AI instruments are increasingly becoming an indispensable component of programmers’ daily work, yet a segment of engineers argues that rather than speeding up development, this leads to a degradation of fundamental skills and a shift in the developer’s role toward merely validating code generated by others.
The integration of AI into developers’ routines is accelerating: corporations are openly announcing increased investments in automation, and internal team structures are emerging to monitor model usage, sometimes including internal metrics ranking employees based on the volume of “AI tokens consumed.”
This transformation is particularly pronounced within major tech firms, where developers are increasingly moving away from authoring code from scratch, transitioning instead into the capacity of reviewers for AI-generated code. In certain instances, according to employee accounts, these tools are incorporated into workflows virtually as the default setting.
Against this backdrop, some engineers voice concerns about a gradual erosion of foundational competencies. Instead of engaging in consistent coding practice, they spend time on repetitive checks of model outputs, which they claim diminishes their capacity for independent design and implementation of intricate systems.
Similar grievances surface in public discourse and on social media: developers report that relying heavily on AI aids makes recalling and recreating even familiar technical solutions increasingly difficult. One engineer characterizes this as regression to a skill level predating their professional career, despite years of experience.
Other developers, though not experiencing direct mandates to use Large Language Models (LLMs), routinely employ tools like Cursor. They note that while these tools save time, they subtly reorient the thought process toward external prompting rather than autonomous analysis.
Some experts frame this situation as a cognitive “outsourcing of thought”: crucial decision-making—both critical and architectural—is increasingly delegated to the model, which they believe risks shallowing problem comprehension when designing complex architectures.
Preliminary studies and observations in related fields are already examining the potential effect of sustained AI reliance on specialists’ critical thinking and skill retention, although the scope and enduring implications of this phenomenon remain undetermined for now.
About the Author
