When Programmers Play Games, They Break Them on

The most telling moment in a recent episode of TheStandup podcast came when one host admitted his wife refuses to play Magic: The Gathering with him anymore. His crime? Building infinite loops that make the game mathematically unwinnable for his opponent, then crashing the digital version of the game for sport.

"Every time you draw a card, you're able to discard a card. Every time you discard a card, you're able to add an extra guy. Every extra guy you have, you can draw a card. Uh-oh. We're in a loop," he explained, clearly delighted. "I crash Magic the Gathering Arena often when I get on purpose. That is the point. That is my only goal playing."

This isn't aberrant behavior for programmers who game. It's practically definitional. Where most players see entertainment, developers see systems to exploit, optimize, or break entirely. The podcast, featuring several tech industry veterans including ThePrimeagen, became an unintentional field study in how coding changes the way you interact with virtual worlds.

Building Computers Inside Games

The conversation started innocuously enough: what's everyone's favorite game right now? One host chose Terraria, a 2D building game that's often called "2D Minecraft" by people who've never played it. But the interesting part wasn't the choice—it's what he's doing with it.

"I've been using ChatGPT to design Terraria circuit designs for me," he said. "It literally prints out exactly the schematic. You tell it what you want, you tell it how you want it to evolve."

Someone built a 32-bit computer inside Terraria. This programmer is working on a 33-bit version. When pressed on whether this represents meaningful progress, he responded with programmer logic: "It has 33 bits. It's one more bit."

This is the fundamental difference between how programmers and civilians approach games. The game provides wires and logic gates. Most players use them to make their character jump higher or a door open faster. Programmers use them to implement Von Neumann architecture.

Terraria's developers made this possible by including what amounts to assembly language in a video game. They gave players the primitives. Programmers did what programmers do—they built abstractions until they could run Doom on it. (Nobody's done that yet, but give them time.)

The Optimization Compulsion

ThePrimeagen's current obsession is Slay the Spire, a deck-building roguelike that he plays exclusively on his phone. He's stuck at Ascension 6 while his colleague has reached Ascension 19, which clearly bothers him more than he'd like to admit.

"I keep losing at it regularly and I feel like I should be better," he said. This is programmer thinking applied to entertainment. The game isn't supposed to be solved. You're supposed to enjoy the experience of occasionally losing. But programmers see failure as a bug in their understanding, not a feature of the design.

The group's disdain for luck-based games like Candy Crush makes sense in this context. "I think you have to have not a lot going on upstairs to enjoy Candy Crush," one host said, which is both condescending and revealing. Candy Crush's appeal is partly because you can't optimize it fully. Programmers hate this. They want deterministic systems where better understanding produces better results.

One host mentioned spending 20 to 30 minutes deciding his next move in Civilization—while in the bathroom. "You would be the absolute worst person to play Civ with," another responded. "When it's his turn, everyone gets up and makes snacks."

This captures something real about how developers interact with complex systems. The turn isn't just a turn—it's a problem to be solved optimally. Every decision tree needs to be evaluated. This makes you unbearable to play board games with but excellent at your job.

The AI Layer

The casual mention of using ChatGPT to design game circuits deserves more attention than it got in the podcast. This represents a shift in how people use AI tools—not for writing marketing copy or summarizing documents, but for solving niche technical problems in recreational contexts.

"The first one worked right away and now I've been going further," the Terraria player said about his AI-generated circuits. This suggests the model actually understands the game's logic system well enough to design functional circuits. That's either impressive or concerning, depending on how you feel about AI capabilities.

The hosts plan to collaborate on Terraria projects, potentially building something together over a week. They discussed using version control for their save files. This is where programmer culture fully colonizes gaming—treating a sandbox game like a software project with sprints and repositories.

What This Says About Both Fields

The gaming industry has spent decades making games more accessible, smoothing rough edges, removing exploits. Meanwhile, a subset of players actively seeks those rough edges out. They want the exploits. They want to find the mathematical break points.

This isn't new. Tool-assisted speedruns have existed for years, where players use frame-perfect inputs to complete games in ways the developers never imagined. But what's different now is the sophistication of the tools. ChatGPT can help you design circuits. You can crash multiplayer games with exponential token loops. The gap between "playing the game" and "reverse-engineering the game" has narrowed.

The question isn't whether this is the "right" way to play. The question is what happens when the people building our software approach their jobs the same way they approach games—looking for edge cases, hunting for exploits, trying to make the system do something it was never designed to do.

Sometimes that produces elegant solutions. Sometimes it produces 33-bit computers in Terraria. Sometimes it produces infinite loops that crash game servers. The mindset is identical. The outcomes vary.

—Bob Reynolds