Robotics and automation have been improving for a long time, and especially recently. Look at the rise in consumer drones, enabled by improvements in batteries, sensors, and computers.
But the main thing holding them back is a lack of AI. Robots can do a rote action over and over again, but they have a hard time identifying where objects are, planning, and reacting to their environment. Just solving machine vision would be a massive step forward and enable a ton of applications.
And that has sort of already happened. The best nets are already exceeding humans at vision tasks. They are learning to play video games at expert level, which is not conceptually distant from robot control. Its taking time to move this research out of the lab and into real applications, but it is happening.
And so I totally believe that at least 50% of current jobs could be automated in 10 to 15 years. How many people are employed doing relatively simple, repetitive tasks, over and over again? Me and most people I know have jobs like that.
But the main thing holding them back is a lack of AI. Robots can do a rote action over and over again, but they have a hard time identifying where objects are, planning, and reacting to their environment. Just solving machine vision would be a massive step forward and enable a ton of applications.
And that has sort of already happened. The best nets are already exceeding humans at vision tasks. They are learning to play video games at expert level, which is not conceptually distant from robot control. Its taking time to move this research out of the lab and into real applications, but it is happening.
And so I totally believe that at least 50% of current jobs could be automated in 10 to 15 years. How many people are employed doing relatively simple, repetitive tasks, over and over again? Me and most people I know have jobs like that.