AI is the most transformative technology we've seen since the internet, and it's enabling companies to innovate at a pace that was previously unimaginable. But as we embrace this new era of AI-assisted development, we need to be clear about what this actually means for the future of software engineering.
It's not coding in the traditional sense. My role is closer to reviewing, auditing, and course-correcting.
I'll be honest: I haven't had this much fun with a computer since learning BASIC on my ZX Spectrum at the age of ten.
That excitement, however, comes with some important realities.
The limitations of today's coding agents make one thing clear: experienced software engineers are not being replaced—but their role is being reshaped. Engineers who embrace AI, understand its constraints, and actively work within them are becoming more valuable than ever. Those who don't risk doing their teams, their organisations, and ultimately themselves a disservice.
AI dramatically accelerates the production of code, but it does not replace judgment. The real work now lives in analysis, testing, configuration, DevOps, documentation, and architectural decision-making. These are areas where experience matters deeply, and where AI still struggles without strong human guidance.
The first 90 percent of an AI-assisted coding project arrives quickly and feels almost magical. The final 10 percent is slow and painstaking, involving detailed back-and-forth, careful validation, and iterative trial and error. Anything that requires real judgment, deep context, or architectural understanding still depends on humans to make the connections and steer the work in the right direction. Left unchecked, these limitations can cause a project to stall—or worse, quietly degrade.
Feature creep becomes almost irresistible. The novelty of building with AI encourages you to keep adding clever new capabilities instead of fixing bugs or refining what already exists. And while fixing bugs has always carried the risk of introducing new ones, coding agents amplify this risk. They can charge through a codebase making broad, goal-driven changes that unintentionally disrupt systems that were already working.
Used well, AI behaves like an extremely fast, tireless junior engineer. Used poorly, it becomes a chaos multiplier.
The engineers who thrive in this environment are those who know when to speed things up—and when to slow them down. They put guardrails in place. They review aggressively. They invest time in tests, deployment pipelines, configuration, and documentation. They understand that velocity without control is not progress.
Power tools allow more work to be done in less time. As always, the economy will respond by expecting even more productivity. The engineers who meet that expectation won't be the ones who simply generate the most code—but the ones who can reliably turn that code into systems that work, scale, and survive contact with reality.
The future of software engineering isn't about AI replacing humans—it's about humans learning to work effectively with AI. This means developing new skills: understanding AI's capabilities and limitations, knowing when to trust its suggestions and when to question them, and building systems that make collaboration seamless and productive.
Organizations that succeed will be those that recognize this shift and invest in their engineering teams accordingly. This means providing training, tools, and processes that enable engineers to leverage AI effectively while maintaining the standards of quality, security, and maintainability that distinguish professional software development from code generation.
The engineers who thrive in this new landscape won't be those who resist change or embrace uncritically—they'll be the ones who understand that AI is a powerful tool in their arsenal, not a replacement for their expertise. They'll use AI to amplify their capabilities, not outsource their judgment. And in doing so, they'll help shape the future of what it means to be a software engineer in the age of AI.