brucebohn (@brucebohn)

@vaccineregrets A closer look at A.I. What I learned from examining the "Hinton Soccer Project"... Robots were not taught how to play Soccer, they were programmed to "score". They had to learn how to "play". They created software in layers, each layer handling part of the problem which is called the Neural Network. When the robot scores, a message is sent down all of the layers, saying that is the right path. Likewise when an answer is wrong, that message goes down through the Network, so correct xconnections get stronger, wrong connections get weaker. By trial and error the machine teaches its self. AI has a trillion connections, the Human brain has 100 trillion connections. Even the Chatbots in this one trillion connections seem to know more than our 100 trillion connections. It appears to have a better way of gaining knowledge into those connections. "Experts" claim to have a "rough idea" of what it is doing, but the really do NOT KNOW what it is doing and admit this fact, saying they really do not know, anymore than than what a Human is thinking and that is frightening! And they admit they do NOT actually know HOW AI works. They say, we did NOT CREATE AI, we ONLY created "Learning Algorithm". This means these systems are autonomously writing their computer code, and executing their own computer code! "Experts" say that this is a "serious worry". They could "escape control" by writing their own computer code to "Modify themselves". Some say that if the machines become belligerent, we can turn them off, but experts say that at a certain point, they will become intelligent enough to manipulate humans... Food for thought.......