5 ms·

There is no good reason to think pre-programmed computers are going to outperform human drivers when it counts. It's a prediction about the future, someting tha

There is no good reason to think pre-programmed computers are going to outperform human drivers when it counts. It's a prediction about the future, someting that has not happened yet. It's most likely wrong. No amount of marketing hype will change that.

Wowfunhappy 7y ago
I can think of lots of theoretical reasons computers might outperform human drivers. A computer won't get distracted, will have nearly instantaneous reaction times, and can literally have eyes in the back of its head. I don't know if that will be enough to make the difference, but I can certainly see how it could be.
- catalogia 7y ago
  > "might" You're still making a prediction about the future. Some predictions are safer than others: I predict I will see the sun rise tomorrow. Some predictions have little rational basis in reality: I predict you will agree with me. Your prediction about automomous cars being safer than human drivers? That falls somewhere in the middle. From my perspective, all computer vision systems are abysmal crap and spend orders of magnitude more power to accomplish results many orders of magnitude less impressive than the human brain is capable of. The human brain uses something like 10-20 watts, how much CV/ML can you do with 10-20 watts? Not much at all. There is clearly a huge disconnect between our current approach to solving the problem technologically and how the evolutionary wetware solves the problem. That gap may be bridged sometime in the future, but I'm not holding my breath. I think the problem is being approached from the wrong direction, like trying to drive a screw with a hammer. Maybe we'll still get the job done, but I do not consider that a foregone conclusion.
  - Wowfunhappy 7y ago
    Let's let all the companies in this space finish their work. Once they do, we can look at the data and decide what should be allowed on the roads. I just hope that, if the time comes, the decision is based on comparative improvement, instead of hysterics about whether self driving cars will ever kill anyone ever.
    - catalogia 7y ago
      By all means, they should be free to spend their money on whatever speculative research they want if they do it safely (it's easy to test cars on closed circuits.)
    - agentdrtran 7y ago
      What would you say is "finished", even?
  - Dylan16807 7y ago
    * Counting watts is probably the weakest argument for whether the tech is doing something well or badly. * A car can spare 100x that many watts. * Why would we even want cars to calculate the way humans do? It's a very clearly flawed method.
    - catalogia 7y ago
      The point is not whether we can afford to power the machine. The point is that the power discrepancy suggests that the methods we're using are radically different than the methods the human brain employs (evidenced by their dramatically different efficiencies). Predicting radical advancement given past performance is a leap of faith. > Why would we even want cars to calculate the way humans do? The fact of the matter is that humans are orders of magnitude better at object recognition than computers. You may believe that will change in the near future, but that's not yet been demonstrated.
  - millettjon 7y ago
    Can the human brain see through fog? Can it observe in 360 degrees in real time?
    - jakeogh 7y ago
      Human sensory augmentation will save lives.
ben_w 7y ago
Would you describe a neural network as “pre-programmed”? What if it’s updated by every vehicle constantly comparing its own predictions with reality? Don’t get me wrong, I am aware that current generation AI is fundamentally worse than a human brain, because if it was already as good as a human then the fact Tesla’s AI has done more miles than the average person has had heartbeats would make it essentially perfect. It isn’t perfect, so I know the AI is flawed. But my question is, does it seem to you that this approach cannot ever become the equal of the best human (if that best human had 360 degree vision, sonar, and radar, and their nerve impulse conduction speed was way faster), and if not, why not?
jsight 7y ago
I personally know of a recent fatal that likely wouldn't have happened even with the (garbage) ADAS in my current car. There are plenty of good reasons to think that pre-programmed computers can outperform human drivers in the cases that lead to many of our current fatal accidents. I take those as being the ones that really count.

Wowfunhappy 7y ago

I can think of lots of theoretical reasons computers might outperform human drivers. A computer won't get distracted, will have nearly instantaneous reaction times, and can literally have eyes in the back of its head. I don't know if that will be enough to make the difference, but I can certainly see how it could be.

catalogia 7y ago
> "might" You're still making a prediction about the future. Some predictions are safer than others: I predict I will see the sun rise tomorrow. Some predictions have little rational basis in reality: I predict you will agree with me. Your prediction about automomous cars being safer than human drivers? That falls somewhere in the middle. From my perspective, all computer vision systems are abysmal crap and spend orders of magnitude more power to accomplish results many orders of magnitude less impressive than the human brain is capable of. The human brain uses something like 10-20 watts, how much CV/ML can you do with 10-20 watts? Not much at all. There is clearly a huge disconnect between our current approach to solving the problem technologically and how the evolutionary wetware solves the problem. That gap may be bridged sometime in the future, but I'm not holding my breath. I think the problem is being approached from the wrong direction, like trying to drive a screw with a hammer. Maybe we'll still get the job done, but I do not consider that a foregone conclusion.
- Wowfunhappy 7y ago
  Let's let all the companies in this space finish their work. Once they do, we can look at the data and decide what should be allowed on the roads. I just hope that, if the time comes, the decision is based on comparative improvement, instead of hysterics about whether self driving cars will ever kill anyone ever.
  - catalogia 7y ago
    By all means, they should be free to spend their money on whatever speculative research they want if they do it safely (it's easy to test cars on closed circuits.)
  - agentdrtran 7y ago
    What would you say is "finished", even?
- Dylan16807 7y ago
  * Counting watts is probably the weakest argument for whether the tech is doing something well or badly. * A car can spare 100x that many watts. * Why would we even want cars to calculate the way humans do? It's a very clearly flawed method.
  - catalogia 7y ago
    The point is not whether we can afford to power the machine. The point is that the power discrepancy suggests that the methods we're using are radically different than the methods the human brain employs (evidenced by their dramatically different efficiencies). Predicting radical advancement given past performance is a leap of faith. > Why would we even want cars to calculate the way humans do? The fact of the matter is that humans are orders of magnitude better at object recognition than computers. You may believe that will change in the near future, but that's not yet been demonstrated.
- millettjon 7y ago
  Can the human brain see through fog? Can it observe in 360 degrees in real time?
  - jakeogh 7y ago
    Human sensory augmentation will save lives.

ben_w 7y ago

Would you describe a neural network as “pre-programmed”? What if it’s updated by every vehicle constantly comparing its own predictions with reality? Don’t get me wrong, I am aware that current generation AI is fundamentally worse than a human brain, because if it was already as good as a human then the fact Tesla’s AI has done more miles than the average person has had heartbeats would make it essentially perfect. It isn’t perfect, so I know the AI is flawed. But my question is, does it seem to you that this approach cannot ever become the equal of the best human (if that best human had 360 degree vision, sonar, and radar, and their nerve impulse conduction speed was way faster), and if not, why not?

jsight 7y ago

I personally know of a recent fatal that likely wouldn't have happened even with the (garbage) ADAS in my current car. There are plenty of good reasons to think that pre-programmed computers can outperform human drivers in the cases that lead to many of our current fatal accidents. I take those as being the ones that really count.