4 ms·

I'm especially excited for their Google X projects[1] - Google Glass 2, Google X Display, and self-driving cars. If reports[2] are true that the self-driving c

by xasos 12y ago

I'm especially excited for their Google X projects[1] - Google Glass 2, Google X Display, and self-driving cars. If reports[2] are true that the self-driving car will start testing next month, I predict 2015 may be the year that autonomous vehicles go big. The true goals of Lyft and Uber will finally be accomplished. I hope there won't be much red tape getting these cars into action (although there probably will be). [1] http://arstechnica.com/gadgets/2014/12/google-tracker-2015-everything-google-is-working-on-for-the-new-year/6/#h1 http://arstechnica.com/gadgets/2014/12/google-tracker-2015-e... [2] http://www.wired.com/2014/12/google-self-driving-car-prototype-2/ http://www.wired.com/2014/12/google-self-driving-car-prototy...

ams6110 12y ago
I think you're at least a decade, more likely two or three, on the optimistic side. But we'll see.
- eru 12y ago
  Do you think the technological progress will take that long, or do we need the time for the regulation to catch up?
  - objclxt 12y ago
    Both. It's clear right now that self driving cars can perform very well in ideal conditions. It is considerably less clear that they perform well in poor conditions. I haven't seen any demos of Google's tech in low visibility / high rain, and demos of other tech that I have seen went poorly.
  - ghaff 12y ago
    Both, also. From a technology perspective, I think a lot of people are greatly underestimating the gap between autonomous vehicles operating under relatively controlled conditions with a human available to take over control within a modest period of time and the general case. I.e., it's easy for me to imagine autonomous cars being possible for highway driving in good weather in the relatively near term. But, unfortunately, I suspect robo-taxis are decades away. And regulations will be slow to change as well, especially to the degree that autonomous vehicles are seen as toys for rich people. One of the issues will be that, even if they really are statistically safer overall than human-driven cars, they will tend to fail in ways that are different. And you can be sure that those different failures will be highly publicized.
nl 12y ago
I predict 2015 may be the year that autonomous vehicles go big. I'd be interested to hear what you define "go big" as. It's going to be years before driverless cars are seen at all regularly outside the areas Google (and others) are testing them in, and at least a decade before they are common enough to start having a real impact on the road. The economics of car buying alone will show you that: eg, how many 5 year-old cars do you see on the road? And it'll be at least 5 years before you can buy a driverless car, right?. I hope there won't be much red tape getting these cars into action Really?!?!?! I had to pass a driving test to be allowed to drive. I hope something similar happens for driverless car implementations. While I'd trust a Google-coded car, I'm not at all sure I'd trust one coded by <insert a company with less experience at building high-reliability machine-learning software>.
- lozf 12y ago
  Self driving cars just have to be safer than Humans -- probably not that difficult. The annoying part will be entering our destination in to Google Maps Car Interface, and having the car tell us: "Because we searched for Chairs the other day, we'll be making a quick stop at our Sponsor: Furniture Warehouse to see their fantastic selection, with free delivery... After all your appointment isn't for another 45 mins and traffic is clear... (you may override this action at a cost of ${LostAdRevenue}"
  - nl 12y ago
    Self driving cars just have to be safer than Humans -- probably not that difficult. Perhaps, on average. But (for example) should someone in the Netherlands trust car trained on US road rules? Or someone in Japan (where they drive on the left hand side of the road) trust software trained in India (where lane discipline is fairly.. loose). Also, saying a system should be permitted because it kills less than humans seems a bit post-hoc. What happens if a bug is introduced that makes it worse in a specific circumstance? Licensing seems a good way to deal with this, though I'm interested in alternatives.
  - DanBC 12y ago
    > Self driving cars just have to be safer than Humans -- probably not that difficult. Sadly, this probably is not true. TERRORISTS are the reason we go through ridiculous security theatre at airports - costig many millions of dollars - when the drive to the airport is very much more risky that flying. People are not rational about risk. They will drive a car because they thinkthey are safe drivers. Giving up control is tricky.

by xasos 12y ago

ams6110 12y ago

I think you're at least a decade, more likely two or three, on the optimistic side. But we'll see.

eru 12y ago
Do you think the technological progress will take that long, or do we need the time for the regulation to catch up?
- objclxt 12y ago
  Both. It's clear right now that self driving cars can perform very well in ideal conditions. It is considerably less clear that they perform well in poor conditions. I haven't seen any demos of Google's tech in low visibility / high rain, and demos of other tech that I have seen went poorly.
- ghaff 12y ago
  Both, also. From a technology perspective, I think a lot of people are greatly underestimating the gap between autonomous vehicles operating under relatively controlled conditions with a human available to take over control within a modest period of time and the general case. I.e., it's easy for me to imagine autonomous cars being possible for highway driving in good weather in the relatively near term. But, unfortunately, I suspect robo-taxis are decades away. And regulations will be slow to change as well, especially to the degree that autonomous vehicles are seen as toys for rich people. One of the issues will be that, even if they really are statistically safer overall than human-driven cars, they will tend to fail in ways that are different. And you can be sure that those different failures will be highly publicized.

nl 12y ago

I predict 2015 may be the year that autonomous vehicles go big. I'd be interested to hear what you define "go big" as. It's going to be years before driverless cars are seen at all regularly outside the areas Google (and others) are testing them in, and at least a decade before they are common enough to start having a real impact on the road. The economics of car buying alone will show you that: eg, how many 5 year-old cars do you see on the road? And it'll be at least 5 years before you can buy a driverless car, right?. I hope there won't be much red tape getting these cars into action Really?!?!?! I had to pass a driving test to be allowed to drive. I hope something similar happens for driverless car implementations. While I'd trust a Google-coded car, I'm not at all sure I'd trust one coded by <insert a company with less experience at building high-reliability machine-learning software>.

lozf 12y ago
Self driving cars just have to be safer than Humans -- probably not that difficult. The annoying part will be entering our destination in to Google Maps Car Interface, and having the car tell us: "Because we searched for Chairs the other day, we'll be making a quick stop at our Sponsor: Furniture Warehouse to see their fantastic selection, with free delivery... After all your appointment isn't for another 45 mins and traffic is clear... (you may override this action at a cost of ${LostAdRevenue}"
- nl 12y ago
  Self driving cars just have to be safer than Humans -- probably not that difficult. Perhaps, on average. But (for example) should someone in the Netherlands trust car trained on US road rules? Or someone in Japan (where they drive on the left hand side of the road) trust software trained in India (where lane discipline is fairly.. loose). Also, saying a system should be permitted because it kills less than humans seems a bit post-hoc. What happens if a bug is introduced that makes it worse in a specific circumstance? Licensing seems a good way to deal with this, though I'm interested in alternatives.
- DanBC 12y ago
  > Self driving cars just have to be safer than Humans -- probably not that difficult. Sadly, this probably is not true. TERRORISTS are the reason we go through ridiculous security theatre at airports - costig many millions of dollars - when the drive to the airport is very much more risky that flying. People are not rational about risk. They will drive a car because they thinkthey are safe drivers. Giving up control is tricky.