Technology, software, tech in general fascinates me. Every day I read the latest updates on AI, on the programming languages I use the most, including important packages. I do this because I find it fascinating, but also to keep on learning and to help me with my career.

And even though I'm constantly learning something new this way, not much necessarily changes in my work or in the technology I use. AI is an exception, but be honest: who wants to read about AI all day ?

"For my team, the cost of compute is far beyond the costs of the employees" - Quote by Bryan Catanzaro, vice president of applied deep learning at Nvidia [1]

I have to say, that's a ridiculous thing to say. Or is it?

The narrative we've been fed for years is that AI will make human workers too expensive to keep around. Turns out, at some organizations, it's already the other way around.

OpenAI made LLMs available to the masses. No question about it.

Before ChatGPT, AI felt like something for researchers and big labs. After ChatGPT, my parents, non-technical friends, and basically everyone started using it.

For a while, "just ask ChatGPT" became almost the same as "just Google it".

That is a massive achievement, but now the picture looks different.

ChatGPT is no longer alone

The quality gap has closed.

Gemini got better. Claude got better. Copilot got better. Open-source models got better.

Using AI to improve workflow and efficiency as a programmer has become a necessity to compete in today’s market. But just how good are LLMs on their own?

This is what the Remote Labor Index aims to measure (link to the study and paper).

Remote Labor Index (RLI), a broadly multi-sector benchmark comprising real-world, economically valuable remote-work projects designed to evaluate end-to-end agent performance in practical settings.

And it actually looks like a very high-quality study. They claim to have compared projects that cost over $140,000 and took more than 6,000 hours of real work with their AI-generated counterparts.

I've recently came across the article I Taught 6,642 Googlers THIS Productivity System by Jeff Su.

This is not a completely new productivity system for me. Instead, it’s something that I’ve been doing for years, at least in a very similar way. I’ve never really thought about defining this workflow or writing it down, but after reading about it, I actually took the time to reflect on it.

Definition: CORE

I am always interested in finding ways to be more productive, how to increase the performance of workflows and software, and how to improve in general.

Recently, LeadDev published their two Software Engineering Performance Reports for 2025, which of course I had to read.

I like how both reports are very compact and only focus on the most important results of their surveys. I can only recommend everyone to have a quick look.

Do you know that funny moment when someone talks about a topic and tries to sound smart, but you actually know more about whatever they’re talking about?

I usually get this feeling when talking to consultants or people in similar fields when they start talking about software development. I think they literally are the embodiment of fake it until you make it. Maybe after reading the following article by McKinsey, you’ll know what I mean.

On my blog, you’ve probably noticed how I outline my workflow to help you become the most productive developer you can be.

The truth is, however, that technical factors play a much smaller role in our productivity than we’d like to admit.

In the paper What Predicts Software Developers’ Productivity?, a wide range of influences on developer productivity are examined.And as it turns out, the top factors are:

1. Job enthusiasm
2. Peer support for new ideas
3. Useful feedback about job performance

After learning how AI and LLMs are built, after calculating backpropagation over several pages by hand, I still didn’t fully understand how they work or why we are doing things the way we are.

I just followed my university courses blindly, without actually asking myself what each step in a neural network is doing.

Then came a book I recently read: A Brief History of Intelligence: Why the Evolution of the Brain Holds the Key to the Future of AI by Max Bennett.

I'm a full-time software engineer, but writing code is only a part of what I do. Much of my time is spent thinking through problems, planning solutions, collaborating with teammates, and doing research. So when it finally comes time to write code, I need to make the most of every minute. That's why I focus heavily on optimizing my workflow: to turn ideas into working code as efficiently as possible.

LLMs take over programming. So that's it ? Years of experience, degrees and certificates are now worthless ? According to non-programmers that is in fact the case.

Experienced developers know the shortcomings though. The more experienced they are, the less stressed they are about it. As outlined in my last Post even the GitHub CEO states that programmers are not going anywhere.

But why is every company pushing AI so much ? Does it make us more efficient ?

AI, probably the buzzword of the last few years, is at the forefront, for better or worse. Big tech wants us to believe that using their LLMs and coding agents makes us so much more productive, that we don't even need many of programmers and software engineers anymore. With the click of a button, everyone can create prototypes, full-scale apps, or even SaaS. While, in my opinion, LLMs do change the industry in some sense, not everything can be replaced by them, and humans are still the backbone of the development process. Much can be said about the quality of LLMs and coding agents: security issues, code quality, cost, dependence, etc., in this article I want to outline the point that is most important to me, personally, that is: learning.