Welsh’s article explores how artificial intelligence (AI) developments may redefine the landscape of the field of software development and make traditional coding methodologies obsolete. Readers should find it interesting, as it forecasts the potential impact on job opportunities, including the shift in computing from manual programming to AI model training, that is currently being observed.
Welsh predicts that traditional programming will be mostly replaced by AI, where it will learn from data instead of following explicit instructions. Therefore, a shift in skills will be needed, such as from manual coding to model tuning, dataset creation, and analyzing interpretability. Currently, software engineering heavily focuses on writing explicit instructions that are carried out by computers. The rapid developments in machine learning and large language models (LLMs) are changing this. Tools like GitHub Copilot, Amazon CodeWhisperer, and many others are able to generate code quickly with simple instructions, which now require guiding AI through data management, quality control, as well as ethical considerations.
This visionary and thought-provoking article examines both theoretical projections and examples from ongoing developments in the industry to speculate about a future where programming, as we know it today, is obsolete. While the entirety of manual programming may not just disappear, such roles will continue to be limited going forward.
It is a valuable read for those closely associated with software engineering, such as programmers, researchers, and educators. It may also interest policymakers who want to ensure that policies stay current with the advancements in the field.