Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register
This site may earn commission on affiliate links.
Even the 5mp HW3 cameras are capturing more data than we know what to do with it.

Pedantic correction: HW3 is 1.2MP, HW4 is 5MP.

(Regardless, I agree with the sentiment that pure camera resolution is far from the limiting factor. Basic decision making is still very problematic, let alone camera positions for cross traffic visibility. Certainly curious to see what HW5 brings but I'm not expecting purely higher resolution cameras)
 
I heard that if it’s been a long time since your vehicle has checked for an update it’s because Tesla is holding back checking your vehicle because they are sending you an update soon so there’s no need for your vehicle to check for an update before then

So the longer it’s been since your vehicle has checked for an update the sooner you will be getting the update GL all

Me and fasteddie have the longest time since last check up I believe
My Model 3 with FSD trial is still waiting around on 2024.14.7, the Y with FSD is still on 2024.3.25
I know which one I'd rather have get FSD12.4.1 😎
 
  • Like
Reactions: FSDtester#1
Good to know in advance this is how attention monitoring works. Once it gets into the blue flashes, simply returning your attention to the road is not sufficient, you need to tug on the wheel:

I see what it is, he got the blue flashing plus the red hands on the wheel not just blue flashing, that’s why eyes alone weren’t enough

If you only get blue flashing then eyes alone are enough
 
  • Like
Reactions: willow_hiller
That is why FSD speed is slowed down in bad vision weather just like how humans do.

That being said, even a car with LiDAR and RADAR will have to slow down in bad weather, not because it cannot detect objects but because it still needs to use vision to identify the road, the lanes, the signs everything else that is made for human vision.
AV's have the advantage of GPS and pre-mapped roads for much of this. Lane lines a couple meters away (to establish current exact position) are much easier to see in bad weather than cars or obstacles ~100 meters away which still might require pre-emptive braking; the latter is a huge advantage of radar (in bad weather) and Lidar (in low lighting) over pure vision.

My concern is that Tesla's cameras tend to become far more degraded in bad weather than human vision does, due to their proximity to the glass. A single water drop can render a Tesla camera completely useless, whereas a human can just move their head around a little. Here is what my 2022 Model Y front camera sees in light drizzle (dashcam capture):
This specific degradation is not strong enough to impede lanekeeping, but it's certainly strong enough to interfere with control planning and distance vision.
 
Last edited:
Sorry, I meant to say left lane. I saw the no left turn sign. But more importantly for item #2, zero chance FSD has contextual awareness and even if it did it's well before the intersection.
Why do you say it has no contextual awareness? FSD is trained on video clips that are long enough to provide context, and certainly takes the last ~30 seconds into account when making decisions. The occupancy network is persistent, providing at least short-term "memory" of objects and places the car previously saw but can't currently see.
 
Pedantic correction: HW3 is 1.2MP, HW4 is 5MP.

(Regardless, I agree with the sentiment that pure camera resolution is far from the limiting factor. Basic decision making is still very problematic, let alone camera positions for cross traffic visibility. Certainly curious to see what HW5 brings but I'm not expecting purely higher resolution cameras)
Completely agreed that resolution is not the limiting factor. Dynamic range and low-light clarity are far more important, as well as positioning (bumper cameras) and resilience to dirt/rain obstruction. On the other hand, once the car is able to read and interpret nontrivial street signs, the increased resolution would allow them to be read from that much further away. I wouldn't be surprised if HW5 may have an articulated telephoto camera for this reason, filling up the third (currently unused) slot in the forward housing.

Pure vision may be just fine for 99% of driving conditions, but 99% or even 99.9% is not enough for L4. It's the other 0.1% where additional sensor modalities (radar/Lidar) can still add a huge amount of value, IMO.
 
My concern is that Tesla's cameras tend to become far more degraded in bad weather than human vision does, due to their proximity to the glass. A single water drop can render a Tesla camera completely useless, whereas a human can just move their head around a little. Here is what my 2022 Model Y front camera sees in light drizzle (dashcam capture):
I had the same doubts until I started tracking the visualizations and reactions, and boy it sees everything needed to stay safe better than I can even in the bad weather.
 
I heard that if it’s been a long time since your vehicle has checked for an update it’s because Tesla is holding back checking your vehicle because they are sending you an update soon so there’s no need for your vehicle to check for an update before then

So the longer it’s been since your vehicle has checked for an update the sooner you will be getting the update GL all

Me and fasteddie have the longest time since last check up I believe
That checks out I usually get all 3 of my cars at the same time very close to when employees do
 
  • Informative
Reactions: FSDtester#1
Good to know in advance this is how attention monitoring works. Once it gets into the blue flashes, simply returning your attention to the road is not sufficient, you need to tug on the wheel:

This drive raises my optimism about 12.3.x-->12.4.x being a definite noticeable step in improvement. We will need to see more videos, but so far, this is what I was expecting/hoping for.