This is a test post — and I mean that in the most honest way possible.

In the current world AI agents can already write code in minutes, review pull requests, propose architecture, and catch bugs you missed at 2am. And this ability is only going to improve.

Maybe one day they will think the way engineers think. But here's the real question — not can they, but should we hand them the wheel entirely? I don't think so.

You can discuss with a machine, learn from it, use it to pressure-test your ideas. But handing over creative decisions to something running autonomously, without your judgment, your context, your skin in the game — that's where I draw the line.

Stand at the base of a 25-storey building and look up. You see glass and steel. You don't see the foundation, the load paths, the years of decisions baked into every beam. The building looks simple from outside. Code is the same. An app that handles a million users isn't magic — it is decisions, stacked.

Think, what happens when:

  • Two people try to buy the last item in a sale at the exact same moment — who gets it, and how does the system decide without charging both?
  • A user in India schedules a meeting for "9am tomorrow" and the person they're meeting is in the USA — whose tomorrow is it?
  • Someone uploads a profile picture that is actually a carefully crafted file designed to crash your image processor?
  • Your database has grown so large that a query which took 10ms six months ago now takes 4 seconds — and nobody changed the code?
  • A food delivery app is halfway through confirming your order and the restaurant's system goes down — is your money gone, is the order placed, or neither?
  • A social media platform tries to show you a feed of people you follow, but you follow 50,000 accounts — does it load instantly or does it make you wait while it reads every single post?
  • You fix a bug by changing how you store a user's age — and 1000000000000000000 existing users still have their data in the old format?

None of that is visible when you tap a button. But someone had to think through every one of those scenarios — not just make it work on a good day, but make it hold under pressure, at scale, in conditions they couldn't fully predict.

This is what makes engineering a creative field. Similar to how a painter is creative — a painting has colors, shapes, and meaning all at once. The colors catch your eye. The shapes give it structure. But the meaning is what stays with you after you've walked away. A system is built the same way. The interface is the color — what people see and feel. The architecture is the shape — how the pieces fit and hold together. The meaning is what it does for someone at 2am when they're trying to send money home, track a delivery, or just find an answer. Engineering is the act of holding all three at once — and making decisions that serve all three, not just the one that's easiest to see.

That's what this blog is about. I write about systems that work at scale — systems that move millions of dollars, serve millions of people, and quietly hold up the world as it is today. Every time you pay someone, order food, send a message, book a flight, or check your bank balance — there is a system underneath that is doing something extraordinarily difficult and making it look effortless. The engineers who built those systems made thousands of decisions you will never see. Some of those decisions were brilliant. Some were mistakes that cost companies millions to fix. All of them were made by people — not by machines running on their own, but by engineers who sat with the weight of the problem and thought it through.

I write about those decisions their proven methods and strategies, The philosophy behind them, the failure modes, the cases nobody saw coming and about anything else I feel a motivated engineer would find interesting to read. In plain language, so that anyone who writes code can understand how systems work — and build them as a free person, not just a pair of hands following instructions.