4 ms·

IMO, the idea that fully self-driving cars are anywhere close to being created was always just a combination of hype and wildly irrational underestimation of ho

IMO, the idea that fully self-driving cars are anywhere close to being created was always just a combination of hype and wildly irrational underestimation of how hard it is to create them. The challenge of building a road-safe fully self-driving car is not far from the challenge of building artificial general intelligence. I think that road-ready fully self-driving cars are many decades away, and they may never actually be created. What really surprises me is how many otherwise savvy and tech-knowledgeable people bought into the hype. That said, I also think that not everyone involved in spreading the hype does it for innocent reasons. For example, people often criticize Elon Musk for making absurd claims about self-driving cars, but then argue that he does it because he is innocently deluded. I suspect it's more likely that he is lying through his teeth.

ghaff 7y ago
IMO, it was a combination of (to be relatively generous): 1. General AI optimism fueled by the almost incredibly fast rate of advance of progress enabled by supervised learning in certain domains and by reinforcement learning in an even narrower set of domains. 2. Adding to the above, progress enabled by super-precise mapping within constrained areas. 3. Probably some California-centrism. What's this weather thing of which you speak? 4. Silicon Valley techno-optimism that problems just need one more version to solve but 5. A great big game of topper in which, once various people started proclaiming that self-driving was just around the corner for whatever combination of over-optimism and because it was in their financial interest to do so, YOU had to fall in line or you'd be perceived as being hopelessly behind your competitors when you said self-driving was decades out. Further fueled by 6. The many young professional urbanites with no current interest in owning a car or even getting a driver's license who so desperately wanted to believe that their personal robotic chauffeur that they believed things that they never would have were they not otherwise so personally invested.
_bxg1 7y ago
It happened because the first 95% is tractable. Waymo probably got that far. But there's this infinitely-long tail of edge cases, and 95% isn't good enough when it's literally a matter of life-and-death.

ghaff 7y ago

IMO, it was a combination of (to be relatively generous): 1. General AI optimism fueled by the almost incredibly fast rate of advance of progress enabled by supervised learning in certain domains and by reinforcement learning in an even narrower set of domains. 2. Adding to the above, progress enabled by super-precise mapping within constrained areas. 3. Probably some California-centrism. What's this weather thing of which you speak? 4. Silicon Valley techno-optimism that problems just need one more version to solve but 5. A great big game of topper in which, once various people started proclaiming that self-driving was just around the corner for whatever combination of over-optimism and because it was in their financial interest to do so, YOU had to fall in line or you'd be perceived as being hopelessly behind your competitors when you said self-driving was decades out. Further fueled by 6. The many young professional urbanites with no current interest in owning a car or even getting a driver's license who so desperately wanted to believe that their personal robotic chauffeur that they believed things that they never would have were they not otherwise so personally invested.

_bxg1 7y ago

It happened because the first 95% is tractable. Waymo probably got that far. But there's this infinitely-long tail of edge cases, and 95% isn't good enough when it's literally a matter of life-and-death.