I started on a laptop I wasn't allowed to touch.
I'm Mohammed Almahbashi, Head of Technology. I lead a team of engineers and technologists who regularly prove me wrong, and most of what I'm proud of at work, they built. I've spent twelve years learning that the hard part of technology is rarely the technology.
My father worked on his laptop every evening. I'd wait for him to leave it on the desk and go to bed, then sneak in and play Solitaire in the dark, one ear on the hallway.
Solitaire turned into the internet, and the internet into Facebook. Eventually he got a new laptop, and the old one was, let's say, left to us. I made sure of that.
Back then I was playing a farming game that made you wait a full day for your crops. I didn't want to wait. So I went looking for a way around it and found Cheat Engine, and with it the idea that a program's state lives in memory, and memory can be changed. The crops were ready in minutes.
I didn't know that was called hacking. I only knew the game trusted my machine, and that was its mistake. I've been looking for assumptions like that ever since.
Build first. Plan second.
Most technical leaders I know want to brainstorm, analyse and plan before anyone touches anything. I'd rather start with whatever idea is on the table, build the roughest version that works, and let it teach us. Add, remove, rebuild, until the prototype settles and the direction is obvious. Then we write the plan, and it's a far better plan, because it's based on something real.
The one condition is that it's safe to fail. Trial and error only works inside a sandbox, so the sandbox is the first thing I build. I still work the way I did on that laptop: try it and see what breaks. The difference is that now, breaking it is allowed.
For twelve years my work has spoken for itself inside the places I've worked. This site is the first time I'm letting it speak outside them.
I still write code. Less than I'd like, more than my calendar allows. This site is the first thing I've built for myself in a long time, and like everything else I build, it started rough and it will keep changing.
- Bringing seven-plus companies, each on its own ERP, into one system, without asking any of them to start over. The integration is the easy part. Agreeing on what a “customer” means across seven companies is not.
- Experimenting with loop engineering: designing the checks and stop rules that let an agent keep working without someone prompting it each time. The hard part isn’t the agent. It’s deciding what “done” means.
- Reading everything I can on governing coding agents: who reviews their work, who owns their permissions, and who answers when they’re wrong. It’s the oldest management question, asked about a new kind of teammate.
- Building this site in public. It’s the first thing I’ve made for myself in years.
