The Greatest Player Retires — Declares Artificial Intelligence Invincible.
Is humanity first or the tipping point happening now? Will our kids be alright?
“Even if I become the number one, Deepmind cannot be defeated”- South Korean Go champion Lee Se-dol
I thought I knew. Now, I am not so sure.
I always thought I would always celebrate an Artificial Intelligence victory as a triumph of the human mind. I felt that way when Deep Blue won against Garry Kasparov in 1997. I was young and earnest.
In 2016, I watched the contest between the smartest player on planet earth, Lee Sedol, for the ancient game of “Go” and a computer algorithm from the stable of Google, AlphaGo. AlphaGo won the game 4–1. My left, logical brain rationalizes man vs. machine contest as a human victory irrespective of outcome. My right brain is not so sure.
When the machine won the first three games straight, the world at large (read news articles) started whispering the big I word (no, it is not an Apple product) that the computer has made forays into — Intuition.
Intuition is our (human) hack and we have not figured out how it works. We just know it works inside us. Newspapers have a way of hijacking us with their headlines.
Has intuition been mastered by a machine? Or is it just hot air in the media?
Here is my narrative in a story format. (I always dreamt of converting a complex topic that means something to humanity and humanize it into a simple, engaging story.)
The setting
Two players are seated across each other with the board game, “Go” staring in the middle. Lee Sedol has the Korean flag in front of him. The other player makes the moves that AlphaGo recommends. The flag in front is United Kingdom’s Union Jack. AlphaGo is Deepmind’s baby, a UK company acquired by the Google family.
This setting repeats itself for all the five games. The roller coaster of emotions and how it panned out is in the diagram below, inspired by Kurt Vonnegut’s story narrative.
What did the computer actually do? (My goal: no Artificial intelligence jargon words)
A story about a student comes to mind. I heard this when I was young. This student always wanted to study at Wharton and get an MBA. So, he made a trip to Philadelphia, took a picture in front of Wharton and kept it in front of his computer from his high school days. When I first read about this story, what stuck me was his clear goal in his mind that was reinforced by a visual in front of him! He did not know how exactly, but he did it — he graduated from Wharton with an MBA.
This is exactly what this AlphaGo did — at least in my opinion. It was trained by games of the past and was primed like a horse with blinkers. This is something we have come to expect machine learning to do — image search in Google being a great example.
What was interesting about AlphaGo is that it has special blinkers that let the eye have a wider periphery to evaluate multiple track options (like levers for changing tracks for trains). AlphaGo has the mental stamina to visualize the likely outcome (like the student with the picture) for different tracks and assign odds. In the process, it imagines its own games and self learns from its own creations!
Being a better bean counter, AlphaGo won the first three games and the fifth. That makes the fourth game, the game that Lee Sedol won interesting.
Unlike Chess, it is not practically possible for the Alphago to simulate every conceivable path- apparently there are more alternatives than the atoms in the observable universe. So, it approximates the odds based on samplings of paths. Being the human genius [that] he is, Lee found one lone path that was hidden in an unlikely route.
After winning that game, he was given a standing ovation. The celebration after the game spoke about how the world felt and which side our hearts are leaning in to find solace.
Like many of us, deep in my heart, I was getting sentimental for a human win. My father used to say in half jest, “As you grow older, you get more sentimental as you have more to look back.”
What It Means To You And Me.
That was 2016. This is 2019. The greatest champion of Go feels comprehensively defeated by the machine and retires.
The consoling grace is that the definition of a win is well defined in this game. In life — not so much. Will machines figure that one out?
If yes, I am worried about the future for our kids. Will they be alright?
