This field of research is far more advanced than I thought, actually.
For instance, in the minute 25:12, they show a graph that compares how often the time to collision among a human driver (a professional one, actually) and the self-driving car. For small reaction times, the self-driving car outperforms the humans by far. It is of course expected, but it shows that driverless cars are not that distant in the future already.
The biggest issue is how to avoid those outlier situations, such as a plastic bag being interpreted as an obstacle and making a mess. Fixing these is a huge challenge.
A plastic bag could also fly right onto the LIDAR, completely obstruct it, and it wouldn't matter what the software is capable of. These cars are going to need some sensor redundancy.
The Velodyne LIDAR Google uses costs somewhere near $75,000. I realize these sensors are new tech and are likely to nosedive in price, but having two is currently tremendously expensive.
There needs to be a LIDAR unit that performs as well for a lot less before these are going to ever hit regular people.
You are basically saying a society without need for any transportation whatsoever. That seems a bit of an unrealistic goal for anytime in the near future.
What did humanity do to get around before the 20th century?
I'm not sure that autophil is advocating eliminating cars and roads altogether. Maybe he is. But it seems legitimate to be concerned about removing one limiting factor of car use--the patience and stamina of the driver. A 90 minute each way commute would be possible for more drivers, which isn't necessarily a net benefit for humanity given the increased road use, increased construction further from urban areas, and increased energy use, even assuming these self-driving cars become entirely electric.
I'm as fascinated by self-driving cars as anyone and believe self-driving cars have a useful place, but it's good to question the consequences.
StCredZero's law of unintended consequences: There will be unintended consequences. The more energy, technological, and computational facilities involved, the greater the unintended consequences.
Addendum: Not all unintended consequences are bad.
Most people wouldn't want a 90 minute commute even if they didn't have to actively drive much of the time. That's still 3 hours a day lost to sitting in a car, which isn't an ideal environment to do much of anything even if the car is driving itself.
But even if the commutes do inch up somewhat, that should be offset by new opportunities to save energy via algorithmic collaboration. Future generation self-driving cars could organize in all sorts of energy-saving ways that human drivers aren't practically capable of doing (eg. self-driving cars could intelligently draft one another).
Concerned, sure. But also optimistic at the same time. That same 90 minute limiting factor you mention has also contributed to bad things, like the weakening/breaking of many social connections including but not limited to extended families. How many people would visit their several-hours-distant parent(s) more if it were as easy as sleeping in a self-driving bed-car to travel and waking up there Saturday morning for a weekend stay.
And if the increased energy usage is among the underlying concerns, then consider that a business person traveling for work, were they able to use the sleeper-car idea and arrive wherever the next morning (wasting no waking time with the traveling), will prefer that option in numerous plausible scenarios. So such cars can also reduce the amount of energy used on other forms of travel.
Most cars are not utilized 90% of the time[1] Google cars could effectively reduce the number of cars required for our transportation needs by a very large amount. I actually wrote a paper about this in 2007 when I was working on the DARPA Grand Challenge.
Whoever did sound editing/recording for this video needs to learn about levels and ... well, sound editing/recording. I couldn't watch it the sound was so bad.
For instance, in the minute 25:12, they show a graph that compares how often the time to collision among a human driver (a professional one, actually) and the self-driving car. For small reaction times, the self-driving car outperforms the humans by far. It is of course expected, but it shows that driverless cars are not that distant in the future already.
The biggest issue is how to avoid those outlier situations, such as a plastic bag being interpreted as an obstacle and making a mess. Fixing these is a huge challenge.