Move 37 and the AI Creativity Revolution

From a baffling early-game move that shocked pros to a broader reckoning with how AI reshapes strategy and science, this episode dives into the 2016 Lee Sedol–AlphaGo match. We unpack move 37, its field-shaping genius, and how AlphaGo’s unconventional intuition foreshadowed AlphaFold—showing how humans and machines can push each other toward new heights of imagination, and what that means for our own habits and breakthroughs.

Note:  This podcast was AI-generated, and sometimes AI can make mistakes.  Please double-check any critical information.

Sponsored by Embersilk LLC

Transcript

SPEAKER_00 0:00

So I was playing this casual strategy board game with a friend over the weekend, and uh I tried to pull off this just massive, incredibly confident bluff.

SPEAKER_01 0:09

Oh no, how did that go?

SPEAKER_00 0:11

I mean terrible. I moved literally all my pieces to one side of the board, you gave them this super smug look, and then I just lost instantly on their very next turn.

SPEAKER_01 0:20

Oh man.

SPEAKER_00 0:20

Yeah, I completely miscalculated the whole thing. It really highlighted for me just how flawed human strategic intuition can be sometimes.

SPEAKER_01 0:29

It really is, and that actually happens to the best of us.

SPEAKER_00 0:32

Which brings us perfectly into today's deep dive into our sources. Because today we are looking at the historic 2016 Go match between the human champion Lisa et al. and the AI, AlphaGo.

SPEAKER_01 0:43

Right. And specifically, we're focusing on this one baffling moment, uh, move 37 that basically redefined how we even understand creativity.

SPEAKER_00 0:51

Aaron Powell Exactly. But to really appreciate why it's such a big deal, you kind of have to understand the insane complexity of Go.

SPEAKER_01 0:58

Oh, definitely. Like this is not chess. The Go board has roughly 10 to the power of 170 possible positions.

SPEAKER_00 1:05

Aaron Powell, which is just a staggering number.

SPEAKER_01 1:07

It is. It's vastly more than the number of atoms in the observable universe. So you simply cannot use brute force calculation to win. You have to rely on deep, long-term planning, and you know what professionals call intuition.

SPEAKER_00 1:20

Aaron Powell Right, which sets the stage perfectly for game two. So AlphaGo plays move 37, which commentators called a shoulder hit on the fifth line.

SPEAKER_01 1:29

Aaron Powell Yeah, and for anyone who isn't playing Go every single weekend, a shoulder hit just means you place your stone diagonally right next to an opponent's stone.

SPEAKER_00 1:36

Aaron Powell But doing that on the fifth line, which is way up high toward the empty center of the board, that is totally unheard of this early in the game, right?

SPEAKER_01 1:44

Completely unheard of. Usually you fight for the edges first to secure some easy territory. The AI system actually calculated that there was only a one in 10,000 chance a human professional would ever play that move.

SPEAKER_00 1:55

Wow, one in ten thousand. And the human reaction on the live broadcast was wild. The professional commentators literally thought it was a blatant mistake.

SPEAKER_01 2:04

They did. To their trained eyes, the move looked too thin, like structurally fragile. It's kind of like building a massive fortress wall, but just leaving these huge, obvious gaps in it.

SPEAKER_00 2:15

And Lee Sedlett actually stepped out for a quick smoke break, right? So he comes back, sees this stone on the board, and he is visibly stunned.

SPEAKER_01 2:22

Oh yeah. It took him 12 minutes of precious clock time just to process what happened and come up with a reply.

SPEAKER_00 2:28

I was actually trying to wrap my head around why this move worked. And it feels a bit like say you're playing a big resource strategy game, and you build this tiny, seemingly useless outpost in the middle of an empty desert.

SPEAKER_01 2:40

Right. It looks like a total waste of resources.

SPEAKER_00 2:42

Exactly. But an hour later, the whole map shifts, and you end up forcing your opponent to retreat directly into that exact outpost.

SPEAKER_01 2:49

That is a brilliant way to visualize it. And speaking of AI systems coming up with totally unexpected solutions, this deep dive is sponsored by Embersilk.

SPEAKER_00 2:58

Yes, if you need help with AI training or automation or integration or even software development, you should really check them out.

SPEAKER_01 3:06

Definitely. Uncovering where agents could make the most impact for your business or personal life. Check out Embersilk.com for AI needs.

SPEAKER_00 3:13

So getting back to that desert outpost analogy, analysts later realized Move 37 was this absolute genius act of field shaping.

SPEAKER_01 3:22

Exactly. AlphaGo just ignored all the immediate, obvious local gains. Instead, it placed a stone that subtly altered the gravity of the entire board.

SPEAKER_00 3:32

By placing that single piece in the center, it ensured that dozens of moves later, when all those edge skirmishes finally spilled into the middle, AlphaGo's pieces would perfectly connect to secure this massive advantage.

SPEAKER_01 3:44

It's incredible.

SPEAKER_00 3:45

But wait, here's where I get hung up. Does this mean that human Go masters who developed this strategy all for centuries were just fundamentally wrong about how the game works?

SPEAKER_01 3:53

Well, no, not wrong, just constrained by tradition. You see, move 37 didn't break the mathematical logic of the game. It just explored this deeper, highly unfashionable part of Go's DNA.

SPEAKER_00 4:05

Because humans rely on these stylistic habits just to manage the overwhelming complexity of it all.

SPEAKER_01 4:10

Precisely. AlphaGo didn't care about what looked weird or thin to a human. It proved that if an intelligence is unburdened by our traditions, it can uncover structurally novel solutions.

SPEAKER_00 4:22

And those novel solutions scale way beyond just a wooden game board. I mean, the exact same logic behind AlphaGo evolved directly into AlphaFold, which is so exciting.

SPEAKER_01 4:31

It really is. The mechanics are incredibly similar. AlphaGo learned to predict the overarching shape of a winning board, and Alpha Fold learned to predict the 3D shape of a protein based purely on its one-dimensional string of amino acids.

SPEAKER_00 4:45

That is just mind-blowing. By recognizing those vast structural patterns, instead of doing slow physical trial and error, AlphaFold successfully mapped 200 million proteins.

SPEAKER_01 4:54

Yeah, and that breakthrough is fundamentally accelerating real-world scientific progress. It's helping develop things like malaria vaccines, and it even earned its creators a 2024 Nobel Prize. It is such a beautiful example of how uncovering hidden structures helps us solve grand challenges.

SPEAKER_00 5:10

It is so hopeful. And here is the best part. We humans are far from obsolete here. Because in game four, that same go match, Lisa et al.

SPEAKER_01 5:18

He did. He played his own miraculous structural anomaly, Move 78. Commentators actually dubbed it God's Touch.

SPEAKER_00 5:26

And it completely shocked AlphaGo and led to a human victory. To me, that proves humans and AI can push each other to entirely new heights of imagination.

SPEAKER_01 5:36

It is a deeply optimistic vision of our future. It is this engaged collaboration where the technology helps us break out of our ruts and then we use that expanded vision to see the world differently.

SPEAKER_00 5:46

Exactly. So think back to my terrible board game bluff at the start. It really makes you wonder what obvious rules in your own life are actually just unquestioned habits.

SPEAKER_01 5:56

What rut are you stuck in, just waiting for your own move 37 breakthrough?

SPEAKER_00 6:00

Something to think about. If you enjoyed this podcast, please subscribe to the show. Hey, leave us a five star review if you can. It really does help get the word out. Thanks for tuning in.