Tesla Autopilot director contradicts Musk’s self-driving timeline(arstechnica.com)
arstechnica.com
Tesla Autopilot director contradicts Musk’s self-driving timeline
https://arstechnica.com/cars/2021/05/tesla-autopilot-director-contradicts-musks-self-driving-timeline/
19 comments
Celebrity worship is that strong.
Musk's law is the observation that Tesla is always two years away from full autonomous driving.
https://twitter.com/elonmusk/status/1387901003664699392
Elon: "A major part of real-world AI has to be solved to make unsupervised, generalized full self-driving work, as the entire road system is designed for biological neural nets with optical imagers"
Seems like he's inching closer to reality by the day.
Elon: "A major part of real-world AI has to be solved to make unsupervised, generalized full self-driving work, as the entire road system is designed for biological neural nets with optical imagers"
Seems like he's inching closer to reality by the day.
Or I could have used better sensors on my car but that doesn't sound as controversial as "anyone using LIDAR is doomed" so I fooled my customers and let my competitors get ahead.
Saying "anyone using LiDAR is doomed" is a controversial statement sure but the self-driving problem is indeed a computer vision problem.
George Hotz puts it quite neatly, companies like Waymo have spent millions so far and someone sitting at home could not use their product highlighting their inefficiency. He further states to think about how a human would drive a car, does a human need LiDAR to drive a car? No, not really. The legitimacy in his statements come from the fact ( imo ) that comma.ai ( his self-driving startup ) outperforms with less funding than one could imagine by solving the problem absolutely as a Computer Vision problem. They treat it slightly different than Tesla but in all, it goes on to show that you don't really need to use LiDAR to solve the problem, it can potentially solve the problem, but it's an inefficient way to do it.
George Hotz puts it quite neatly, companies like Waymo have spent millions so far and someone sitting at home could not use their product highlighting their inefficiency. He further states to think about how a human would drive a car, does a human need LiDAR to drive a car? No, not really. The legitimacy in his statements come from the fact ( imo ) that comma.ai ( his self-driving startup ) outperforms with less funding than one could imagine by solving the problem absolutely as a Computer Vision problem. They treat it slightly different than Tesla but in all, it goes on to show that you don't really need to use LiDAR to solve the problem, it can potentially solve the problem, but it's an inefficient way to do it.
> He further states to think about how a human would drive a car, does a human need LiDAR to drive a car?
I hate this argument. It doesn't hold for several reasons:
- A human driver has blind spots (which btw Tesla avoids by having cameras surrounding the car)
- Human reactions times are very poor at even moderate speeds
- Human vision is affected by glare
- Humans are terrible at driving in low light conditions
- Humans are terrible at driving in inclement weather - snow storms or heavy rainfall
LIDAR doesn't solve many of these problems but to claim that self driving is solvable by vision alone is a bad take. Multiple sensors add redundancies and reduce errors. Teslas are already equipped with ultrasound sensors and radar so it's not technically solving a Computer Vision only problem.
I hate this argument. It doesn't hold for several reasons:
- A human driver has blind spots (which btw Tesla avoids by having cameras surrounding the car)
- Human reactions times are very poor at even moderate speeds
- Human vision is affected by glare
- Humans are terrible at driving in low light conditions
- Humans are terrible at driving in inclement weather - snow storms or heavy rainfall
LIDAR doesn't solve many of these problems but to claim that self driving is solvable by vision alone is a bad take. Multiple sensors add redundancies and reduce errors. Teslas are already equipped with ultrasound sensors and radar so it's not technically solving a Computer Vision only problem.
> LIDAR doesn't solve many of these problems
Radar+sonar do.
Radar+sonar do.
It's never about being just able to solve the problems, it's more about solving the problem without spending millions and millions of dollars and still not having a product everyone can use after years.
Can you give an example of competitors that use Lidar and are further ahead than Tesla?
Doesn’t Waymo use LIDAR? They’re not as widespread as Tesla is, but some suggest they’re better than Teslas.
Furthermore the entire road system is designed to work with simply two rotating cameras primarily facing forward.
Seems to me that maybe smart roads would be better? If there's sensors that communicate back and forth provide additional wifi boosts for cars on the grid, etc... L5 could come via cheating...expecting a car to do everything itself is going to be rough... but when a street can give you data about everything on it -- pedestrians, bikes, etc...it'd be easier to build to l4.9 automation...
Musk needs to team up with the Biden Administration and work in tandem on infrastructure - that may be the only way we see self-driving road-ready cars before 2040.
Musk needs to team up with the Biden Administration and work in tandem on infrastructure - that may be the only way we see self-driving road-ready cars before 2040.
I had a boss who liked to announce various new software features in email updates to clients, but frequently didn't bother to tell the engineering team to implement those features. I got tickets where a customer asked, "how do I make feature X work?" for a feature that didn't actually exist, so the task for the ticket was to create it.
At least Musk is just exaggerating the capability of an actually existing feature. On the other hand none of our missing features misled anyone to think that a dangerous device is safer than it is.
At least Musk is just exaggerating the capability of an actually existing feature. On the other hand none of our missing features misled anyone to think that a dangerous device is safer than it is.
I think we'll crack fusion power before level 5 self-driving cars. And that's been ten years away since 1950. :-)
Not if we had smart infrastructure/roads. That's the only way, Musk needs to bite his pride and team up with Toyota, GM, BMW, VW etc go to D.C. talk to Biden, go pay off McConnel, and get an infrastructure bill that includes all roads in America replaced with smart roads in next 10 years.
Roads that:
- Self-repair via drones / automation
- Self-fix weather conditions like flooding (pumps in locations where flooding is likely?), or ice (lasers that heat melt it, drones that drop salt/sand, self-driving plow or drones that plow and sand, heated roads with deep drains or that slant on either side with a nice drop off for excess fluid to drain.
- Mapping/tracking of EVERY object on the road ..if a deer crosses the highway in Wyoming all cars within a mile are aware of it.
- All cars around you know your speed and can adjust.
- Your license/speed/etc are all known by the grid at all times.
- Break a law no police will pull you over, you'll just be fined. However, speeds can be adjusted or removed in places on highways while in self-driving mode, the car will never go faster than is safe in an area/weather but can drive faster and cars could "link up" i.e. drive exact same speed and move like a "train" to increase aerodynamics.
- You'd be able to drive yourself, but speed limits/etc would still be tracked and dangerous driving would be finable.
Some of these have many ethical / dystopian concerns... frankly the "authorities" always knowing our location/speed, but on the other hand less police stops means less likelihood people are killed for being black/minority... it's a catch 22 on liberty vs equality.
I mean if you haven't recently robbed a bank or something I think there's no problem and it's not like there's not traffic cams everywhere where we're already tracked by, at least this way you can take a road trip and snooze through all the "Dad are we there yet?" complaints lol.
Roads that:
- Self-repair via drones / automation
- Self-fix weather conditions like flooding (pumps in locations where flooding is likely?), or ice (lasers that heat melt it, drones that drop salt/sand, self-driving plow or drones that plow and sand, heated roads with deep drains or that slant on either side with a nice drop off for excess fluid to drain.
- Mapping/tracking of EVERY object on the road ..if a deer crosses the highway in Wyoming all cars within a mile are aware of it.
- All cars around you know your speed and can adjust.
- Your license/speed/etc are all known by the grid at all times.
- Break a law no police will pull you over, you'll just be fined. However, speeds can be adjusted or removed in places on highways while in self-driving mode, the car will never go faster than is safe in an area/weather but can drive faster and cars could "link up" i.e. drive exact same speed and move like a "train" to increase aerodynamics.
- You'd be able to drive yourself, but speed limits/etc would still be tracked and dangerous driving would be finable.
Some of these have many ethical / dystopian concerns... frankly the "authorities" always knowing our location/speed, but on the other hand less police stops means less likelihood people are killed for being black/minority... it's a catch 22 on liberty vs equality.
I mean if you haven't recently robbed a bank or something I think there's no problem and it's not like there's not traffic cams everywhere where we're already tracked by, at least this way you can take a road trip and snooze through all the "Dad are we there yet?" complaints lol.
Pretty sure Rush wrote a song about the future you’re describing.
https://en.m.wikipedia.org/wiki/Red_Barchetta
https://en.m.wikipedia.org/wiki/Red_Barchetta
Tesla seemingly gets off the hook rather easily regarding this stuff anyway, wonder if that's just because they have Musk at the top.