Jules GAY--DONAT | Systems & Engineering

The Era of AI-Assisted Development: Why Coding Is No Longer Enough

Sun, 19 Jul 2026 00:00:00 GMT

An Indispensable Tool for Competitiveness

In the software industry, a debate persists: is using artificial intelligence to generate, audit, or restructure code a form of "cheating" or a natural evolution? Stubbornly hand-coding every single line has now become not only a misplaced act of resistance but a dangerously counterproductive stance.

The use of LLMs (Large Language Models) and autonomous agents makes it possible to automate boilerplate code, drastically speed up testing phases, and simplify massive refactors. Refusing to leverage these tools is equivalent to a mathematician forbidding themselves from using a calculator to solve a complex equation system. To remain competitive—whether as a freelancer, a startup, or a tech giant—AI-assisted development must become the absolute standard.

The End of Programming as a Sole Competency

For decades, the "ability to program" (knowing the syntax of a language, knowing how to implement a loop or a SQL query) was an end in itself. With generative AI, this skill has become heavily commoditized. Writing raw code is no longer the true bottleneck of software engineering.

If anyone can generate a Python or Rust function with a simple prompt, syntax inherently loses its value. What gains value instead is architecture and true engineering.

The Rise of the Systems Engineer

In this new paradigm, the cardinal skill shifts toward two extremes: high-level abstraction and low-level understanding. Tomorrow's developers will no longer be mere "coders," but system architects and critics.

True experts will distinguish themselves through:

A deep understanding of algorithms: Knowing how to evaluate the time and space complexity of AI-generated code and identifying its mathematical or logical flaws.
Systemic vision: Knowing how to make services interact flawlessly, guarantee the security of an entire infrastructure, and optimize data flows at scale.
Low-level mastery: While AI excels at generating high-level boilerplate, we will desperately need engineers capable of diving into memory management, multithreading, assembly, or system calls to solve critical performance bottlenecks that models cannot comprehend.

Embracing AI in our daily work does not mean abandoning our craft; it means rising above simple syntax to focus on what makes computer science fascinating: true engineering.

Artificial Intelligence, Ecology, and the Sovereignty Gamble

Sun, 19 Jul 2026 00:00:00 GMT

The False Dilemma of Data Centers

Lately, a recurring narrative has taken root in public debate: Artificial Intelligence consumes too much energy, and building new data centers is an ecological heresy. While concerns about energy consumption and carbon footprints are fundamentally legitimate, the systematic rejection of these infrastructures hides a much more insidious danger.

We are treating AI as if it were a mere digital convenience or a passing trend. In reality, Artificial Intelligence is rapidly becoming one of the most critical strategic resources of the 21st century.

AI as a Sovereignty Resource

Refusing to build our own AI infrastructures in the name of environmental sobriety is tantamount to outsourcing our cognitive, technological, and industrial future. If we do not possess our own models and compute power, we will be forced to rely on those of others.

Not possessing this resource can cause extremely serious problems in the future:

Loss of competitiveness: Companies that fail to use AI will be swept aside. Those relying on foreign AI will be entirely dependent on the pricing and terms of service of external entities.
Interference and cultural bias: AI models reflect the values, norms, and sometimes the propaganda of their creators.
Censorship and technological blackmail: We already had a taste of this with the recent "Fable" outage, where dependency on a single provider exposed our sheer vulnerability. What will happen when AI is deeply integrated into our critical infrastructures?

A Mandatory Investment

We do not have the luxury of waiting. Investing in Artificial Intelligence and building the data centers needed to run it is mandatory, whether we like it or not.

This is not about abandoning ecology, but rethinking it within the framework of technological realpolitik. Rather than rejecting these infrastructures, our challenge must be to make them as green as possible: optimizing water consumption, utilizing low-carbon energy, and recovering waste heat to warm our cities.

Surrendering our technological sovereignty today for symbolic ecological victories means ensuring we have absolutely no say in the future of tomorrow's world.

Why I'm passionate about computers and hardware

Mon, 20 Jul 2026 00:00:00 GMT

It all started when I was 5 years old. My dad bought me my very first PC, a modest machine running Linux. Even before I could read properly, I was completely hooked. I literally haven't let go of a keyboard since.

At age 11, I reached a milestone for any tech enthusiast: using my own savings, I built my first PC from scratch. The specs were humble, but it was my machine:

An AMD Ryzen 5 4600G CPU (the integrated graphics were a lifesaver, as I couldn't afford a dedicated GPU).
16GB of DDR4 RAM.
A rather low-end MSI motherboard.
And the famous "Amazon mistake" that worked in my favor: I received a 1TB SSD instead of the smaller one I ordered!

Since that first build, my PC has become a continuous project. Thanks to successive upgrades gathered every Christmas and birthday, the machine has slowly transformed. Today, I work (and game) on a Ryzen 5 7600X paired with an RTX 5070 and 32GB of DDR5. It's safe to say it's come a long way from the little 4600G.

The love for Linux and the Dual Boot burden

Having started on Linux at such a young age, the OS holds a special place in my heart. Today, I'm an avid Arch Linux user (I use Arch, btw). I love the absolute control, the lightweight footprint, and the elegance of the ecosystem.

Yet, a necessary evil persists: the Windows dual boot. The culprit, of course, is gaming. Even though tools like Proton have made giant leaps, some titles (especially those with invasive anti-cheats) still force me to keep a Windows partition on my drive. It breaks my heart a little, but it's the price to pay to play...

The family "Tech Guy"

My passion for systems and hardware comes with what I jokingly call the IT curse. I've inevitably become the official PC repairman for my friends and family. A stubborn printer? A Windows install that won't boot? A virus on a laptop? That's on me. I do it often—maybe a bit too often—but deep down, it's what taught me patience, how to diagnose unlikely hardware failures, and how to tinker with motherboards.

This relentless curiosity about hardware is what pushed me toward systems engineering. Understanding exactly how hardware and software interact has become an everyday learning process for me. And it all started with a little Linux machine, 11 years ago.