The Head of CMU's Robotics Lab Says Self-Driving Cars Are ‘Not Even Close’

While technically correct, calling NREC a CMU robotics lab is really a stretch. They are surprisingly independent entities - I know this as a former student of Carnegie Mellon and employee of NREC. Most people at CMU, including some of the robotics students, had never heard of NREC (although I imagine that's no longer true after Uber).

Anyway, onto the subject at hand. Yep, Herman is right. Everybody thinks it's just around the corner because they see things move so fast in Silicon Valley. But most of the companies that grow exponentially are based on a sound business and relatively (compared to robotics, nuclear fusion, new silicon chip manufacturing) simple technologies. I love robots and I think they will have a profound impact on the world moving forward, but holy fuck it is hard to make a good robot. Self-driving cars are a long way from happening. I think the gains are so great, and the technology so difficult, that they will augment the problem by changing public infrastructure. Making smart traffic lights that can talk to cars wirelessly, and cars that can communicate with each other, and simple road markers specifically for autonomous vehicles, will drastically reduce the complexity of the problem and allow for self-driving cars to happen far sooner (and safer).

This doesn't surprise me based on my own reading. Perception is a very hard problem to crack. Also, even a seemingly simple task like driving involves situations requiring "human-level" decision making and judgement.

Also interesting was this link from the article: http://motherboard.vice.com/read/robots-are-caring-for-elder... The obvious answer to the robots-taking-the-jobs argument is that the world population is aging and there's already a shortage of medical care and nursing home staff.