Born a month before the Moon landing, I’ve spent 56 years watching the world shift from vacuum tubes to LLMs. While others are chasing the latest JS framework, I’m busy rebuilding an 1859 mudbrick house in Blender and writing a WASM compiler that feels like Z80 assembly. This is not just a memoir of a GitHub Copilot Challenge winner; it’s a manifesto for the "Vibe Archeologist"—the bridge between the raw silicon of the past and the fluid AI of the future.
https://dev.to/devteam/congrats-to-the-github-copilot-cli-challenge-winners-2240
1969
It is high time I started writing my memoirs. I don’t know if winning this challenge means I’ve dug my own grave or found the key to a brighter future, but understandably, I choose to remain optimistic. To put it bluntly: I was born exactly one month before humanity landed on the moon (1969). I don’t feel old, but saying "56 years" out loud... well, it’s quite a lot.
The story is fascinating in hindsight because it began in those heroic days before the internet, back when I felt I was born on the "unluckier" side of the planet: in Hungary, on the wrong side of the Iron Curtain. At the same time, it had its perks. As a child, there were no gadgets to steal the magic of a childhood painted with drifting clouds.
My relevance with computers started with a Scramble arcade machine installed in a local pub. At 11 years old, it took real courage just to enter that "csehó" (dive bar) with my friend. That same year, I was lucky enough to see Star Wars: A New Hope in the cinema. I was absolutely stunned by the opening scene as the Star Destroyer thundered over our heads—we didn’t need 3D; the experience was perfect.
All of this had an impact. With grid paper, scissors, and markers, I began crafting tabletop strategy games that didn't exist in their current form back then. My friends and I enthusiastically fought galactic wars with hundreds of paper ships—the only danger was my mother opening the door too quickly, sending our entire fleet scattering across the room.
Eventually, I learned the basics of BASIC from an early book that was more of a general guide to computers, only touching on programming as a side note. Fortunately, at a Budapest International Fair (BNV), I found a retro mainframe at an exhibition booth where I could actually sit down and run some code. That was the first time I felt what it was like to program.
Later, after dropping out of university (I was supposed to be a silicate chemical engineer), I bought my first computer: a Videoton TV Computer. Back then, I used a cassette tape recorder for storage and wrote games in Z80 assembly. In those days, even multiplication was considered a luxury.
Forgive this lengthy memoir section; I just wanted to introduce the times when computing was far from being as trivial as it is today. But now, with a hyperspace jump—passing through owning my own RPG shop and developing physical CCG card games—I eventually became a professional programmer.
...years passed, until...
2024: Dawn of AI
AI kicked the door down, and I couldn't avoid it. My first year was about diving into image generation. I tried to create something unique, but after a year and 60,000 images, I tossed the whole thing in the bin. However, I learned a lot about prompt engineering—how crucial it is to express ourselves precisely. This applies to any form of prompting.
But the real test for me is when I get an LLM to be a Tabletop RPG Game Master while I play. I believe this is a true measure of how well a model can think complexly and play a role within a consistent world. I do this test periodically; it's practically my only "prompt-based game." I highly recommend it: single prompt RPG - boring long full conversation.
Returning to the "peak climb," in 2024 I started entering challenges: A.L.I.E.N. S.O.L.I.T.A.I.R.E.. This was my first, where I tested how well my JSDoc-based, type-supported React state handler npm module worked. I later moved away from it because I began shifting toward a "no framework, pure code" direction. Regardless, that 65-line JSDoc module is truly useful, though these days I find even the useReducer React state management to be an excessive abstraction.
2025
After a disappointment in September, I used the "hair of the dog" tactic and signed up for two more hackathons. One was on Reddit, the other focused on AI extensions. Back then, I was still far from "vibe coding." Preparing for these was exciting. I poured a lot of energy into them, and in the process, a Markdown viewer was born... along with a minimal framework hidden within Flogon Galaxy.
https://dev.to/playfulprogramming/javascript-frameworks-heading-into-2026-2hel
Added to this was a flat 3D CSS game "engine" and an original sci-fi world. This led me to enter an AI video competition, where I visualized the backstory of my previous game in a 6.5-minute AI sci-fi video. During development, the protagonist (a punk-metal singer named Wora Shard) even got her own full music album (available on my YouTube channel). The story was very "round," and even though I didn't place in any competition, it did two things: it boosted my confidence regarding hackathons and solidified my commitment to minimalism.
1859
Eventually, that massive self-confidence crashed into reality. We were living in a rental, but I found the ruined little house of my dreams in Csobánka. It’s a traditional "vályog" (mudbrick) house, bleeding from several wounds, built in 1859 according to the cellar attached to it. We managed to snag it, and though it required a loan, we bought the cottage. Now, it’s just a matter of time and even more money before we can move in. For the plans, I started modeling the house in low-poly, using a "single texture" approach. This keeps the model around 4MB, so I don't have to struggle with material settings.
2026
Vibe coding has arrived. Today's tools are true partners in accelerating the development process. Now, the opportunity is before me to sift through my memories and dust off a good few ideas to implement them with a fresh approach.
Vibe Archeologist
Of course, I can't stop, because different forms of creativity keep launching me onto new paths. My next plan is to reach back into the world of Z80 assembly, but with the aim of making WebAssembly (WASM) programming tangible for people. I'm creating a language that compiles to WASM but looks like assembly, where you can see exactly what is happening and why in the debugger in real-time. Because I believe we must know the foundations—while providing an experience for programmers where hex numbers still matter.
Top comments (0)