This is a post-script to my recent post on Eric Schmidt saying that the Turing Test would be passed within five years. An interview with a pioneer in quantum computing suggests that Google just might be hoping to build a human-level AI in that sort of time frame.
Google has recently bought a quantum computer from D-Wave, and during a long but fascinating webcast interview on Singularity 1 on 1, Geordie Rose, founder and Chief Technology Officer at D-Wave Computers, talked about how D-Wave is a key partner in Google’s programme to develop machine intelligence. Rose did not commit to any timelines, but when asked whether 2029 is the date when he expects human-level AI to arrive, he said “machine learning is progressing faster than you think and will become more broadly available on shorter timescales than you might have hoped.”
I remain sceptical, but Google is a company that expects and often delivers remarkable achievements. The worrying thing is that neither Google nor D-Wave seems to be paying much attention to the question of whether the arrival of human-level AI is an unmitigated Good Thing. These are very smart people, but even smart people can get complacent.