Is this the future of DreamDrive?

hydbob

Referral Code - R0YBCKIJ
Admin
Verified Owner
Supporting Member
Joined
Nov 14, 2021
Messages
10,250
Location
Los Angeles, CA, USA, Earth, Milky Way
Cars
Dream P
DE Number
33
Referral Code
R0YBCKIJ
Seems like NVIDIA is becoming popular is autonomous driving. I hear Volvo and Jaguar are using them as well.
 
I must be the odd one out. The pleasure of being mobile for all these years is the actual interaction of driving a vehicle. The last thing I want it to do is drive itself.
 
I’m really keen to see how DD pro develops. I love the drive on the GT but not every drive is for pleasure and so having the option of an advanced autopilot would be awesome imo.
 
I must be the odd one out. The pleasure of being mobile for all these years is the actual interaction of driving a vehicle. The last thing I want it to do is drive itself.
I’m with you
I want to drive… unless for some reason I am unable to. Eg 97 and blind :) Or asleep at the wheel driving back from scottsdale at 3am
 
...Or asleep at the wheel driving back from scottsdale at 3am
My aunt was crippled by a similar accident after visiting her husband in the hospital. I'm very, very eager for workable self-driving with LIDAR object detection (Tesla still can't detect fixed objects on the road), even if it's only on freeways.
 
The cruise cars have a metric ass-ton of LiDAR so the Air is definitely not getting that level of autonomy with the current tech anytime soon :D

That said, Lucid is the only current production car with LiDAR (the Lotus Eletre will have four when it launches year). I do think we will get level 3 sooner rather than later, hopefully at highway speeds. There are lucid LiDAR mapping cars going on major California freeways right now so I’m optimistic it’ll be this year.

The Tesla camera-only system I’m extremely skeptical of. Why artificially constrain the inputs to the system? By not having any long range active sensors they’re going to have to rely on a very very very well trained neural network but it will always be at a disadvantage vs. cars with better long range active sensors. It baffles me.
 
NVIDIA provides software development kits SDKs that start out not knowing specific attributes of the car. The SDKs have all of the subroutines needed for ADAS level ?. Lucid has to write software to integrate the NVIDIA SDKs with the Air. It requires a lot of testing to be safe and effective.
 
The cruise cars have a metric ass-ton of LiDAR so the Air is definitely not getting that level of autonomy with the current tech anytime soon :D

That said, Lucid is the only current production car with LiDAR (the Lotus Eletre will have four when it launches year). I do think we will get level 3 sooner rather than later, hopefully at highway speeds. There are lucid LiDAR mapping cars going on major California freeways right now so I’m optimistic it’ll be this year.

The Tesla camera-only system I’m extremely skeptical of. Why artificially constrain the inputs to the system? By not having any long range active sensors they’re going to have to rely on a very very very well trained neural network but it will always be at a disadvantage vs. cars with better long range active sensors. It baffles me.

This. I do not want my car to drive with the skill and vision of the “average” driver as it is bound to revert to the mean when using only cameras and learning from existing driving. The “average driver” drives like garbage which, incidentally, helps explain why my experiences with FSD have, by and large, been terrifying.
 
My aunt was crippled by a similar accident after visiting her husband in the hospital. I'm very, very eager for workable self-driving with LIDAR object detection (Tesla still can't detect fixed objects on the road), even if it's only on freeways.
AND … if it was good enough, the car might even be (part of) a solution for her

so sorry to hear this happened to your aunt. I had a scare 30+ years ago, and it taught me a lesson. Sleep impacts driving as much or more than (a lot of) alcohol
 
...The Tesla camera-only system I’m extremely skeptical of.....
I agree- it seems that they haven't solved 3D object segmentation or ranging even though they have stereo forward-facing cameras. To rely on AI for object detection on a freeway, they'd have to train it to recognize every possibly dangerous object at any angle: mattress, ladder, goose, armchair, TV set, etc. LIDAR is a big reason that I'm considering the Air, or Polestar 3 or Volvo's sister version of the same.
 
Last edited:
Or
The cruise cars have a metric ass-ton of LiDAR so the Air is definitely not getting that level of autonomy with the current tech anytime soon :D

That said, Lucid is the only current production car with LiDAR (the Lotus Eletre will have four when it launches year). I do think we will get level 3 sooner rather than later, hopefully at highway speeds. There are lucid LiDAR mapping cars going on major California freeways right now so I’m optimistic it’ll be this year.

The Tesla camera-only system I’m extremely skeptical of. Why artificially constrain the inputs to the system? By not having any long range active sensors they’re going to have to rely on a very very very well trained neural network but it will always be at a disadvantage vs. cars with better long range active sensors. It baffles me.
Signed a purchase order and deposit made for an ELETRE! Can’t wait for the gold interior accents! Then can’t wait for and Electric version of Emira or something like it from Lotus.
 
The cruise cars have a metric ass-ton of LiDAR so the Air is definitely not getting that level of autonomy with the current tech anytime soon :D

That said, Lucid is the only current production car with LiDAR (the Lotus Eletre will have four when it launches year). I do think we will get level 3 sooner rather than later, hopefully at highway speeds. There are lucid LiDAR mapping cars going on major California freeways right now so I’m optimistic it’ll be this year.

The Tesla camera-only system I’m extremely skeptical of. Why artificially constrain the inputs to the system? By not having any long range active sensors they’re going to have to rely on a very very very well trained neural network but it will always be at a disadvantage vs. cars with better long range active sensors. It baffles me.
It’s constrained because Elon insisted it can be done and has refused to allow them to add in LiDAR.

As time goes on, they will get left behind unless they change their perspective here.
 
It’s constrained because Elon insisted it can be done and has refused to allow them to add in LiDAR.

As time goes on, they will get left behind unless they change their perspective here.
Elon is brilliant and has assembled one of the best engineering teams on the planet, so if anyone can do it with vision alone it is Tesla. But still, my skepticism remains.
Or

Signed a purchase order and deposit made for an ELETRE! Can’t wait for the gold interior accents! Then can’t wait for and Electric version of Emira or something like it from Lotus.
The eletre looks absolutely stunning. I am absolutely a fan of it!
 
I agree- it seems that they haven't solved 3D object segmentation or ranging even though they have stereo forward-facing cameras. To rely on AI for object detection on a freeway, they'd have to train it to recognize every possibly dangerous object at any angle: mattress, ladder, goose, armchair, TV set, etc. LIDAR is a big reason that I'm considering the Air, or Polestar 3 or Volvo's sister version of the same.
Volvo owns half of Polestar and private Chinese company owns both. FYI. I’m sure most know this. I just read it today. Geely also owns Lotus.
 
Back
Top