Cowritten by ChatGPT o1

Few things are cozier than the occasional super-nerdy round of Age of Empires II with friends on cold winter nights with our own set of rules. Even after 20 years, it still brings me so much joy.

It’s also taught me a few lessons—not just about resourcefulness or the classic cheat code (“how do you turn this on”), but deeper ideas that stuck with me as a designer and strategist. Every game starts the same: the mini-map is hidden, the resources are equal, and you’re left with your small, familiar starting point. If you play competitively, the start of the game becomes an optimized sequence of clicks—a perfected build order. But what always fascinated me was the blank map itself. There are lessons in that emptiness.

The map begins completely hidden, with two layers of uncertainty: the unknown (what you’ve never seen) and the fog of war (what you’ve seen but can no longer track). This idea has become a common metaphor for how I approach new projects and emerging fields. Every project feels like starting on a blank map. You don’t know the lay of the land. Sure, you can get started, do the work, and begin building. But you also need to send out scouts to understand what’s out there.

Scouting in Age of Empires is a multitasking skill. You’re balancing your base’s build order while ensuring your scout keeps moving forward. There are three objectives for my scout: understanding the map, discovering resources, and locating allies or enemies. It’s a balancing act—keeping your base productive while constantly uncovering more of the terrain. In projects, it’s much the same. While you’re developing your “base,” you also have to explore the landscape, map opportunities, and identify strategic areas where you might stake a claim.

The fog of war takes this a step further. Even after you’ve explored, areas slip back into obscurity if you’re not watching closely. In Age of Empires, you can build scout towers or send troops to keep tabs on strategic points—but both cost resources, especially early in the game. In work, it’s no different. What’s worth observing? Where do you need to focus your attention? Where is it essential to show presence—whether that’s launching a product or positioning yourself before someone else does?

Playing with friends has always added another layer to these lessons. Over the years, we developed our own house rules to make the game more collaborative among us. Age of Empires can be played fast, but that’s not our style. Instead, we team up against the AI. It’s fascinating to watch how the AI’s strategies remain hardcoded—rigid and biased in ways that don’t align with how we want to play. Our goal is to win, but with fun over time. Sometimes it feels like the AI just doesn’t understand that.

This disconnect reminds me of the challenges I face with AI today. I project human traits onto these systems, but they operate with completely different logic. Prompts are meant to express my intent, but the feedback feels one-sided, like a hyperlink—linear, directional. As AI becomes more capable, this gap widens. How do we create systems where intent flows both ways? Where machines can reflect their goals back in ways we can interpret? With humans and AI speaking entirely different “languages”—each with its own cultural, intellectual, and ideological coding—how do we even begin to build mutual understanding? Our conversations flow in parallel from different directions but often do not fully match.

In Age of Empires, building and optimizing your base is just one part of the game. An equally important part is observing the land—knowing what’s out there and what’s happening. In design, I’ve approached this through research and prototypes, similar to placing scout towers. But with AI, some of these tasks shift. Agents can handle build orders better than I can—they optimize and execute flawlessly. Maybe this is why I still lose to the hardest AI after all those years. (Maybe I am just bad.)

As AI takes over base-building, my role shifts to observer and systems interventionist. I place scout towers, send out troops, and take strategic action. My job becomes about decision-making, observing cause and effect, and curating systems that respond dynamically. It’s a pattern I see in my own work with AI tools like Replit too. Instead of managing infrastructure, I focus on exploration, testing, and shaping outcomes through prompts in a live environment.

This brings back questions though. Agents aren’t just better at optimizing—they’re better at observing, multitasking, and simulating all potential decisions. They can predict outcomes faster and at a scale that I can. In many ways, they’ve taken over the observer and curator roles I once held. At least slowly, but possible then all at once.

Which brings me back to playing Age of Empires. My friends and I create house rules because we want to play differently. We want to collaborate, to shape the game in a way that aligns with how we think. When agents take over the grind and the systems work, what’s left for us isn’t playing the same game.

Which leaves me wondering: what’s the game this new game we play?