The Risky Experiment: Building a Game Without Writing Code

I’m building a game, and I’m not writing the code.

Not because I can’t. I’ve spent over 20 years writing code for large companies and startups alike. I understand how software gets built, on small teams and at scale. But something is shifting. AI is fundamentally changing how we write code.

This project, Vibing Risk, is my way of exploring what that shift actually means: where AI is strong, where it breaks down, and what the role of a developer becomes when you’re no longer the one writing the code.

What is “vibing”?

“Vibe coding” gets thrown around a lot right now. At one extreme, it means describing an idea and letting AI do everything. At the other, it’s a collaboration: developer and AI working together.

This project is firmly in the second camp.

I’m not just asking AI to write or tweak functions. I’m treating it more like a very capable junior engineer: I set direction, define constraints, and shape the design. The AI produces the implementation, and I review it, correct it, and push it toward the outcome I want.

Less typing. More directing, reviewing, and steering.

Why try this?

The world is changing, fast. Two years ago, I barely considered AI for writing code. Today, I use it constantly to understand systems, explore approaches, and improve implementations.

But I’ve still been using it like a traditional developer: I write, and AI assists.

This project flips that. What happens if AI comes first? What does development look like when I’m directing instead of typing?

I want to find the boundary: how far AI can go today, where it starts to struggle, and what still needs to improve.

I don’t have answers yet. That’s the point.

What is Vibing Risk?

Vibing Risk uses the traditional game of Risk as a testbed. I’m building an online version you can play with friends, automating things like dice rolls and turn flow. Nothing special, and that’s the point.

I chose a game intentionally. Games force you to deal with real problems: timing, synchronization, and state management across players. The UI is also more demanding than typical apps. This isn’t just forms and CRUD.

It also pushes AI into more creative territory, where there isn’t always an obvious solution. And it lets me push one step further: AI writing the systems that power the game, including AI that builds other AI.

The goal isn’t the game itself. It’s to use something complex enough to expose where this approach works, and where it doesn’t.

What to expect

This isn’t a guide or a finished system.

It’s an exploration. I’ll be building in public and sharing what I learn as I go.

I don’t know where this ends yet, but I’m going to find out.

Follow the project on GitHub: lithylabs/vibing_risk.