We discuss the state of AI, what we have achieved so far and what AI cannot do today. Gary underlines the fact that we are far from so called Artificial General Intelligence (AGI) where computers have cognitive abilities that match the humans.
He argues that Deep Learning alone will most likely not solve the problem and that we need to use tools from other fields within AI and combine them with Deep Learning in order to get closer to AGI.
Gary reasons around the danger of AI, and discusses why he thinks that an AI that could be a direct threat to the human race is not likely to happen in the next decades.
Läs mer: https://ai-podden.se/