“Intelligence and creativity, it would appear, are not a human monopoly.”
—Alvin Toffler
I have been in the AI field for over 55 years, shortly after it was named. 
Yet it is only in the last two years that we now have enough computational resource that neural nets can do everything that humans can do, where either (i) we have examples of how humans solve problems or (ii) we can simulate the world in which the problem lies. That is basically everything we do.
If we create a graph with the major neural net models since 2012 it looks like my usual models where computational density goes up over time, except that it is not doubling every year (which is normal for chips) but rather it is doubling every 3.5 months. That’s an increase of 300,000 fold in six years.
Until recently, neural nets have not done very well because they were not given sufficient computation. Now they are over the threshold that is sufficient for just about every task we present to them.
When IBM’s Deep Blue won the chess competition in 1997, the super computer at that time was filled with everything we could gather from people about chess. The machine was not useful for anything else. It was a chess machine.
In contrast, when Alpha Go Zero won at Go in 2017 it was not given any information from humans about Go except for the rules and in about three days played itself to be able to defeat every human player. But that was not the most significant thing about Alpha Go Zero. Although it played Go better than any human, its design had nothing to do with Go. When the Go rules were simply changed to chess rules, again, with no human ideas about chess, it was able to create a chess playing program by playing itself, and in only four hours it defeated every human.
Similar neural nets from various companies have exceeded human doctors at radiology. They can also predict how a new linear amino acid sequence will create a three-dimensional protein, crucial for finding medications. Yet another neural net created trillions of chemical compounds that activated the human immune system, while another net determined if each one would be useful against flu and created an optimal flu vaccine.
Ultimately, we will be able to resolve most failures of our bodies using these biological simulators, testing them in hours rather than years. The FDA is actually now accepting simulator results instead of human results in testing new vaccines such as this year’s flu vaccine, since we can’t wait a year or more to approve it.
These successes are coming every week. Well within this decade AI will do everything that any human can do, only much better. We will be able to simulate a human by recreating the thousands of human skills and combining them as another intelligent task.
To pass a Turing test (to determine if an AI will “pass” for a human), we will need to make it less smart, as its capability in any area will be too impressive for an unenhanced human.
It is important to understand that technology is used to go beyond human capabilities. Most professions today would be impossible without the digital enhancements we already have. In the 2030s we will use nanobots to connect our neocortexes to the cloud and thereby enhance our intellectual capability with these major nonbiological enhancements. Our biological intelligence is more or less fixed, whereas our nonbiological intelligence will grow indefinitely.
Ray Kurzweil is one of the world’s leading inventors, thinkers, and futurists, with a thirty-year track record of accurate predictions. Ray was the principal inventor of many firsts, including the CCD flat-bed scanner, omni-font optical character recognition, print-to-speech reading machine for the blind, text-to-speech synthesizer, music synthesizer capable of recreating the grand piano and other orchestral instruments, and commercially marketed large-vocabulary speech recognition. He received a Grammy Award for outstanding achievements in music technology, the National Medal of Technology, was inducted into the National Inventors Hall of Fame, holds twenty-one honorary Doctorates, and honors from three U.S. presidents. Ray has written six national best-selling books, including New York Times best seller The Singularity Is Near (2005). He is Co-Founder and Chancellor of Singularity University and a Director of Engineering at Google.