Move 37 and what humans can do in the “age of AI”.

Pretext and Context:

Day 7 of the lockdown and in this next article I take you back 4 years and try to show you what the future looks like. The day is 10th March 2016, the venue is Seoul, South Korea, the occasion is the second match of the now historic 5-match GO series between AlphaGo (Google’s GO playing AI) and Lee Sedol, the reigning world champion. 

Sedol had claimed pre-series that he’d win 5-0 or 4-1 and that machines could never get close to playing at the level of man when it came to GO. And with good reason too. If you thought that chess was complex, think again. With more combinations in the game than the number of atoms in the universe (go ahead, do the calculation, we all have a lot of time on our hands), GO is probably as complex as it gets. For a computer to be able to logically deduce all the moves and then play the best one against a grandmaster of the game was thought to be impossible. 

Move 37!:

What happened over the course of the next week was unprecedented. AlphaGo beat Lee Sedol 4-1, and this was, by Sedol’s own admission, a complete wipeout where he never felt in control. For an AI agent to beat the world champion of GO was touted by some as the most astonishing achievement in the field of AI to date. This was 2016. Of course, we’ve progressed by leaps and bounds in the application of AI since then, from driverless cars to identifying/diagnosing deadly diseases (this particular COVID-19 pandemic included). We’ve pushed the boundaries of image recognition, image segmentation, transfer learning, additive networks, reinforcement learning, and a bunch of new fields too. However, that series will remain in the minds of people forever, irrespective of how much farther we’ve come. It will remain in the minds of people not just for the way AI beat the most accomplished human in the field, but also for one specific move: Move 37!

In Game Two of AlphaGo versus Lee Sedol, the AI made a move no human would ever think of doing. And it was beautiful. “Move 37” was unimaginable in the more than three thousand year history of the game. By taking a position on the “fifth line”, AlphaGo pushed the boundaries of human intuition. "That's a very strange move," said one commentator, himself a nine dan Go player, the highest rank there is. "I thought it was a mistake," said the other. Lee Sedol, after leaving the match room, took nearly fifteen minutes to formulate a response. As the world looked on, Move 37 so perfectly demonstrated the enormously powerful and rather mysterious talents of modern artificial intelligence. 

1 in 10,000:

AlphaGo learns from human moves, and then it learns from moves made when it plays itself. It understands how humans play, but it can also look beyond how humans play to an entirely different level of the game. This is what happened with Move 37. AlphaGo had calculated that there was a one-in-ten-thousand chance that a human would make that move. But when it drew on all the knowledge it had accumulated by playing itself so many times—and looked ahead in the future of the game—it decided to make the move anyway. And the move was genius to some, and scary to others. 

That is the reason why Elon Musk (among others) is concerned about the future role of AI and made his famous quote: “With artificial intelligence, we’re summoning the demon.” Because when you play against a system that is smarter than you are, all you can predict is that you are going to lose. However, rather than creating an independent intelligent agent, our implementations of AI are more like a way of extending our own intelligence, which seems a much more optimistic and interesting way of framing AlphaGo’s victory. These are all tools that make us better. And make Lee Sedol better it (AlphaGo) did. 

Move 78! and 1 in 10,000, again:

After losing the first three matches, with tens of millions of people watching, Sedol had nothing but his pride to play for in the fourth. Midway through the game, the Korean's prospects didn't look good. But after considering his next move for a good 30 minutes, he delivered something special. It was Move 78, a "wedge" play in the middle of the board, and it immediately turned the game around. AlphaGo made a disastrous play with its very next move, and just minutes later, after analyzing the board position, the machine determined that its chances of winning had suddenly fallen off a cliff and resigned. Among Go players, the move was dubbed "God's Touch." It was high praise indeed. But then the higher praise came from AlphaGo. Sedol had now done something that wasn’t done in the three thousand year history of the game. Drawing on its months and months of training, AlphaGo decided there was a one-in-ten-thousand chance of that happening, exactly the same tiny chance that a human would have played AlphaGo's Move 37 in Game Two.

The symmetry of these two moves is more beautiful than anything else. One-in-ten-thousand and one-in-ten-thousand. This is what I take away from the series. DeepMind built a machine capable of something super-human. But at the same time, it's flawed. It can't do everything we humans can do. In fact, it can't even come close. It can’t write this article (yet). It can't carry on a conversation (yet). It can't play charades (yet). It doesn’t have the common sense of a teenager (yet). It can't account for God's Touch (yet).

But think about what happens when you put these two things together. Human and machine. This isn't human versus machine. It's human and machine. Move 37 was beyond what any of us could fathom. But then came Move 78. And we have to ask: If Lee Sedol hadn't played those first three games against AlphaGo, would he have found God's Touch? The machine that defeated him had also helped him find the way.

Lessons for us (What can Humans do?):  

As we enter the “age of AI”, an age in which thought/research leaders predict that machines will be capable of achieving “general intelligence”, the role of humans and certain jobs will come under the gun. A lot of this is already happening with automation, machines, and AI “replacing” significant portions of the workforce. The question then is, what can humans do? There are a few core areas where humans have a huge edge over machines, a chasm so deep, it might not be bridged in our lifetime (or so I think, could be wrong). Decision-making, for example, is more than just applying rules. It’s using discretion, instinct, experience, playfulness, and emotion — all aspects of human judgment. Empathy is another gift humans have that we need to practice now more than ever. It won’t just help us feel more engaged at work, it’s how we’ll make real progress in today’s society. Curiosity to be able to push boundaries is what makes us human and using this curiosity right now to understand trends and reacting is what separates us from the machines. Invest in learning the skills that are required to usher in the “age of AI” and make sure that you’ll not be left behind. 

Oh, and also, learn AI. Schedule your free consultation with Gradvine to know more about specific paths to learning AI through undergraduate and graduate degrees.