Photo of Alex Arvanitidis
Alex Arvanitidis

Machine Learning Engineer

The luxury of letting ideas marinate

Published 27 days ago

The luxury of letting ideas marinate

In the past, I worked for companies where speed was everything. Deliver fast, deliver often. That’s how business works. I get it — money doesn’t wait.

But recently, I led a project called Jaqpot, where I implemented most of the frontend and backend myself. For the first time, I had full control. No pressure. Just time.

When a tough problem came up — something my brain couldn't fully hold — I paused. I left it alone and jumped to a task I understood well. I built that, delivered it. And days later, without trying, the first problem untangled itself. A lamp lit up in my head. It just made sense.

We’ve all seen this. As engineers, we know there’s usually one best solution. But when you’re under pressure, you patch things up. You make it work. You move on. Linus Torvalds has fought hard for the “best” solutions in Linux — because that’s what makes great software.

This was the first time in my career I could let ideas breathe. I now think it’s essential. A clear mind is the first step. So many solutions are right in front of us, but we can’t see them when our brain is cluttered. And then one morning, with a calm head, you fix it in 5 minutes.

In the age of AI, this matters even more. AI reflects what you give it. It doesn’t replace the thinking. It helps fine-tune, sure. But it won’t shift your approach. You need to do that yourself.

One day I spent hours tweaking an algorithm called Attentive FP. The results got worse. The next day, I changed my approach, and within 5' the model converged. A simple fix I couldn't see before. Because I was tired. Because creativity needs space. John Cleese said Monty Python relied on that space. I’ve seen it in improv too.

Time, calm, space — that’s what good solutions need.

That’s something I’ll carry with me in any project from now on.