Nick Bostrom starts his article pessimistically by saying that scientific progress has been slow since the advent of the computer in regards to artificial intelligence. He surmises that an AI that matches human intelligence may be beyond what the human mind can comprehend. He adds that this might be a good thing, as plenty of people have predicted that artificial intelligence surpassing our own could spell the end for humanity. Regardless of either of these points of view, the quest to improve AI should never be relented, according to Bostrom. He even claims that, if humanity applies itself, the task could be achieved in the next 50 years. The technology required for these machines to be created already exists. The problem lies in human thinking on the subject, as well as the necessary hardware to house whatever software we can come up with. However, while we wait for technology to become more efficient, we can continue to apply ourselves and think of new ways to improve AI using current technology as well as anticipating new ones.
In regards to this topic, I am in the camp that believes that truly human artificial intelligence is impossible. I have no doubt that there have been, are, and will always be computers that can do things 100 times better than any human can. However, for a computer to reach a level of self-awareness and sentience necessary for true intelligence is just the stuff of science fiction. Which is good, because sci-fi with sentient robots is some of my favourite!