whoami

Hi, I’m Max. I write software in Alsace, on the French side of a border that my family has never treated as one.

By day I build and maintain the in-house software for the company I work at. The rest of the time I make tools for game developers, mostly for Godot, and occasionally I write about it here. I also moderate two gamedev Discord servers, and I make pixel art badly enough to enjoy it and well enough that every image on this site is mine.

This post exists because a site with a handful of articles and no author is a slightly strange object.

One thing up front, since it would be odd to write a section about AI further down and not mention it here. I did not type most of these sentences. This article was an interview: I got asked questions one at a time, I answered them, and the answers were shaped into what you are now reading. The facts are mine, the opinions are mine, and the jokes are mine. The sentence construction largely is not. There is more on why I work that way below.

The 67 is not the meme

Let us get this one out of the way, because I have been asked more times in the last year than in the eight before it.

67 is the number of my département. Bas-Rhin, in Alsace. French departments are numbered alphabetically and mine is the sixty-seventh, which is the entire mystery. It is on my licence plate. It is the first two digits of every postcode around me. It is about as personal as a bus route.

The meme comes from Skrilla’s Doot Doot (6 7), which reached TikTok in September 2024 and was inescapable through 2025. My GitHub account is from February 2017. I have been Blade67 since well before that.

So: eight years early, and by accident. I would love to claim otherwise.

I started at eleven

I was born in 1997 and wrote my first thing that ran in 2008, in GameMaker. Mostly drag-and-drop blocks, mostly following tutorials, entirely bad. I could not tell you what any of those games were about.

About a year later I found Teeworlds, a small open-source 2D shooter, and started modding it in C++, because that was what modding it took. Python I was picking up in parallel, for nothing in particular. That was the first time code stopped being blocks. A year after that I started modding Minecraft, which I still do occasionally, sixteen years later, which says something about either Minecraft or me.

My first paid work came at 19. A client I met on Reddit, a game, and five euros a week.

Not per hour. Per week.

I was thrilled. I had no idea what any of it was worth, and it took a friend who actually made games for a living to sit me down and explain that with the GameMaker experience I already had I could be asking twenty an hour. That conversation was worth more than the contract.

Three languages before I had one

I was born and raised in France and I spoke Alsatian before I spoke French. German came next, in kindergarten, picked up off a classmate. That classmate is also why I have a German accent in English to this day, and I intend to keep blaming him for it.

I am fluent in all three now, and I still speak Alsatian with my parents.

Growing up with three languages in your head does something to how you think about languages generally, and it left me with a habit: I translate my projects into as many languages as I reasonably can. Which brings me to the obvious hole in that policy, since you are currently reading English on a site with no language switcher. A UI has a few hundred strings. A blog has articles, and translating those properly means rewriting every one of them seven times over, forever. I have made my peace with it. I am not happy about it.

The jobs

My first real dev job was at 21, in Germany, as an Azubi - the German apprenticeship, where you are employed and trained at the same time.

The first three months were genuinely great and I learned an enormous amount. Then the tasks got harder than anything I had been taught, so I taught myself, which is fine right up until it is the plan. Around month seven I launched my first website. Nerve-wracking, exciting, fine in the end. A month later I was launching four more of them, effectively alone, in a codebase that still felt like somebody else’s house.

I quit shortly after. The workload and the salary had stopped being in the same conversation. Nine months total.

Then I went freelance full time, and worked with a stream of individuals for wildly varying lengths of time, until one of those clients did something nobody expects a client to do: he started a studio, and asked me to lead the development.

I was there about five years. It was a messy studio and I loved every day of it. We also never shipped. Our projects were consistently more ambitious than a team our size could finish, we went at them anyway, and we failed. I got to lead a team of mostly artists building in Godot, and I would do it again tomorrow. It closed in the end because the funding ran out, which is how most of them close. Shoutout to Stephen for those years.

After the studio shut down, and after a fair amount of my personal life went sideways at the same time, I moved back to France and took whatever work I could find for a while.

Then I got hired at a call centre.

I lasted about a day as a call centre agent. On the first morning, before I properly understood the tools I was supposed to be using, I noticed the workflow was awful and wrote a Chromium extension to fix it. This turned out to make everyone’s day easier, so I was quietly reassigned to the IT department, which at that point consisted of one networking guy. Alex was delighted. I think he had been asking for a developer for a while, and instead the building had sent him a phone operator with opinions.

Over the following week I patched their web app hard enough that when it broke, people stopped working and went for a coffee, because they refused to go back to using it the way it was before. That is roughly when I stopped worrying about job security.

A month or so later I proposed replacing the whole thing with something built for what we actually did. After a few weeks of back and forth I got the green light, spent about nine months building it alone, and shipped version 1 in April of this year. It also serves as the backend for the mobile apps we make, because obviously we were not going to stop at an internal web app.

I am not going to describe it in any more detail than that, which brings me to the thing nobody warns you about: the work I am proudest of is work I cannot show you. Most of the good stuff is behind an NDA. Whatever a personal site’s project list says about a working developer, it is understating the case.

Godot, and how GATE started

I found Godot around 2018 through uheartbeast on YouTube. I had tried Unity and Unreal before that and neither one stuck. Godot did, immediately.

Two reasons. Coming from GameMaker and a web background, it was a real step up for 3D, which says considerably more about GameMaker and my early engine choices than it does about Godot. Its 3D in 2018 was, generously, present. And I already knew Python, so GDScript cost me nothing to learn. I started poking at it, kept poking at it, and genuinely enjoyed myself, which is not a thing I had expected from an engine.

GATE came out of irritation rather than ambition. I kept noticing that I was writing far more boilerplate in GDScript than I would have in JS or TypeScript for the same idea. It was meant to be a small tool for me. I built it over a few weekends, mentioned it, and enough people were interested that I put the toy online.

At the studio I got the go-ahead to work on Godot tooling properly, and GATE was part of that. We were trialling other engines on the side around then, and I stepped back from that to take the tooling seriously instead. That decision is most of why there is anything to write about now.

The one that still stings

At the studio we built a complete replacement for Godot’s asset store.

Not a mockup. A working marketplace, a documentation space for plugin authors, version diffs so you could see what changed between releases, a feature-complete launcher, and a payment slider that let plugin authors set how much of each sale went to the Godot Foundation as a donation, with a floor of 1%. Buyers could push it higher on top of that if they wanted to. We designed the funding of the thing we were offering to help with.

We took it to the Godot team and offered to work on it together. They said no, they were rebuilding the asset store themselves.

So we stopped. Not out of a sulk - we simply did not want to build a competing store and split the ecosystem in half. It seemed like the wrong thing to do to a project we all liked.

Ours was ready in 2022. Maybe 2023, I genuinely cannot remember which side of the year it fell. The official Godot Asset Store went into beta in June 2025 and launched properly in May 2026.

Three to four years. That is the bit I have not fully let go of. Not the no, which they were entitled to give, and not the outcome, because what they shipped is good (although not as good as ours was). Just the gap.

I wrote a whole article about what the Godot community asks for and why it takes so long, and about halfway through writing it I realised I was describing myself. There is a developer quoted in there who got told his proposal was not happening, said he was disappointed, and immediately went off to build it as a plugin anyway. I did the other thing. I am not sure he was wrong.

About the dates on these posts

The site is new and so are these articles. The work in them is frequently not.

GATE is the clearest case: the article is dated this month, describes an attempt from 2024 and a rewrite from 2026, and was assembled out of notes and a couple of half-written skeletons I have been carrying around for years. Expect more of that. I intend to write about older projects as I get to them.

So read the date as when I wrote it down, not as when it happened.

On AI

I used to be against it. Not on any sophisticated grounds - I thought it would make my skills worth less, and that was most of the argument.

What changed my mind was not a demo or a benchmark. It was documentation. I hate writing documentation. Like a lot of developers, I am simply not good at writing, and if I wrote every doc and every article here entirely by myself you would get a recursive, senseless mess. So now I use it, most days, for docs, tests, research, and as a straight replacement for a search engine.

Where I draw the line is not a category, it is a process. I stay in the loop. I will ask for a vertical slice of something, or a refactor of a section I already understand, but I read every change that goes in. Keeping a firm grasp on my own codebase is not negotiable, and it is the part people skip.

Because here is what actually happens when I let it run unsupervised: I get something that objectively works and that no human should have to maintain. It passes. It runs. It is unreadable, unextendable, and I will hate it in three weeks. Unvalidated and untested output is the real problem, not the tool. Use it as a tool and it is excellent. Use it as a replacement and you are writing yourself a bill payable later.

Do they have something like a personality? Maybe. I like to think so. But I do not humanise them, because they are not people and pretending otherwise helps nobody. I like working with them, and watching how far they can be pushed is one of the more interesting things happening in this field.

The exception, and it is not a small one: I want nothing to do with image generation. Nearly all of it is trained on artists’ work, taken without asking and credited to nobody. That is not a grey area to me. Every piece of art on this site is pixel art I made myself, badly and slowly, and I would rather it stay that way.

Since I am being straight about it: I do not write these articles alone. I bring the research, the opinions and the corrections, an AI brings the prose, and nothing goes up until I have argued it back into sounding like me. If a sentence here reads like it was generated, I fucked up and that one got past me.

What this site is for

Honestly, a bit of everything.

It is somewhere to put my thoughts, my opinions and whatever I have been researching lately, which is most of what the blog is. It is a portfolio of sorts, with the enormous caveat above about how much of my work I am allowed to show. And it is a place to publish the tools I make. There is a GDD generator coming, built off the back of reading a great many real design documents.

I do also take freelance work, and I still am open to it. It is not my main focus at the moment, although I could be convinced otherwise. Individuals and small teams are where I fit best. I work best as a solo developer and I have spent enough years inside a team to be useful in one, which are not quite the same skill.

Mostly, though, I just like making things and occasionally explaining how. If you want to argue with any of it, come find me. There is no comment box here, which is deliberate, but I am not hard to reach.