Senior engineer publicly abandons AI coding tools citing environmental, societal, and mental health costs

The engineer’s experience traces a common adoption curve for AI coding tools. Early versions like GitHub Copilot were dismissed as distracting or error-prone, but later iterations, particularly agentic workflows, became capable of automating complex tasks. This progression suggests that the tools’ utility improved significantly over time, but so did their potential to displace human judgment. The ability to generate entire projects without manual code edits raises questions about whether such automation preserves or erodes engineering rigor

The decision to abandon AI tools was not based on technical limitations but on broader concerns. The engineer cites environmental costs, likely referring to the energy demands of training and running large models. Societal impacts may include job displacement or the devaluation of human expertise. Mental health concerns could stem from over-reliance on tools that reduce cognitive engagement or create pressure to work at an unnatural pace. These trade-offs are rarely discussed in product marketing but are critical for engineers evaluating long-term adoption

The engineer’s resistance to mandated adoption highlights a tension between top-down directives and individual workflows. Management’s enthusiasm for AI tools was framed as a competitive necessity, but the engineer found them incompatible with his preference for deliberate, test-driven development. This disconnect suggests that AI tools may not suit all engineering roles equally. Teams adopting these tools should consider whether they align with their existing processes or require a fundamental shift in how work is performed

The personal use of AI chat tools for advice, including medical and career guidance, adds another layer to the critique. While the engineer initially treated the AI as a supportive collaborator, the potential for over-reliance on unvalidated suggestions could have unintended consequences. This underscores the need for clear boundaries when integrating AI into professional or personal decision-making. For engineers, the risk is not just technical errors but the erosion of critical thinking when tools are treated as authoritative sources