Hello World!
I think it's worth sharing how this all started.
At first, it was just an experiment. I had early access to Moonlake AI, a new tool designed for creating virtual worlds — mostly in 3D. But for me, it quickly became equally great for 2D games and a useful playground for practicing game design skills.
So I decided to test a simple idea: could Moonlake help me create a small arcade game just by prompting?
I wanted to make something old-school, inspired by the classics I grew up playing on the NES.
And somehow, that led to shooting ducks.
The first goal was simple: create a game where you aim, shoot, chase a better score, and immediately understand what is going on.
Moonlake turned out to be surprisingly capable, but it was definitely not a “one prompt and you get a game” situation. It could hallucinate (ducks with three wings do look funny), misunderstand instructions (did you know that birdsin ai generated game fly backwards by default), introduce strange bugs, or build systems that looked right at first glance but did not really work in practice.
The interesting part was that most of those problems could be fixed with one or two more prompts — plus testing and refining, of course.
After around 12 hours, the final V1 version of Pluck Em existed.
It was rough and had bugs, sure. But the core loop worked, shooting birds felt satisfying, it ran smoothly, and I would rate the overall experience around 6.5/10. Which, I guess, was not too bad for around 12 hours of work.
Also, what is worth noting is that even though V1 began as a simple duck-shooting experiment, it slowly turned into an arcade shooter with roguelike elements, varied equipment, a shop, upgrades, an online leaderboard, and a dynamic day/night mode.
Everything was generated from scratch with Moonlake AI.
The music was created with Suno.
At that point, Pluck Em was still an experiment — but it no longer felt like just a tech demo.
After creating the first version, I shared Pluck Em! on Reddit.
I still treated it mostly as a curiosity and wanted to see if people would find it interesting or worth discussing.
The feedback was much more positive than I expected.
That was the first moment when I started thinking that maybe this AI-made duck-shooting experiment could actually become a proper arcade game.
So I submitted V1 to the AI Festival 2026 by Runway.
That was the moment when I decided to take the project further.
Pluck Em is now being expanded into a fuller arcade shooter with more content, better balance, more polish, and more reasons to keep playing.
I'm still using AI to create it, and I won't hide that.
The upcoming demo will be the first real chance for more players to try the expanded version, and that is where I would love to hear your feedback.
Let's see where we can take this game together.