Professional skepticism: the key skill for developers in the age of AI

The importance of professional skepticism

In the current technological landscape, dominated by the advance of artificial intelligence, a fundamental skill for developers emerges forcefully: professional skepticism. As discussed by David Burns, Head of Developer Advocacy and Open Source at BrowserStack, in an interview withStackOverflow, this critical approach is not only useful, but essential for navigating the complexities of modern software development. Professional skepticism, understood as the ability to question, verify and validateEverything is fineaspect of your work, becomes a shield against errors, inefficiencies and potential vulnerabilities, especially when integrating AI-based systems into existing workflows. The tendency to blindly trust new technologies or AI-generated results can lead to unintended consequences, making a methodical and critical approach essential.

This critical attitude is particularly relevant when addressing the challenges posed by the integration of artificial intelligence systems. AI, while offering enormous potential, can introduce new forms of complexity and unpredictability. Therefore, professional developer skepticism becomes an indispensable tool to ensure the robustness, reliability and security of the software. This is not distrust as an end in itself, but a healthy exercise in continuous verification, which leads to a more solid code and more effective solutions.

Apply TDD to Agentic Engineering

A key aspect discussed in the interview concerns the application of Test-Driven Development (TDD) principles to the field of agent engineering. Agentic engineering deals with the creation and management of AI-based autonomous agents, systems that can operate and make decisions independently. Integrating TDD practices in this field means writing tests before implementing the agent code, ensuring that every component and every expected behavior is verified from the beginning. This approach helps clearly define goals and expectations for AI agents, reducing the risk of unexpected or unwanted behavior.

Professional skepticism manifests itself here in the desire to rigorously test every hypothesis about how the agent works. Instead of passively accepting the AI's output, developers are encouraged to create test scenarios that cover a wide range of inputs and conditions, pushing the agent to its limits to uncover any weaknesses. This iterative process of writing tests, implementing, and verifying is critical to building reliable and secure AI agents.

Manage application state for robust testing

Another significant challenge mentioned is managing application state, a critical factor for successful testing, especially in complex contexts such as those involving AI. Flaky tests, those that sometimes pass and sometimes fail for no apparent reason, are often caused by poor state management. Professional skepticism drives developers to thoroughly investigate the causes of these unreliable tests, trying to isolate and control all the variables that can influence the outcome of a test.

This involves a deep understanding of how the application manages data, configurations and interactions between its components. Adopting rigorous practices for initializing and restoring state between tests is essential. In this sense, professional developer skepticism encourages a meticulous approach to test design and software architecture, ensuring that each test is run in a clean and predictable environment. The ability to solve flaky tests is ultimately a demonstration of competence in managing the complexity of the application state, an increasingly valuable skill.