Intellectually Curious
Intellectually Curious is a podcast by Mike Breault featuring over 1,800 AI-powered explorations across science, mathematics, philosophy, and personal growth. Each short-form episode is generated, refined, and published with the help of large language models—turning curiosity into an ongoing audio encyclopedia. Designed for anyone who loves learning, it offers quick dives into everything from combinatorics and cryptography to systems thinking and psychology.
Inspiration for this podcast:
"Muad'Dib learned rapidly because his first training was in how to learn. And the first lesson of all was the basic trust that he could learn. It's shocking to find how many people do not believe they can learn, and how many more believe learning to be difficult. Muad'Dib knew that every experience carries its lesson."
― Frank Herbert, Dune
Note: These podcasts were made with NotebookLM. AI can make mistakes. Please double-check any critical information.
Intellectually Curious
WallZero: Mastering WallGo with Strategic AI Analysis
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
We dive into the WallGo breakthrough where an AI called WallZero uses a reachability mindset to plan future moves on a shifting 7x7 board, defeating top players and revealing new depths of strategic game design. From endgame point sacrifices that flip turn order to millions of self-play insights testing fairness of different starting setups, we explore how this AI collaboration reframes how we think about board control and real-world systems like urban planning and resource reachability.
Note: This podcast was AI-generated, and sometimes AI can make mistakes. Please double-check any critical information.
Sponsored by Embersilk LLC
I was actually yelling at my screen last night while I was watching The Devil's Plan. You know, just seeing them sweat over Walgo.
SPEAKER_01Oh, yeah, it gets so incredibly intense.
SPEAKER_00Right. And since so many of you out there play or watch the game, you know exactly how fast that tiny seven by seven board just gets completely claustrophobic.
SPEAKER_01It really does.
SPEAKER_00But um what totally blew my mind today is this newly published research paper showing how an AI basically just broke the game wide open. But uh before we get into how these researchers exposed all these hidden strategies, I want to quickly mention that if you're feeling inspired to build your own AI to solve complex puzzles, today's sponsor Embersilk can actually help with that.
SPEAKER_01Yeah, they are fantastic for that sort of thing.
SPEAKER_00Exactly. Whether you need AI training, software development, or you just want to uncover where AI agents could make the most positive impact in your business or personal life, check out Embersilk.com.
SPEAKER_01And you definitely want an agent like this, wall zero AI on your side. I mean, looking at the data, it didn't just master Walgo. No, it uncovered this whole new level of strategic depth that completely reframes uh how we even think about board control.
SPEAKER_00Aaron Powell Okay, so let's unpack this a bit. Because the paper mentions that the dual action in the game, you know, where you move a stone and then you build a wall that creates a game tree complexity of 10 to the power of 87.
SPEAKER_01Yeah, which is just astronomically huge. Aaron Powell Right.
SPEAKER_00It's like trying to play chess, but you're simultaneously building the walls of a maze around your own pieces. So how does an AI even begin to parse a maze that's constantly shifting?
SPEAKER_01Aaron Powell Well, it all comes down to a specific mechanism the researchers programmed. They call it reachability. So instead of just looking at where the walls are right now, Wall Zero evaluates where its pieces can potentially go in the future.
SPEAKER_00Oh, I see.
SPEAKER_01Yeah, you can think of it like water flowing through a maze. It isn't just looking at the current barriers, it's actually calculating where the water will ultimately pool over, say, the next 10 turns.
SPEAKER_00Okay, so it values keeping its own pipes open, basically, over just trying to close off the opponent.
SPEAKER_01Exactly. It prioritizes future flexibility over grabbing immediate territory.
SPEAKER_00But wait, does that kind of abstract water flowing logic actually hold up against human intuition? I mean, humans are pretty good at this.
SPEAKER_01It more than holds up. The researchers actually tested Wall Zero against two Taiwanese professional Go players.
SPEAKER_00Oh, wow.
SPEAKER_01Yeah. And one was a nine Dan, which is essentially grandmaster level, and Wall Zero won all eight formal matches.
SPEAKER_00All eight. That is wild.
SPEAKER_01And it secured, on average, almost twice as much territory as the humans, like 1.9, 8 times more.
SPEAKER_00Wait, I was reading the match logs, and it was doing this by using its own stones as temporary walls, right? It's exactly like it wasn't even building physical walls to block the Grandmasters, it was just body blocking them while keeping its own reachability totally wide open.
SPEAKER_01Precisely. And honestly, this brings up the most optimistic part of the whole research. By playing millions of games against itself with this reachability mindset, Wall Zero essentially turned into an automated game designer.
SPEAKER_00Oh, right. Because it was analyzing the fundamental fairness of the game board itself.
SPEAKER_01Exactly. It actually proved mathematically that the four-stone mode they play on the TV show, where the four pieces are pre-placed, that actually creates a more balanced game than starting on an empty board.
SPEAKER_00That is so cool. But hold on, here is where it gets really interesting for me. I was a bit confused by that implicit passing strategy it invented for the end game.
SPEAKER_01Oh, the point sacrifice strategy.
SPEAKER_00Yeah. Because if the AI is intentionally sacrificing points just to change the turn order, doesn't that put it behind on the scoreboard? Like, how does giving up points actually lead to a win?
SPEAKER_01So it's a brilliant manipulation of force moves. Yeah. Since you absolutely have to move and build a wall every single turn, entering a highly contested area first can sometimes trap you.
SPEAKER_00Right, because you run out of space.
SPEAKER_01Exactly. So while Zero realized that if it sacrifices one point early on, it flips the turn order. That forces the human player to make the mandatory move into that tight space instead.
SPEAKER_00Oh, I get it. So the human is forced to build a wall that effectively traps their own pieces.
SPEAKER_01Yeah. Which ends up costing them way more than the single point the AI sacrificed in the first place.
SPEAKER_00That is wild. It's like giving up a penny to force your opponent to drop a dollar.
SPEAKER_01It really is. And I think this just shows how incredibly positive AI collaboration can be. I mean, it isn't just about defeating humans.
SPEAKER_00No, not at all.
SPEAKER_01It's this incredibly uplifting tool that can analyze fairness, validate rule sets, and just help us unlock solutions to massively complex problems we couldn't even visualize on our own.
SPEAKER_00Definitely. Knowing that Wall Zero can act as an automated designer, testing the fairness of a system with millions of simulations, it really makes you wonder about the bigger picture. Like what happens when we start applying this technology to real world systems?
SPEAKER_01Oh, there's so many possibilities.
SPEAKER_00Right. Imagine using this kind of AI to optimize urban planning so everyone has better access to resources. We might soon let AI help us design the systems we use in real life so everyone has better reachability. So it leaves you with a thought how often in your own life do you prioritize an immediate small win over your long term reachability and keeping your future options open?
SPEAKER_01That is a fascinating application to think about.
SPEAKER_00It really is. Well, if you enjoyed this intellectually curious exploration, please subscribe to the show and hey, leave us a five star review if you can. It really does help get the word out. Thanks for tuning in.