Back in 1993 Vernor Vinge, the legendary science fiction author and one of the godfathers of the concept of the technological singularity, wrote in his classic essay The Coming Technological Singularity that he would be surprised if the singularity occurred before the year 2005 or after the year 2030. It was a radical prediction at the time and considered to be hopelessly optimistic by all but a few artificial intelligence researchers and futurists.
Now, twenty four years after making his original predictions Vernor Vinge is still sticking with them. In a recent interview with the journal Research-Technology Management he said,
“I would be surprised if it [The technological singularity] hasn’t happened by 2030. I think we are pretty much on schedule for it to happen.” – Vernor Vinge, Research-Technology Management, 2017
Progress in the last 24 years in artificial intelligence research has occurred faster, especially in the last few years, than most AI researchers would have predicted. In the years since 1993, bowling pin after blowing pin argument of the critics of artificial intelligence has been knocked down. In 1997 IBM’s Deep Blue supercomputer defeated the world chess champion, Garry Kasparov, in 2005 Stanley, Stanford University’s self-driving car, won the DARPA Grand Challenge by driving 150 miles across the Mojave Desert without any human intervention, in 2010 deep learning blasted onto the scene by winning the ImageNet challenge. In 2011 IBM’s Watson supercomputer defeated the two best human players at the game of Jeopardy, Ken Jennings and Brad Rutter and in 2016 DeepMind’s AlphaGo defeated Go champion Lee Sedol a full 10 years before experts had predicted it to be possible.
I suspect that with all this recent progress we’ve witnessed in machine learning and the explosion in interest and hype around the field it has given him confidence that his date of 2030 is not unreasonable. They’re are only a handful of AI researchers and futurists who think that greater than human-level artificial general intelligence is likely before 2030, one of them being DeepMind co-founder Shane Legg who has said he believes it will happen sometime in the mid-2020s.
I was at first surprised to see Vernor Vinge sticking with an outer-limit date of 2030, 15 years sooner than Ray Kurzweil’s date of 2045 for the singularity, although I guess I shouldn’t have been considering he was willing to make the prediction in 1993 at a time when it was viewed as outlandish and considering the strides in his favor that have been made since. Still, even as someone who has followed the discussion around the singularity for the last 15 years or so it’s still quite astounding to me to contemplate, even if it’s only a small chance, that in less than 13 years from now the greatest revolution in the history of technology, and biology for that matter, could occur.