Rocket's Red Glare: Real intelligence needed to pick NHL playoffs in haiku form

That was the most difficult headline I've ever had to write, not because it pained me but because distilling the things commenter Rocket wrote in his guest post into a smattering of words is not easy. He's a complicated, hirsute man.

The upshot? He saw a movie and he picked the second round of the NHL playoffs in haiku form. Rocket?

————

Not so long ago, I got a hankering to see one of those moving picture stories that all the kids are talking about. Rockette and I took the horseless carriage down to the local movie palace to take in one of these celluloid fantasies. The name of this particular talkie was Ex Machina.

Like any work worthy of the title of "art," the film has stuck with me (this is also the reason that Slap Shot is the greatest film in the history of mankind – nearly forty years after it was made and at least thirty years since I saw it for the first time it still makes me laugh just thinking about it). It will not spoil the film to note that the questions that drive it are when will artificial intelligence be "good enough" and what would happen if we got it wrong?

Naturally the movie addressed the Turing Test and a theoretical capacity of a machine to mimic a human being effectively enough to fool actual humans. The very nature of the Turing Test and the film's treatment of its central conceit left me reflecting on the question of artificial intelligence and at what point it would need to reach to be "good enough."

Eventually I came to a surprising conclusion: artificial intelligence will never be "good enough." This is not because I naively believe that we will never develop the sufficient technology – this is a patently absurd assertion that cannot possibly be made on a blog that you might be reading on your phone. Rather, artificial intelligence will never be good enough because human intelligence has never been good enough.

Put differently, humans are defined by their faults. We recognize that our flaws are the very things that which make us human. The stories that we tell about ourselves and each other are not about the times when things go smoothly or as planned. Rather, we seek to amuse each other with stories of when things went awry or were off-kilter or out of the ordinary. We bond over the tough times or when we had to struggle. And when we run across someone who only wants to talk about their successes, we consider that behavior to be their flaw.

And since we are defined by our faults, any machine that would be capable of passing the Turing Test would also have to be equally flawed. Yet, such a possibility seems utterly and completely horrific. Do we really want robots that are or can become jealous or depressed or schizophrenic or engage in domestic violence or homicide or genocide or any and all of the awful things that human beings continue to do to each other? Because human intelligence is, by its very nature, deeply flawed – because we are not "good enough" – we can never make artificial intelligence "good enough" to pass the Turing Test.

This is, of course, an interesting, thought-provoking, well-written way of noting that I got a few things wrong in my last batch of predictions. While I did get six of the eight series correct, including the Jets getting swept, I erroneously – and to my great regret – picked against the Wild. On occasion, it is really nice to be human.

Here are the second round predictions, in haiku form:

Montreal Canadians v. Tampa Bay Lightning

Everybody knows

Carey Price is MVP

Yet his season ends

New York Rangers v. Washington Capitals

Even Ovechkin

Cannot escape the allure

Of Hank's dreamy eyes

Anaheim Ducks v. Calgary Flames

Unfortunately

The sad ghost of Hakan Loob