Randall Munroe of xkcd reminds us that we humans are gradually losing our competitive advantage with computers. Our only hope is look at Calvinball as not only a game, but a way of life.
One two-day session with one giganticly powerful supercomputer seems fair enough reason to straddle the line between "can beat top humans" and "still lose to top humans" is fair enough. The designers weren't _confident_ Watson was going to win -- which would seem necessary to me to justify moving the entry towards checkers.
I never did learn how they fed questions to Watson: did it OCR the screen? Did it speech-to-text Trebek? Did they feed it the raw text over a socket? If the latter, then _definitely_ it doesn't deserve to be moved towards "easy".
Umm, In 2011 IBM Watson beat the top 2 humans of all time, very badly. Chart needs updating.
I never did learn how they fed questions to Watson: did it OCR the screen? Did it speech-to-text Trebek? Did they feed it the raw text over a socket? If the latter, then _definitely_ it doesn't deserve to be moved towards "easy".