3 ms·

Seems like a pretty biased article which seems to largely ignore quickly increasing sensor ability (cameras, pairs of cameras, radar, and lidar). They also ign

by sliken 9y ago

Seems like a pretty biased article which seems to largely ignore quickly increasing sensor ability (cameras, pairs of cameras, radar, and lidar). They also ignore quickly increasing amount of CPU power for vision like tasks (parallel) like the Nvidia TX2. No reason that shouldn't keep improving over the next few years. The article also repeats the popular refrain that lidar is the only reasonable solution. Funny how human eyes (just 2) aren't bad. I've also read reports that lidar is particularly susceptible to things like fog and dust. I can't see why a dozen or so cameras can't do quite well compared to a lessor number of lidar units. Especially if you use pairs of cameras which makes it much easier to pick out 3D features. Human eyes are only separated by a few inches, no reason pairs of cameras couldn't be separated by 6 feet if depth perception is a limiting factor. Machine learning has been increasing quite quickly (things like recognizing human speech, identifying images, computer vision, tracking objects, etc). Current trajectories seem to imply pretty large improvements in the next 3-4 years. Things are somewhat delayed because Tesla just switched from pretty poor sensors to processors to a significantly improved hardware last year. The software is still in the process of catching up. I expect quite nice autonmous driving solutions will be available in the discussed 3-4 year time frame and the roads will be significantly safer because of them. The article also seems to exaggerate the difficulty in testing. Tesla (a fairly small car company) already has quite a few miles under it's belt, even if you ignore all the miles on the previous generation hardware. The previous reported number of dropping airbag deployments by 40% is a huge step in the right direction. With even a pessimistic number of model 3's shipping there will be ample chance to show if tesla's system is better than the average human driver. I expect pretty much all the large car manufacturers to update their current systems that provide predictive braking to include similar capabilities and catch up and surpass Tesla's 100's of millions of miles driven under computer supervision.

frgtpsswrdlame 9y ago
If this article is biased because it shows some pessimism towards self-driving hype, what are the dozens of articles that blindly accept Musks near-term prophecy? Here's another good source for some pessimism: http://www.nakedcapitalism.com/2016/10/self-driving-cars-how-badly-is-the-technology-hyped.html http://www.nakedcapitalism.com/2016/10/self-driving-cars-how... Here's the best parts: Contrary to Musk and many of the most prominent advocates of autonomous cars, Shladover insists that so-called Level 5 vehicles—robocars that require no human input—are not on the horizon. “I tell adult audiences not to expect it in their lifetimes. And I say the same thing to students,” he says. “Merely dealing with lighting conditions, weather conditions, and traffic conditions is immensely complicated. The software requirements are extremely daunting. Nobody even has the ability to verify and validate the software. I estimate that the challenge of fully automated cars is 10 orders of magnitude more complicated than [fully automated] commercial aviation.” Herman Herman, director of the Carnegie-Mellon University Robotics Institute, disagrees as well: "With autonomous cars, you see these videos from Google and Uber showing a car driving around, but people have not taken it past 80 percent. It’s one of those problems where it’s easy to get to the first 80 percent, but it’s incredibly difficult to solve the last 20 percent. If you have a good GPS, nicely marked roads like in California, and nice weather without snow or rain, it’s actually not that hard. But guess what? To solve the real problem, for you or me to buy a car that can drive autonomously from point A to point B—it’s not even close. There are fundamental problems that need to be solved."
- rwmj 9y ago
  But a car that solves the 80% (say it only works on sunny days) is still incredibly useful, and I'd buy one. If it only halves the number of times I have to drive, it's providing real value.
- sliken 9y ago
  Can you really take an article claiming flying is 10 orders of magnitude harder than driving? Because flying hugely more complicated machines in 3D at speeds 10 times those of driving is so simple? 2^10th easier? Or was it 10^10th? Wow, if an 18 year old can be safe enough to drive, that should make piloting easy at what 3 years old?

by sliken 9y ago

frgtpsswrdlame 9y ago

If this article is biased because it shows some pessimism towards self-driving hype, what are the dozens of articles that blindly accept Musks near-term prophecy? Here's another good source for some pessimism: http://www.nakedcapitalism.com/2016/10/self-driving-cars-how-badly-is-the-technology-hyped.html http://www.nakedcapitalism.com/2016/10/self-driving-cars-how... Here's the best parts: Contrary to Musk and many of the most prominent advocates of autonomous cars, Shladover insists that so-called Level 5 vehicles—robocars that require no human input—are not on the horizon. “I tell adult audiences not to expect it in their lifetimes. And I say the same thing to students,” he says. “Merely dealing with lighting conditions, weather conditions, and traffic conditions is immensely complicated. The software requirements are extremely daunting. Nobody even has the ability to verify and validate the software. I estimate that the challenge of fully automated cars is 10 orders of magnitude more complicated than [fully automated] commercial aviation.” Herman Herman, director of the Carnegie-Mellon University Robotics Institute, disagrees as well: "With autonomous cars, you see these videos from Google and Uber showing a car driving around, but people have not taken it past 80 percent. It’s one of those problems where it’s easy to get to the first 80 percent, but it’s incredibly difficult to solve the last 20 percent. If you have a good GPS, nicely marked roads like in California, and nice weather without snow or rain, it’s actually not that hard. But guess what? To solve the real problem, for you or me to buy a car that can drive autonomously from point A to point B—it’s not even close. There are fundamental problems that need to be solved."

rwmj 9y ago
But a car that solves the 80% (say it only works on sunny days) is still incredibly useful, and I'd buy one. If it only halves the number of times I have to drive, it's providing real value.
sliken 9y ago
Can you really take an article claiming flying is 10 orders of magnitude harder than driving? Because flying hugely more complicated machines in 3D at speeds 10 times those of driving is so simple? 2^10th easier? Or was it 10^10th? Wow, if an 18 year old can be safe enough to drive, that should make piloting easy at what 3 years old?