AI’s Double-Edged Sword: Speed vs. Discipline in Software Engineering, According to Linas Beliūnas

In a recent LinkedIn post, Linas Beliūnas offers a compelling perspective on the role of Artificial Intelligence in software development, cautioning against prioritizing speed over disciplined engineering practices. Beliūnas, who identifies as a non-coder, draws a stark contrast between AI-generated code and what he terms “actual engineering,” urging professionals to maintain rigorous development standards even as AI tools accelerate the process.

The Perils of Unchecked AI-Driven Speed

Linas Beliūnas highlights a potential pitfall of modern AI tools: their ability to generate code rapidly, which can inadvertently foster technical debt. He suggests that the left side of an implied visual (though not provided in the text) represents a chaotic, AI-driven approach, stating:

The left side looks exactly like what happens when you let AI dump code until something vaguely works and nobody can explain the wires later.

This imagery underscores Beliūnas’s concern that a reliance on AI for quick code generation, without proper oversight and understanding, can lead to complex, unmaintainable systems. As Beliūnas argues, the true cost of such an approach becomes apparent when the code needs to be revisited or expanded upon.

Engineering Discipline as a Counterbalance to AI

In contrast to the potentially messy output of unchecked AI, Beliūnas outlines the core tenets of what he considers sound engineering practice. He presents a clear, sequential process that prioritizes understanding and quality over sheer velocity.

The Pillars of Engineering Excellence

According to Linas Beliūnas, the essential steps in robust software engineering include:

• Diagnose the problem.
• Plan the structure.
• Write it.
• Test it.
• Review it.
• Ship something you can still maintain next month.

This structured approach, Beliūnas implies, is crucial for long-term project health and maintainability. He acknowledges the undeniable speed benefits AI offers:

Today, AI definitely makes you dramatically faster.

However, he immediately follows this with a critical caveat, emphasizing that speed alone is insufficient and can be detrimental if not guided by discipline. Beliūnas posits that the efficiency gains from AI can be a double-edged sword.

The Escalating Cost of Bad Judgment

Linas Beliūnas’s central argument revolves around the concept that increased speed, when divorced from sound judgment and engineering principles, magnifies the consequences of poor decisions. He elaborates on this point:

But speed without discipline just means you can create technical debt faster too.

In Beliūnas’s view, the faster code is produced, the more expensive and difficult it becomes to rectify errors or structural weaknesses that arise from a lack of careful planning and testing. This perspective serves as a vital reminder for development teams integrating AI tools, urging them to implement robust processes that complement, rather than bypass, established engineering best practices. While AI can undoubtedly accelerate development cycles, Beliūnas’s insights underscore the enduring importance of human oversight, critical thinking, and disciplined execution in building sustainable, high-quality software.