Author: coding agents increase the complexity of software engineering

In a note dated 24 September 2026, the author argues that using coding agents actually makes software engineering more complex. Although these agents can achieve impressive results, the author says exploiting their full potential requires extraordinary discipline and deep professional expertise. The note emphasizes that automation introduces new sources of error and maintenance challenges, necessitating stronger processes and testing practices, and highlights the importance of configuration, oversight, and compliance around the agents. This has implications for development workflows, team training, and software quality assurance, and implicitly signals risks to security and reliability if the necessary expertise is lacking. Overall, the author urges caution: while the promise of coding agents is great, successful adoption will require more structured procedures, better documentation, and a higher level of professional skill within the developer community.