How much smarter is AI than us? The Rubik’s Cube test
Bottom Line: I’ll forever marvel at CBS 12’s Lauren Oleski’s amazing weather forecast. The one in which she delivered the forecast and solved the Rubik’s Cube in under five minutes. Who does that? LOL does, that’s who. And probably AI given the opportunity. If you’re wondering about the power of AI already, and the implications to our society in the future... This story will paint the picture.
Researchers at the University of California took their latest AI creation and ran it through the Rubik’s Cube test. Something that provides a good baseline for comparison to us human types. To give you an idea...the top 1% of Rubik’s cube solvers complete the puzzle in an average of 50 moves. A perfect score is 17 moves. The human world record is just over 7 seconds. As impressive as that sounds get this...
60% of the time the University of California’s AI was perfect. The average length of time to completion,under 1 second. Wow... Right? It’s literally not humanly possible to do what AI is capable of doing already. The question becomes what happens next. The goal for these researchers is to build on the depth of comprehension to create the next round of digital assistants. If they’re successful, you can count on fewer frustrations with Alexa, Cortana and Siri. The question will be whether they’re too smart for their own, or our own, good.