, it physically drove me off the carriageway. I was testing and was aware how bad it was at that turnoff, but how bad it was on that occasion surprised me. If I hadn't taken control, I'm not convinced it wouldn't have clipped the sign and taken me into the Armco. Lane splits, on ramp and offramp definitely not fit for purpose.
 
  • Like
Reactions: Simavon and Roy W.