Live data from Hacker News

Transitioning to Tesla Vision

tesla.com

51–60 of 62 posts

Re: Transitioning to Tesla Vision

#51
post #31

It's incredible that a company as irresponsible as Tesla is allowed to operate vehicles on public roadways. "Well, we dropped a bunch of safety features, we're going to try to add them back via software update in the near future" is not the behavior of a company that should be entrusted with public safety. Combine this with the fact that this is a choice to deliberately cripple the vehicle by removing a standard driv…

Claiming that Tesla is an irresponsible company is pretty strong in the absence of evidence. Do you have evidence that I can review to see in what ways Tesla vehicles are more dangerous than other vehicles? From what I can find Tesla vehicles are significantly safer in terms of accidents per mile, deaths per mile, and they have the highest safety rating of any vehicle. Of course I could be misled and you have evidenc…

It's irresponsible to heavily promote "full self driving" when the reality is that the cars don't have that capability.

The advertising works and some people truly believe in this supposed "full self driving". They start saying things like "Coming home from LA after work thank god for self drive" and "What would I do without my full self driving tesla after a long day at work".

And the result is: https://insideevs.com/news/507038/nhtsa-investigate-tesla-cr...

Re: Transitioning to Tesla Vision

#52
post #38

It's incredible that a company as irresponsible as Tesla is allowed to operate vehicles on public roadways. "Well, we dropped a bunch of safety features, we're going to try to add them back via software update in the near future" is not the behavior of a company that should be entrusted with public safety. Combine this with the fact that this is a choice to deliberately cripple the vehicle by removing a standard driv…

Someone is against the future here.

The future is coming, but Elon Musk will not be the guy who ushers it in.

For one, this is just a dumb decision: It both means a crippled safety package today, and one less sensor one can use tomorrow. Considering Tesla hasn't actually developed self-driving capability, they can't actually say for certain if the hardware that is on the cars today is enough or not. (Any time they promise you've purchased a car that has everything it needs to do this, Tesla is lying, because they can't know the requirements of a system they haven't yet produced.) It certainly is only going backwards on the likelihood that there is once they start removing sensors.

Re: Transitioning to Tesla Vision

#53
Tesla announced, or leaked, last year that they were transitioning from an older Continental radar to one from Arbe Robotics in Israel.[1] Apparently that's not happening.

Arbe claims to have a much higher resolution radar than anybody else in automotive. That's a significant step forward. About 10x less resolution than LIDAR, but better than most radar systems. Good enough to see humans, they claim. You can't actually order the thing, which would be a useful component for many robotic systems, and there don't seem to be third party reports on it.

Their web site is all about "partnerships" and "announcements". They're currently about to go public through a reverse merger with an SPAC from Texas.[2] They've never publicly stated that they had a deal with Tesla. They have a deal with AutoX, which builds self-driving taxis in China, and some kind of arrangement with NVidia. So some real companies think they're real.

[1] https://electrek.co/2020/10/22/tesla-4d-radar-twice-range-se...

[2] https://www.sec.gov/edgar/browse/?CIK=0001816696

Re: Transitioning to Tesla Vision

#54
post #31

Earlier quoted context omitted.

Claiming that Tesla is an irresponsible company is pretty strong in the absence of evidence. Do you have evidence that I can review to see in what ways Tesla vehicles are more dangerous than other vehicles? From what I can find Tesla vehicles are significantly safer in terms of accidents per mile, deaths per mile, and they have the highest safety rating of any vehicle. Of course I could be misled and you have evidenc…

It's irresponsible to heavily promote "full self driving" when the reality is that the cars don't have that capability. The advertising works and some people truly believe in this supposed "full self driving". They start saying things like "Coming home from LA after work thank god for self drive" and "What would I do without my full self driving tesla after a long day at work". And the result is: https://insideevs.co…

It was a TikTok accident, if you want them to stop happening without bannking TikTok, you need to ban knives:

https://www.indiatoday.in/trending-news/story/tiktok-acciden...

Re: Transitioning to Tesla Vision

#55

Earlier quoted context omitted.

It's irresponsible to heavily promote "full self driving" when the reality is that the cars don't have that capability. The advertising works and some people truly believe in this supposed "full self driving". They start saying things like "Coming home from LA after work thank god for self drive" and "What would I do without my full self driving tesla after a long day at work". And the result is: https://insideevs.co…

It was a TikTok accident, if you want them to stop happening without bannking TikTok, you need to ban knives: https://www.indiatoday.in/trending-news/story/tiktok-acciden...

No, it was overconfidence in "full self driving".

Musk says this with a megaphone to his 55.9 million followers: https://www.theverge.com/2019/4/22/18510828/tesla-elon-musk-...

But Tesla privately tells regulators that "full self driving" is only level 2: https://www.thedrive.com/tech/39647/tesla-admits-current-ful...

And Tesla privately says that Musk's comments don't match engineering reality: https://www.theverge.com/2021/5/7/22424592/tesla-elon-musk-a...

You can't have it both ways.

The most responsible thing Musk can do is to publicly admit that "full self driving" is no such thing and change the name to something more realistic.

Re: Transitioning to Tesla Vision

#56
post #31

It's incredible that a company as irresponsible as Tesla is allowed to operate vehicles on public roadways. "Well, we dropped a bunch of safety features, we're going to try to add them back via software update in the near future" is not the behavior of a company that should be entrusted with public safety. Combine this with the fact that this is a choice to deliberately cripple the vehicle by removing a standard driv…

Claiming that Tesla is an irresponsible company is pretty strong in the absence of evidence. Do you have evidence that I can review to see in what ways Tesla vehicles are more dangerous than other vehicles? From what I can find Tesla vehicles are significantly safer in terms of accidents per mile, deaths per mile, and they have the highest safety rating of any vehicle. Of course I could be misled and you have evidenc…

> Claiming that Tesla is an irresponsible company is pretty strong in the absence of evidence.

There is evidence though. Look at the beginning of this Tesla marketing video (in particular the "person in the driver's seat is only there for legal reasons" quote):

https://vimeo.com/192179726

Re: Transitioning to Tesla Vision

#57
post #39

Earlier quoted context omitted.

Yes, Tesla's "Full Self-Driving" is not autonomous enough to operate without you paying attention to the road at all times. But they advertise it as if you can sleep in the driver's seat (I think they literally have said that in the past).

lol they absolutely do not. where do you get this stuff?

https://www.wired.com/story/elon-musk-tesla-full-self-drivin...

>The CEO says his Autopilot system will be "feature-complete" this year, and ready to ferry snoozing passengers by the end of next year.

Re: Transitioning to Tesla Vision

#58
post #27

I'm curious if any AI/vision researchers without a dog in the fight can share their thoughts if Tesla has any chance of a pure vision system where every other company I'm aware of is decking out their cars with LIDAR? My personal impression is that given how cost prohibitive it is of putting LIDAR in every Tesla to their business model, they're relying on safety assist features with the hopes that ~~eventually~~ data…

My highly speculative hypothesis: Tesla doesn't actually plan to sell cars capable of full FSD, at least in the near term. Partially to keep costs low and partially to keep the gate open for operating a fleet of self driving cars that are capable of full FSD.

Waymo so far seems to be inclined towards maintaining a fleet of autonomous vehicles for rides, deliveries and trucking. I suspect it's easier to justify higher hardware costs per vehicle in that case.

Re: Transitioning to Tesla Vision

#59
post #27

I'm curious if any AI/vision researchers without a dog in the fight can share their thoughts if Tesla has any chance of a pure vision system where every other company I'm aware of is decking out their cars with LIDAR? My personal impression is that given how cost prohibitive it is of putting LIDAR in every Tesla to their business model, they're relying on safety assist features with the hopes that ~~eventually~~ data…

Consider this more a wild speculative rambling than actual thorough evaluation, so take that grain of salt (I do work in AI/vision research):

The probability that they can safely pull this off is very low. Right now progress in vision-based ML is incredible. If you look at AK's twitter feed [1], you'll be amazed almost every day. But I work in planning, and for planning detecting / segmenting objects, even recreating 3d models is not enough, you need a full contextual model of the world. Right now, I have 0 clue how this could be done, I don't think anyone else does either. Here's an example of what I mean: you see an image, and can say 'there's a person there' - that's detection. you see an image, and can sort of imagine the person's shape from another viewpoint - that's 3d reconstruction / completion. you see an image, and can imagine the person moving, obeying social and physical laws, waiting to cross a street, look at a car and decide to wait. You know trains will move on rails, trees will stay still unless cut, etc. You're mixing physics engine, statistical inference, heuristics in your head, and this gives you a full, contextual understanding of the world.

I think to get there, one approach is to focus on prediction ability: create a model that learns to 'see', then 'predict' and train both at once (see for example World Models [2]). This prediction task alone was also enough for GPT-3 to seemingly learn really good models of the world. Then the hope is, if you can predict the future really, really well, somewhere in you must be a really perception / simulation / contextual understanding model. In public research we are so incredibly far from being there, prediction ([5], [7]) and prediction-based planning works (eg [3], [4], [6]) are still in the proof-of-concept stage - we're nowhere near predicting high-res images or 3d models of the world with any accuracy. Still, I do see some chance of a GPT-scale endeavor by a well-funded group like tesla leading to some big leap forward. But I'm not holding my breath. To make things bleaker, some think prediction is not enough, and you actually need embodiment (skin-in-the-game), to truly develop good contextual models (I used to be in this camp, GPT made me reconsider). This is even further away, adding hardware, safe exploration and co. is just not something we're in any way close to. Another common thought is that there's a limit to current NN architectures, and that a drastic step is needed to get to the next level, though no one knows exactly what (moving away from feed-forward architectures and backprop, towards something like spike-time-dependent plasticity? higher level meta network coordinators, that decide when to connect trained models to each other's inputs and outputs, when to start and stop modifying their weights, etc, continually?)

You could argue that it might be possible to skip this entirely, and just focus on training an algorithm to be good at driving from start to finish, but there I'd say that we are even further from having the approaches to reach that goal. I'll leave that topic for someone else.

And if you say you don't need this, then you better have some magical over-engineered planning method that can perfectly account for all possible errors / missing info in your perception pipeline. As far as I'm aware, Tesla has no edge there. I'd expect that's the approach other big manufacturers were throwing millions at for all these years, before ML suddenly became more than just edge detection / segmentation.

[1]: https://twitter.com/ak92501?ref_src=twsrc%5Egoogle%7Ctwcamp%... [2]: https://worldmodels.github.io/ [3]: https://www.youtube.com/watch?v=w6TLbv_54GY [4]: https://arxiv.org/pdf/2012.04406.pdf [5]: https://arxiv.org/pdf/1808.06601.pdf [6]: https://arxiv.org/pdf/1812.00568.pdf [7]: https://www.youtube.com/watch?v=3UZzu4UQLcI

Note: I do think they could pull FSD on motorways in a realistic amount of time, though, because there's a lot less variation there. You could even add a little bit of external control, like road-side accident detection cameras, full walls to avoid animals / objects going in and out, etc.

Re: Transitioning to Tesla Vision

#60
post #3

> For a short period during this transition, cars with Tesla Vision may be delivered with some features temporarily limited or inactive I don't understand this, if its only a few weeks why not just wait? Tesla has been years late on this in the past so it's not very reassuring.

I assume that ‘short period’ here is relative to the age of the universe.
Post reply on HN