Live data from Hacker News

Tesla FSD no longer offered for purchase

notateslaapp.com

71–80 of 82 posts

Re: Tesla FSD no longer offered for purchase

#71
post #27

Earlier quoted context omitted.

The hypothesis was that LIDAR was a crutch, humans manage with just vision. Whether this is correct for delivering self driving cars, we will find out soon enough. Long term though, it definitely makes sense. We just don't know what the missing pieces of the puzzle are.

> humans manage with just vision this is commonly repeated but very obviously untrue. We don't only have vision. We have a general intelligence, coupled with vision. In the absence of AGI, the base assumption has to be the sensor apparatus needs to be significantly superior to humans for an FSD system to drive at a comparable level.

You are right.

Would you agree then, that if the goal was to develop AGI, just relying on vision is a credible choice?

Re: Tesla FSD no longer offered for purchase

#72
post #33
post #27

Earlier quoted context omitted.

The hypothesis was that LIDAR was a crutch, humans manage with just vision. Whether this is correct for delivering self driving cars, we will find out soon enough. Long term though, it definitely makes sense. We just don't know what the missing pieces of the puzzle are.

> humans manage with just vision. But they don't. I can't see how anyone could look at modern driving and see an optimal state. Driving isn't being managed at all, it's killing droves of humans. If we put the same restrictions on airplanes (flying by instrument is a crutch), everyone would rightfully find that ridiculous. They appear to have bet on the wrong technology. The failure happened back in the design phase.

Aircrafts often don't have vision at all, in regular operation.

If a driver doesn't have vision, the right decision is to figure out how to safely stop.

Re: Tesla FSD no longer offered for purchase

#73
post #27

Earlier quoted context omitted.

The hypothesis was that LIDAR was a crutch, humans manage with just vision. Whether this is correct for delivering self driving cars, we will find out soon enough. Long term though, it definitely makes sense. We just don't know what the missing pieces of the puzzle are.

The missing piece may be a different mode of transport like trains. Humans adapted from creatures that lived in trees over millions of years, a computer has nothing on that evolutionary process of the bad tree jumper getting eaten or breaking a leg and dying. Spend a few million years programming a computer to swing through trees and they'll probably get something that can drive a car.

We have close to that much in training data in the form of cars driven by humans.

What we lack is (still) the fundamental algorithms to learn from video. Tokenization like LLMs or diffusion are starting to fall short of this goal.

Re: Tesla FSD no longer offered for purchase

#74
post #49
post #27

Earlier quoted context omitted.

The hypothesis was that LIDAR was a crutch, humans manage with just vision. Whether this is correct for delivering self driving cars, we will find out soon enough. Long term though, it definitely makes sense. We just don't know what the missing pieces of the puzzle are.

The missing piece is LIDAR.

Great, I hope I can someday be this confident about predicting the future!

Re: Tesla FSD no longer offered for purchase

#75
post #70

Earlier quoted context omitted.

Not to mention it is also untrue because we use senses other than just vision when we drive. We use our ears for acceleration information, sometimes hearing, and the feeling of the wheel when we drive.

We don't have anything close to LIDAR though

And a car doesn't have anything close to a human brain.

Humans process sensory data in a fundamentally different way to anything that's possible for a self-driving car. The idea that we should base the decision about the sensors on what humans have just fundamentally makes no sense.

Lidar substitutes hardware for something which humans find easy and CV systems find hard - creating a map of the environment. Humans do that by using a brain. CV systems based purely on video really struggle to do that in lots of edge cases. You can shortcut that in a car by using something like lidar.

Re: Tesla FSD no longer offered for purchase

#76
post #71

Earlier quoted context omitted.

> humans manage with just vision this is commonly repeated but very obviously untrue. We don't only have vision. We have a general intelligence, coupled with vision. In the absence of AGI, the base assumption has to be the sensor apparatus needs to be significantly superior to humans for an FSD system to drive at a comparable level.

You are right. Would you agree then, that if the goal was to develop AGI, just relying on vision is a credible choice?

No. Why should the design parameters for AGI be limited by what a human can do? If the goal was AGI then I'd want all kinds of additional sensor input that humans don't have.

Re: Tesla FSD no longer offered for purchase

#77

Earlier quoted context omitted.

Have you tried it? it's not there but it's close. Even Ford's Blue Cruise is half way there.

the problem is, even if you solve 80% of cases 100% of the time, those 20% of outliers are not easy to get 100%. How should a autonomous car behave when it sees a basketball bounce across the road near a playground? Any human would assume: "there might be kids, I should slow down even if the ball isn't obstructing my path" a car isn't going to be able to make those decisions.

It depends on what you define success. Only Waymo has actual real, live robotaxis right now; Tesla's FSD very much isn't that, despite promises that it'll be here next year for the past 9 years. But in terms of "how much does it suck to drive 2 hours in bumper-to-bumper traffic on the freeway", we've made leaps and bounds of progress since the original Darpa Grand Challenge, 20 years ago in 2004. Yeah I'd love to own a Tesla and rent it out as a robotaxi and have it earn money while I'm not using it. I'd also like a flying car. And a pony. To say that we "don't have anything even remotely close" when we have actually made progress, just because some pie-in-the-sky goal hasn't been meet, is where I take issue with. I don't know how much that last 20% will take, but if we've only made it 80% of the way, that's still not "don't have anything even remotely close" territory.

Re: Tesla FSD no longer offered for purchase

#78
post #23

Earlier quoted context omitted.

Have you tried it? it's not there but it's close. Even Ford's Blue Cruise is half way there.

Like many things in life, "it's close" is not worth celebrating, especially if it's not clear how the gap will be solved.

My point is that, in this case, it is. Totally fine if you disagree, but in terms of how much it sucks to sit in bumper-to-bumper, stop-and-go-traffic, or just any 3 hour drive on the freeway, the SOTA is so much better compared to 20 years ago that it is worth celebrating. Complain about not having robotaxis all you want. What already exists is miles better than OG cruise control from the 90's, which just kept you at a constant speed, and didn't even have radar to automatically slow down if the vehicle in front of you slowed down.

Re: Tesla FSD no longer offered for purchase

#79
post #51

Earlier quoted context omitted.

> something like lvl 3 autonomy before any other company did I'm quite sure Mercedes-Benz was the first to bring lvl 3 autonomy on the market. https://arstechnica.com/cars/2023/09/mercedes-benzs-level-3-... It is also the only carmaker confident enough in the system that it takes full liability over it > Confidence in Drive Pilot is high within Mercedes-Benz, as the system has been active in Germany for over a year w…

According to that article Mercedes-Benz's system is exclusively highway driving. Technically level three, but not "full self driving" as most people would understand it, or as Tesla defines the term.

Level 3 autonomous driving is not FSD.

Re: Tesla FSD no longer offered for purchase

#80
post #71

Earlier quoted context omitted.

You are right. Would you agree then, that if the goal was to develop AGI, just relying on vision is a credible choice?

No. Why should the design parameters for AGI be limited by what a human can do? If the goal was AGI then I'd want all kinds of additional sensor input that humans don't have.

Once it's a solved problem, yes, it makes sense to think about design parameters.

When learning how to solve problems, that is not as helpful.

Post reply on HN