How I Built a Puzzle Game with AI – 1,050 Hours Later

Published 10 August 2026 · Updated 21 September 2026 · By Ivo Čierny

Beauty of Symmetry began as a simple idea for a logic game. I did not know how to program it, I did not have a development team, and I certainly did not expect the project to consume more than a thousand hours of my time.

What eventually made the game possible was artificial intelligence. This is the story of how a paper puzzle idea became an Android and web game with AI acting as my development partner.

The idea started long before the app

I have always enjoyed inventing games. Years before Beauty of Symmetry existed, Silvia and I made physical logic games from wood, including tic-tac-toe, Nine Men's Morris, 3D tic-tac-toe and several other ideas.

We tried offering them to small shops. From a business point of view, it was not particularly successful. We probably underestimated how important marketing is. Creating a product is difficult, but getting people to notice and buy it can be even harder.

Still, I enjoyed the process enormously. For me, inventing games was never mainly about becoming rich. It was about creating something that did not exist before.

12 December 2024 – the first Beauty of Symmetry board

The idea that eventually became Beauty of Symmetry had been developing in my mind for a long time.

On 12 December 2024, I finally sat down and drew the first version of the game board.

From that point, I started experimenting with the rules, moving pieces around and asking a simple question: could disorder be transformed into a visually pleasing, symmetrical arrangement through a small number of logical moves?

Over the following weeks, the basic rules became clearer. The central idea remained simple: move groups of balls between columns until the board becomes symmetrical around a vertical axis.

At first, I considered turning the idea into a physical game. But producing, packaging and distributing a physical product would require a significant investment.

Beauty of Symmetry also seemed particularly well suited to a digital format, where puzzles could be generated automatically and players could receive a new challenge every time.

So the physical version was put aside. I decided that Beauty of Symmetry should become a computer game.

There was one problem: I couldn't build it

Having an idea for a game and turning that idea into working software are two completely different things.

When I decided to create Beauty of Symmetry as an application, I did not know how to program the game.

I began learning JavaScript and tried to understand enough programming to make the idea possible.

Learning was useful, but I gradually realized that becoming experienced enough to create the complete game on my own through traditional programming study would take a very long time.

I considered finding programmers and designers who could help. But building a small independent project with other people was more difficult than I expected.

The project could easily have ended there.

February 2025 – AI became my development partner

Then I decided to try a very different approach: instead of waiting until I had become an experienced programmer, I would use artificial intelligence to help me build the game while I learned.

AI did not magically create Beauty of Symmetry with one command. The process was much more practical.

I described what I wanted to achieve. AI suggested code. I tested it. Something failed. We diagnosed the problem, changed the code and tested it again.

Sometimes a new feature broke something that had worked before. Sometimes we spent hours solving a problem that looked simple. Other times, something that I thought would be difficult worked surprisingly quickly.

This became my method of learning. Instead of studying programming only in theory, I learned while solving real problems inside my own game.

Building more than just the puzzle

A finished game requires much more than its basic idea.

Beauty of Symmetry gradually grew to include different board sizes, one-color, two-color and three-color puzzles, Race competitions, leaderboards, Pattern mode and multiplayer games.

I also had to work on the website, Android application, graphics, menus, mobile layouts, sound, Google Play publishing, promotional materials and many small technical details that players normally never see.

AI helped with many parts of that process, but every feature still required decisions, testing and repeated corrections.

1,050 hours later

By the time Beauty of Symmetry reached Google Play, I estimated that I had spent approximately 1,050 hours working on the project.

Today I can take my phone out of my pocket, open an application and play a game that once existed only as an idea in my head.

That still feels strange.

Beauty of Symmetry is not a game with millions of downloads. It is a small independent project competing for attention in an enormous market.

But the game exists. It works on Android and on the web. It has several game modes, online competition, multiplayer, its own website and its own community channels.

For me, that alone makes the project worthwhile.

AI changed while I was building the game

One unexpected part of the experience was seeing how quickly artificial intelligence itself improved.

Development lasted long enough that I could notice a difference between the AI tools I used near the beginning of the project and those available later.

Tasks that initially required long explanations and many corrections gradually became easier. The systems became better at understanding larger pieces of code, identifying problems and proposing useful changes.

So while Beauty of Symmetry was evolving, the technology helping me create it was evolving as well.

Creating the game was only half the challenge

Building a game and getting people to discover a game are very different skills.

After development came another learning process: Google Play, websites, search engines, YouTube, Instagram, Facebook, TikTok, videos, screenshots, trailers and advertising.

In some ways, this brought me back to the lesson from our old wooden games: creating something is only the first part. People still need a way to discover it.

What I learned from building a game with AI

My experience changed the way I think about software development.

A person can now attempt projects that previously would have required skills they did not have, a team they could not assemble or a budget they could not afford.

That does not mean knowledge no longer matters. In fact, the more I learned, the better I became at explaining problems, recognizing mistakes and deciding whether a proposed solution made sense.

AI reduced one important barrier: it allowed me to keep moving whenever I reached something I did not yet know how to do.

There is another side to the story

There is also something slightly uncomfortable about how convenient working with AI can become.

Human collaboration involves schedules, disagreements, different personalities and sometimes frustration. AI does not get bored because you ask the same question again. It does not decide that your small puzzle game is no longer interesting enough to continue working on.

For an independent creator, that can feel remarkably comfortable.

But it also raises an interesting question.

I do not know the answer.

What I do know is that AI allowed me to create Beauty of Symmetry, a project that I might otherwise never have completed.

And we are still at the beginning of this technological change.

What comes next?

Beauty of Symmetry continues to develop, and I am still learning how to introduce it to new players.

I also already have ideas for another game.

Whether I am ready for another thousand hours remains a different question. :-)