Virtually everyone listening to this has played a game against a computer at least once in their life time.
Whether it be a simple game of checkers or a high definition $60 video game you have been trying to beat for weeks. Humans play computers all the time. And we’ve been doing it for decades.
At first, we humans marveled at our ability to defeat our own creation on whatever game we chose.
But we kept improving our computers and that make playing against them more challenging and entertaining.
This trend continued until a pivotal moment occurred back in 2017
GO: Now allow me to shift your attention to a game you may never have heard of: Go.
The game GO is a much older creation of mankind than computers. So old that it’s origins are unknown, but it is believed to be over 2,500 years old. With the earliest written account dating back to 548 BCE in China.
It is an elegant game. They say the rules are so simple, it could probably be taught to intelligent alien life. But that doesn’t mean it is easy
Go is the most complex game devised by mankind. It has 10 to the power of 80 possible outcomes.
For a reference, some games much more familiar to western listeners:
Checkers has 10 to the power of 20 possible outcomes
And the game of chess only has 10 to the power of 40 outcomes.
In multiple Asian countries Go Officials test young children’s aptitude at the game. If a child shows enough promise, they are taken to special institutes where the children study and play Go as a full time career.
Professional Go players may become state sponsored and can be payed handsomely for their prowess.
This game has had entire cultures dedicate countless lives to becoming better at playing it for hundreds of years.
After thousands of years, millions of players, and possibly the most effort spent to perfect any kind of game play by humankind….
In May of 2017, an AI called AlphaGo was pitted against the world champion GO player.
The 19 year old Kie Je from Go’s Birthplace, China, was defeated 3 games in a row by an Artificial Intelligence created by Google.
Thousands of years of human knowledge was accumulated by a computer in a matter of days. Then that computer executed that knowledge flawlessly to beat the best human Go player in history.
Google’s Deep Mind project created AlphaGo to learn. They taught the AI the rules to the game and then made it play against itself millions of times all at once. After a few short days of learning how to play Go from scratch (no institutes, no ancient techniques) AlphaGo had become the greatest player of the game.
The Chinese Government was so infuriated by the fact that their champion was being defeated by a product from an American company, that they banned the live stream after the first game.
But China wasn’t the only culture to be shocked by the outcome. To many in the Go Community, this was a devastating blow. The game itself has lost many players. The collective human motivation to keep playing Go took a hit.
In the years since Kie Je’s loss in 2017 there has been a drop in the number of people willing to practice the game.
If a computer program can beat arguably the greatest player of all time after just a few days of existing, then what is the point?
Why waste any more resources on a battle that human kind will surely lose? What if they let AlphaGo practice for months or years?
As humans, We’ll never be able to catch up.
But playing a game isn’t always just about winning or losing.
As I said earlier, Although AlphaGo was the most interesting event in Game AI history, it definitely was not the first time humans used computers to play games.
A man who calls himself Tom Seven created a much less sophisticated AI than AlphaGo for a research paper he was writing.
Tom Seven’s AI named SIGBOVIK 2013 was designed to play Nintendo Entertainment System (or NES) games like Super Mario. With all the games the NES had, there were a lot of different scenarios that SIGBOVIK 2013 ran through.
You can watch his videos on YouTube and watch this algorithm work its way through different games through trial and error, trying every single option it could. It is intriguing to watch.
The most notable scenario for me was when SIGBOVIK was tasked with playing Tetris.
====Audio clip 1===Top Seven said :”Playing Tetris well requires some thinking ahead and this algorithm does not think very far ahead.”
Then SIGBOVIK tested out the Pause feature in the middle of a game.
======Audio clip 2=== “there it was pausing the game for no reason”
SIGBOVIK 2013 realizes it cannot win. It is a learning AI, but it isn’t as advanced as what Google’s Deep Thought Project can create. SIGBOVIK 2013 cannot win this game of Tetris. So what does it do?
======Audio clip 3=====”Fast Forward a bit to see how this all ends. It’s not good. So now it is almost done and it pauses the game. Because as soon as he un-pauses he will lose, and really the only winning move is not to play. Thank you.”
SIGBOVIK 2013 decided that instead of losing, it would just pause the game indefinitely. It gave up. It quit instead of going on.
Some saw this as comical that the poor little AI just decided to give up, but I saw it as tragic.
And while human beings have done their fair share of giving up, we also have something called hope.
When AI is given a task like Tetris or Go, and is then told Win=Good and Lose=Bad, it will show great intelligence in accomplishing a win.
But when the odds are stacked so that an AI cannot win, the AI simply pauses the game, refusing to play. It gives up.
THAT is the difference between Human intelligence and Artificial Intelligence.
Some players of Go may have given up their dream of being the next best player of the most complex board game, but there are many other players who continue to perfect their play style regardless of AlphaGo’s shocking accomplishment.
We have hope. Humans will keep trying even in the face of unsurmountable adversity.
Humans may find an unforeseen opportunity in the face of the impossible to overcome an obstacle.
Where an AI will simply give up.
Kie Je made a statement after he lost his match with AlphaGo: “With AlphaGo, I think my understanding of not only the game of Go, but life, has changed.”
His words are left up to interpretation, but I like to think it means he isn’t giving up.
Hope you enjoyed my nerdy pep talk there. I’ve included the accompanying blog post in the description if you prefer to read this podcast. I’ve also included my sources there.
Thanks for listening Who’d a Thunkers! Until next week.