Over the past few years, since the introduction of LLM-based coding assistants into software development workflows, the practice of software development has been experiencing a significant revolution. Like the combination of DevOps, cloud computing, and agile development which radically changed the craft over the last two decades, the age of AI is yet again reinventing what it means to be a software builder.
The story is more complex, however, than simply predicting that software development jobs will be replaced with AI agents. The history of innovation has shown that these warnings are almost always exaggerated. What's just as clearly true, however, is that roles will evolve as technologists move up the value chain and replace toilsome tasks with higher-productivity workflows.
In addition, the prevalence of AI-based code generators ensures that the complexity of software will continue to increase, and therefore so will its risks, particularly given the poor security track record of AI-generated software thus far.
