For the public, A.I. first started gaining recognition after it was tested in the world of chess.
Chess was the perfect testing ground for A.I… because playing chess, well is ultimately about recognizing patterns and thinking through a series of moves and “what-if” scenarios to arrive at a single goal… checkmating the king, it’s the most important part.
But, when designers of A.I.-powered chess playing programs started testing their systems years ago, they noticed something really weird about the strategies their programs employed. They were extremely bizarre… they were strategies human players would never come up with… and in many cases would be outright ridiculed by chess grandmasters.
Here’s a snipet that caught my attention from a recent conversation at Tradesmith, a company which has spend millions building their own AI stock picking machine:
(bold highlights mine)
CHRIS:
Interesting... So, what would one of these bizarre chess strategies look like?
KEITH:
Well, for example, in chess, a player can “sacrifice” a key piece if they believe that sacrifice will lead to ultimate victory. Sacrificing pieces in the pursuit of victory has been a strategy in chess for centuries.
However, to the surprise of human players, A.I. chess programs often make sacrifices that seemed bizarre and nonsensical. A.I. chess programs would create wild and complex strategies humans would never think of. These A.I.-created chess strategies have been called “alien” and “chess from another dimension.” But they ended up crushing human players with those unconventional strategies... even grandmasters like the famous Garry Kasparov.
CHRIS:
So, what you’re saying is that when a chess grandmaster was observing the moves the A.I. was making, to him they may have seemed completely outlandish.
But in the end, it was hard for him to criticize because the A.I. kept winning.
The A.I. was arriving at its ultimate goal… but it got there in a way no one would think of.
The process (the latest machine learning enabled AI) goes through to make these predictions isn’t something you could even begin to fathom.
And that’s not a knock on your intelligence.
It’s not something I could begin to understand either.
Or any other human being for that matter… even if you’re Warren Buffett, Garry Kasparov, Stephen Hawking, or any one of the brightest minds who’s ever lived.
And this is one of the biggest lessons I learned working on our project that I want to pass onto the viewers at home.
And that is… how superior the processing power of A.I. is to the human brain when it comes to certain tasks.
Don’t get me wrong our brains are incredibly complex… but when it comes to making sense of a lot of data… whether that data is in the form of historical stock prices, moves on a chess board, or baseball statistics… even a chess grandmaster like Garry Kasparov or a financial genius like Warren Buffett doesn’t have one thousandth of the computational ability that an A.I. program does…
My takeaways (for good or for bad)
1. It's not even a contest.
So, unless you have the option to go “full AMISH”, don’t try to compete with A.I.
You cannot afford to ignore it. USE IT TO YOUR ADVANTAGE.
As policy makers start using AI, expect some bizarre “incomprehensible” moves in an attempt to “win”, much like AI playing chess.
If you think your job is immune, or can’t somehow be enhanced by AI, you’re kidding yourself. You want to be the person in the room that understands this. If you remember life before the Internet, imagine how much it has changed and how understanding this fully in the early 90’s was critical. This is as big, or bigger.
Get your feet wet here:
—
Why can’t AI learn to be irrational.
Life is not a game of chess, poker or any other game. In any game, there are rules and limits and computers can always run the gamut of possibilities and outcomes, even though there might be millions. But A/I will never incorporate the human factor into its mind and actions and conclusions. In order to survive, A/I must be rational...whereas humans don't have to be.
Any dummy can make up an A/I confloberation, and like with apps, most are completely useless beyond their kindergarten mind sets. A recent bout on the squawk box with a health insurance A/I minded robot answering service gave me all the assurance I need that A/I is doomed.