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.
Someone has shown that even the older design that has some light bleed doesn't normally impact the vision system. (What is displayed on the MCU is impacted much more because of compression/etc. than the raw data AP uses.)

Yeah you can totally tell the camera is not picking up the light at all, it's just image compression. I love how the cameras always capture two images the one everyone sees and the magic one that is much better that they actually use but no one gets to see. :)
 
Tesla and their stubborn insistence on vision is weird. It just goes against ever principle of design. Why would you ever just say we are sticking with this path no matter what. Why would you not continually improve and compare both options. It's just like their decision to use vision to detect rain, it is laughable, over complicated and silly. So many easier better ways to do it. This rain sensor cost a few cents per car, nope we can do it with AI and our vision and spend years of man hours working on it. I wonder how many cycles they piss away in the Ai trying to detect rain. The blinker cameras are another amazing Tesla moment. Lets put the camera right next to the blinker so at night when you change lanes the cameras will be blinded 50% of the time.
It’s easy to make suggestions, or not understand business decisions when you’re not at the helm.
Same way we sit back watching sports, and tell the quarterback we could have done that pass differently or better. :)

A .50 cent sensor has a huge bottom line. We also know that without innovation, nothing will change.

Not saying things can’t be improved, or even that it is the right decision. Time will tell, but Tesla feels it’s the right decision, and they’ve bet billions on it.
 
  • Helpful
Reactions: CanadaEV
Yeah you can totally tell the camera is not picking up the light at all, it's just image compression. I love how the cameras always capture two images the one everyone sees and the magic one that is much better that they actually use but no one gets to see. :)
Here is a link where someone shows the uncompressed 10-bit feed:


Notice that you can sort of see the light, but most of the time it isn't a problem.
 
I was not on autopilot, and there were nothing on the road, no car nor human in front, nothing. And the car just suddenly brake itself, it brake so hard that if there was a car behind me, I’m sure he would have hit me.
So your foot was on the accelerator, and the car just decided to brake out of nowhere? That's odd, I guess it triggered the crash detection?
 
What a F N tease to get the opt in, drive the car perfectly for 200 km. Park your car so your score can’t drop. Then. Nothing.

Elon just Berlin Germany’d us with needless red tape and old world bureaucracy.

Dude if you can reprogram a new chip in 5 days for your cars, you can roll out a beta program in the same time.
 
  • Disagree
Reactions: Vawlkus
so, i guess we are looking at anywhere between 3 days and 9 days lol

i may just opt out for the weekend and opt back in to reset. I've been stuck at 97 now for a while and cant seem to climb out of it. had 3 days (200km worth) of driving at 100, while keeping 3 seconds behind any car i could find. lol this is stress :)
 
Why wouldn't Elon overestimate the date than underestimate it and get everyone's hopes up?
Now I have 1500km since 2 weeks ago with this stupid safety score thing on. Wish I could've turned it on today so I only need to go through this stress for a week.
He should at least send out invites to people who already have good safety scores with decent mileage, then I don't care if FSD beta comes out next week or next month.
 
Why wouldn't Elon overestimate the date than underestimate it and get everyone's hopes up?
Now I have 1500km since 2 weeks ago with this stupid safety score thing on. Wish I could've turned it on today so I only need to go through this stress for a week.
He should at least send out invites to people who already have good safety scores with decent mileage, then I don't care if FSD beta comes out next week or next month.
You must be new to this whole Elon/Tesla estimating technique :)
 
Will decelerating with regen only ding you for hard braking? I thought it would not, but I think it must after my 97 drive today.

Anybody know for sure?
Some people say they have been dinged by it. In my Model 3 I feather the accelerator off for regen as my only braking, I’ve never been penalized. I have the most aggressive regen setting- I think it is at about -0.2g
 
Last edited:
Will decelerating with regen only ding you for hard braking? I thought it would not, but I think it must after my 97 drive today.

Anybody know for sure?
The Tesla safety score page defines the rate of deceleration the triggers hard braking. If your rate of deceleration was in excess of that rate then yes it would trigger hard braking.
 
Will decelerating with regen only ding you for hard braking? I thought it would not, but I think it must after my 97 drive today.

Anybody know for sure?
My safety score was so bad because I thought regen braking would be considered hard braking so I tried to go easy on regen braking and ended up having to actually hard brake to stop. I learned it the hard way that regen braking never counts. I now do full regen braking (like take the foot off accelerator completely) and it's really easy to maintain 100% score.
 
  • Like
Reactions: thesteve