How should we prepare for the invention of artificial general intelligence?

There has been much debate about the potential risks and benefits of developing Artificial General Intelligence (AGI). Is it as dangerous as people like Stuart Russell, Nick Bostrom and Eliezer Yudkowsky think? Or do you agree with people like Rodney Brooks, Oren Etzioni and Neil Lawrence that there’s no real reason to worry about existential threats? Should we instead be concerned about employment, equality or fairness? Or should we stop focusing so much on potential downsides? Is AGI close or far away? Is it even possible? If it’s potentially dangerous, can we already start to do useful work to mitigate the danger of a technology we don’t yet have or should we wait until we learn more? What do you think are the most promising technologies and policy methods to deal with AGI?