Back to Home

The Great Erosion: Are We Watching Our Developers Unlearn Their Craft?

By Robert Welch • January 10, 2025

A troubling pattern is emerging from the front lines of tech hiring, and it points to a foundational shift in how software engineers are developing their craft. I've conducted close to 50 interviews for my innovation department over the last couple of months, and I'm witnessing a subtle but persistent erosion of skills—a slow chipping away at the deep, analytical thinking that has long defined exceptional coders.

The Telltale Signs of a Divided Focus

The observation is not as overt as catching someone copying and pasting. It's more nuanced, and frankly, more concerning. During live coding sessions, I've noticed a recurring theme: rapid eye movement. The candidate's eyes dart back and forth between their main monitor and an unseen, external screen. Their coding process itself feels like a transcription.

They aren't building a solution from the ground up; they are meticulously encoding what appears to be a pre-existing answer. The code is written in a way I'd consider somewhat backwards, where it seems to error by default because it's being built from the top down rather than from a solid logical foundation. This behavior is often preceded by another curious red flag: a sudden, initial "difficulty" in sharing their screen, as if they need a moment to arrange their digital workspace out of view.

A Widespread and Concerning Phenomenon

This isn't an isolated trend limited to a specific type of candidate. I've taken care to make my ingestion of candidates as diverse as possible, sourcing them both independently and through various recruiting agencies. The fact that this behavior appears across such a wide spectrum of applicants suggests it's not an anomaly but an emerging norm.

It's Not About AI, It's About Blind Reliance

Let me be clear: this isn't a screed against artificial intelligence in software development. AI's power to debug complex issues or help a developer dive deeper into the "why" of a particular implementation is undeniable. It can be a potent tool for augmenting a developer's abilities.

However, the trend I'm observing is not one of augmentation, but of unthinking reliance. This blind transcription further erodes the essential knowledge of our software engineers. The act of wrestling with a problem, of breaking it down into its fundamental components and building a solution from scratch, is where true learning and mastery occur. By circumventing this critical process, we are creating a generation of developers who can replicate solutions but not truly comprehend them.

The danger lies in the long-term consequences. A field built on a foundation of superficial knowledge is a field ripe for stagnation. The ability to innovate and create truly novel solutions is predicated on a deep and intuitive understanding of the fundamentals.

The Path Forward: AI as a Tool, Not a Crutch

I remain a firm believer in the power of AI. It can be an invaluable partner in the development process. Imagine using it to rapidly prototype ideas, explore different algorithmic approaches, or handle the boilerplate code that can bog down a project. Used correctly, it can free up developers to focus on higher-level architectural decisions and the creative aspects of software design.

The key is to foster a culture of critical engagement with these powerful new tools. We must encourage aspiring and current engineers to use AI not as a crutch, but as a springboard for deeper learning. Let's champion its use for:

The great erosion has begun, but it is not irreversible. By recognizing the specific red flags and promoting a more mindful and intentional use of AI, we can ensure that the next generation of software engineers are not just transcribers of code, but true architects of the future.