Waymo will beat Tesla FSD because of LIDAR

[REM]

It is also a lot more accurate for distance than 2 cameras could ever be.

This is not correct lol. The definition losses alone in the grade of radar that consumer vehicles have is not high enough to cover the +or- accuracy comparison between the two.

But the argument here is that adding radar, make it better and more accurate, it takes away nothing.

There are several reasons why they got rid of the radar; chief among them was phantom breaking introduced by radar artifacting. Mostly due to reflections in the environment. I experienced this first hand because my M3 has a radar. Had at several hard braking phantoms that disappeared when they removed it from the stack.

Both radar and lidar add too much noise to be useful for autonomous driving. Both technologies are fantastic to use in a lot of other applications. Driving is a very unique nut to crack.

Sponsored

 

[REM]

More absolute data, absolutely equates better outcomes.

I can tell you have never done anything remotely related to Ai in a professional setting.

You can add exabytes of data to the equation and it will not help you if the data is dirty. ?‍

You can absolutely bake a turkey at 1300 degrees for an hour, if you break it up into pieces and fan it out.

Can you explain to me the exact process in which you can bake a turkey at 1300 degrees for an hour and walk away with anything more than charcoal.

I think you are just trolling me here.

 

[REM]

If you need millimeter-level accuracy for your driving, you're doing it wrong

Or you're doing it right, and will be a very, very rich and famous F1 driver.

 

[cybertruckvegas]

Cameras are enough for driving as driving isn't super precision to mm. Monocular depth is already pretty good and they've made giant strides in last few months I've noticed in my own work. Eg Apple vision pro can get decent depth from single zero shot image to create 3d stereo image prob using depth anything v2 model.

People can close one eye and still throw a basketball or shoot a gun and hit small targets.

I really don't think Lidar cost is that big of a deal. The laser diode fundamentally is not that expensive and ASIC is probably $1 in silicon. I bet Lidar sensor adds $2 in silicone and cost is comparable to normal CMOS camera except camera is made in large volumes.

Sure Lidar companies want to make money and you don't want to make your own lidar and cameras are cheap and tons of opensource camera code on huggingspace you can use without development makes it a lot cheaper but fundamentally it's a programming mass scale eg guys at facebook/meta doing free work for you that you can reuse vs nobody has a lidar sensor so nobody writes free opensource code for lidar to give away.

I don't really know too much about lidar but again I doubt it's super expensive special process eg standard silicon on 25um process node is good enough.

 

[Crissa]

No, you are completely wrong.

Self driving needs to be based on accurate and real world dimensions. Telsa has decided, and said that their vision based system can make up the difference in accuracy. That's not up for debate.

That's why your argument is a fallacy.

It's not the measurements it's getting wrong.

More absolute data, absolutely equates better outcomes.

No it does not.

You hit a complexity limit with how much information you can process in the time slice required. And some of of this data becomes more complex geometrically in multiple dimensions at once.

There is a thing called 'decision paralysis'. That's what too much information causes.

You can absolutely bake a turkey at 1300 degrees for an hour, if you break it up into pieces and fan it out.

No you can not. To cook a meat, you have a target endpoint - and then that meat gives you a limit on how much the energy will transfer into it without exceeding your target. 1300F for an hour would give you nothing back, no matter how you slice it, because it would burn, dehydrate, and char - and charred flesh is an insulator.

And I have no idea what this analogy was supposed to mean! ?‍?

-Crissa

 

[cardad]

Just FYI, Tesla was Luminar’s largest Lidar customer in Q1. So the bet is that robotaxi will integrate Lidar and everyone with HW3 and HW4 will be like WTF???

https://www.theverge.com/2024/5/7/24151497/tesla-lidar-luminar-elon-musk-sensor-autonomous

 

[SlegMD]

All this talk about sensory limitations. The real limitation is the brain. That will change things.

 

[cyberos]

Just FYI, Tesla was Luminar’s largest Lidar customer in Q1. So the bet is that robotaxi will integrate Lidar and everyone with HW3 and HW4 will be like WTF???

https://www.theverge.com/2024/5/7/24151497/tesla-lidar-luminar-elon-musk-sensor-autonomous

Damn it man! You just ruined the clap back ? I was going to post during this evening’s “We, Robot” Robotaxi reveal ?

 

[Deleted member 17810]

If you need millimeter-level accuracy for your driving, you're doing it wrong

Camera vision isn't getting within inches.

 

[Deleted member 17810]

That's why your argument is a fallacy.

It's not the measurements it's getting wrong.


No it does not.

You hit a complexity limit with how much information you can process in the time slice required. And some of of this data becomes more complex geometrically in multiple dimensions at once.

There is a thing called 'decision paralysis'. That's what too much information causes.


No you can not. To cook a meat, you have a target endpoint - and then that meat gives you a limit on how much the energy will transfer into it without exceeding your target. 1300F for an hour would give you nothing back, no matter how you slice it, because it would burn, dehydrate, and char - and charred flesh is an insulator.

And I have no idea what this analogy was supposed to mean! ?‍?

-Crissa

First of all this cooking analogy got out of hand. But you can bring meat to safe temp depending on the thickness of cuts and cooking vessel at any temperature.

My point is that tesla is throwing away data because it's "good enough" for most things.

Except when it's not and FSD runs into emergency vehicles, or doesn't have object permeance. Something that 4-7 months babies have.

 

[Deleted member 17810]

I can tell you have never done anything remotely related to Ai in a professional setting.

You can add exabytes of data to the equation and it will not help you if the data is dirty. ?‍


Can you explain to me the exact process in which you can bake a turkey at 1300 degrees for an hour and walk away with anything more than charcoal.

I think you are just trolling me here.

It's not quite 1300 but pachamanca typically reaches around 800 degrees in ground cooking, loaded with proteins.

 

[CyberGus]

you can bring meat to safe temp depending on the thickness of cuts and cooking vessel at any temperature.

Thanks for the invite, but I'll be skipping Thanksgiving at your place this year

 

[Deleted member 17810]

Thank for the invite, but I'll be skipping Thanksgiving at your place this year

fun fact: i hate turkey

 

[evnow]

People can close one eye and still throw a basketball or shoot a gun and hit small targets.

Famously, Pataudi, an Indian cricketer lost an eye but he was still able to play international cricket and handle 90 mph bowling, with one eye and went to become a celebrated captain.

Mansoor Ali Khan Pataudi - Wikipedia

 

[cybertruckvegas]

Tesla spent $2m in lidar. Stock drops $10b in a day...

$2m is nothing. Prob equivalent to 1 prototype cybertruck wheel or one AI engineer's salary...

Interesting article saying lidar is $1000 so cheaper than I thought.

Radar prob even cheaper $10. I mean I spent $110k on a truck, $1000 for lidar is better than $2000 for the tent.

It's all about the code to make it useful. They already have radar on older models which is free but basically stopped using it because it cost money to pay coders and train the AI engine vs use free image depth algorithms that for some crazy reason people "open source" for free while paying their software engineers $10m a year: https://www.theinformation.com/arti...-google-promising-recruits-a-quick-stock-bump

Guess Lidar also fugly and reduces range as not aero at all but again it's probably all to do with what code you can get to work and AI camera vision / video is actively developed.

Just FYI, Tesla was Luminar’s largest Lidar customer in Q1. So the bet is that robotaxi will integrate Lidar and everyone with HW3 and HW4 will be like WTF???

https://www.theverge.com/2024/5/7/24151497/tesla-lidar-luminar-elon-musk-sensor-autonomous

Sponsored