Extended Duckmon Devlog: How 12 Months of Solo Game Dev Led to an Automotive Security Project
Duckmon has been on a short pause over the last month while I’ve been deep in the trenches on another project: a serious, patent-pending automotive security prototype.
Because it is tied to formal intellectual property filings, I can’t discuss the exact technical mechanisms publicly yet. But building it has pushed my engineering and problem-solving skills harder than anything I have ever tackled before.
And during this sprint, I had a major realization:
I wouldn’t have been capable of architecting or building this automotive prototype if I hadn’t spent the last 12 months grinding through Unity with Duckmon, and before that, TextZ-Ombie.
🧩 The Solo Journey: How Game Dev Became My Training Ground
Everything I’ve built over the last year has been done solo. No team, no mentor, no formal training, just stubbornness, curiosity, and a willingness to break things until they finally worked.
When I first opened Unity, I didn’t even know what a C# script was. I didn’t know what a component was, or a prefab, or a coroutine, or a state machine. I didn’t know how games were structured, how data flowed, or how systems communicated. I was starting from absolute zero.
My first attempts, especially TextZ-Ombie, were messy, chaotic, and full of dead ends. But every failed attempt taught me something. Every broken script forced me to understand a new part of Unity. Every meltdown over UI anchors or world-space canvases pushed me one step closer to competence.
Duckmon was where everything finally clicked. It forced me to build:
multi-state systems
dynamic UI pipelines
inventory logic
battle flow
data-driven card systems
modular architecture
clean separation of concerns
It was the first project where I wasn’t just using Unity, I was thinking like a Unity developer.
And that familiarity, the muscle memory of building systems, debugging interactions, structuring logic, became the foundation for everything that came next.
🛠️ Unity as an Engineering Sandbox
Just over 12 months ago, I opened Unity with zero coding background, no computer science degree, and zero game development experience.
Armed with an old entry-level laptop running an Intel Pentium Gold and 4GB of RAM, I tried an AI LLM for the first time and thought: "This is pretty cool. I bet this could help me build a game." I installed Unity and didn't look back, although I did quickly upgrade to 32GB of RAM or my laptop might have exploded by now.
When you build an indie game solo, you are forced to become a generalist overnight. You have to learn finite state machines, event-driven architectures, asynchronous data streams, memory hygiene, mathematical transformations, and user interface design. You also drink more coffee than should safely be ingested by any human in a single day.
It turns out that those exact same systems are the building blocks of real-world physical simulations and industrial engineering projects.
Over the past month, I used Unity not as a game engine, but as a full Software-in-the-Loop (SIL) simulation environment to model and test a complete automotive security feature from scratch:
A multi-node automotive gateway and hardware state machine
Custom real-time physical-layer signal processing and verification pipelines
Multi-device mobile companion applications communicating with the simulated vehicle
Real-world hardware-in-the-loop over-the-air testing across physical microphones and speakers
Complete fail-safe logic, anti-tamper cooldowns, and cryptographic audit ledgers along with full forensic logging
A comprehensive technical specification and patent dossier that is now over 350 pages long
Going from not knowing what a C# script was, or even writing a basic "Hello World", to watching a physical vehicle security prototype execute cleanly on real hardware with a 100 percent success rate has been a surreal experience. It is one of those moments where you just sit back and go: "Oh my god, it actually works."
🤖 Human Architect + AI Co-Pilot: Building at Scale
A massive part of making this possible was treating artificial intelligence as a true architectural partner rather than a shortcut.
For the past month straight, I’ve used Google AI Studio (specifically leveraging massive 1M plus token context windows) to manage the sheer scale of the project. Moving from using standard web chat interfaces to Google AI Studio (thanks to discoveries here on DEV.to) was a game-changer. Being able to feed an entire 27-module codebase, state-machine flowcharts, and a 350 plus page technical specification into a single session meant I could stress-test edge cases, audit mathematical logic, and ensure total consistency across the whole system.
I cross-checked outputs across various AI models to ensure the data was as accurate as possible, iterating continuously until the system functioned in my physical testing environment with zero failure rate.
The key takeaway for me: The human has to be the architect.
AI doesn't build complex systems on autopilot. You have to direct the vision, define the constraints, debug the physical hardware behavior, and know what questions to ask. Anyone can ask surface-level questions to an AI model, but you need to push back, cross-check, and dig deep. When you combine stubborn human execution with high-context AI tooling, a solo builder can produce the output of an entire engineering team.
🦆 Returning Home to Duckmon
With the heavy lifting on the patent documentation and simulation suite now stabilized, I am genuinely excited to get back into Duckmon.
Coming back to the game after building an industrial-grade simulation feels amazing. Duckmon will make me exactly £0 in profit, but I am building it as a pure passion project. Returning to it now, my code architecture habits are cleaner, my understanding of Unity's internal pipelines is ten times sharper, and I have a ton of fresh momentum.
Duckmon (along with TextZ-Ombie) was never just a card game project. It was the masterclass that made everything else possible. For me, skipping the endless tutorial hell and just getting stuck into building was definitely the way forward.
Top comments (0)