Hidden Risk - Everyone’s A Software Developer

There has been a surge of articles like this one, and as a Quality Engineer, I can imagine how our community must feel. Competent developers know better than to get too worried over this type of nonsense, especially if they’ve genuinely put these tools to the test. Yet unskilled and novice developers will take this hook, line, and sinker and begin to panic that their jobs are doomed for the scrap heap.

Will AI replace a developer's role? Well, it’s possible somewhere down the road if you specialize in “Hello World” as your development aspersions. I’ve been putting a couple of these tools (Copilot & Gemini) to the test in this area (I’ve developed software for over 40 years), and the results were not totally surprising.

What you need to keep in mind at this moment is that AI sells. It’s getting a bit worn out already, but it still sells, especially for those lacking in-depth experience developing complex software. Can you get it to write a single-page application or an Excel macro? Well, yes, this is pretty easy and not what I would call complex software development.

I’m working on a more extensive article documenting a two-month project in which I truly tested AI. If you think about what auto-tune has led to in the music world, where any individual with a PC and software suddenly now can make music, AI and software development, I suspect, will follow a similar trajectory. Remember when you write that first webpage or your first JavaScript function? It seemed incredible, and you might have considered yourself a developer for the first time. It’s likely you soon realized you were scratching the surface of what could be, and each successive project led to more complex problems and challenges.

If I had one overarching takeaway from turning AI completely loose on developing a software application, it was this. We are on the cusp of some of the most significant financial and information breaches in the history of software development. The article I’ve included here actually gives a hit to why that’s about to become a reality. Non-developers developing software. Nothing about great architecture has been mentioned in any of these articles. I watched my AI tools consistently introduce scaling and architectural issues in their attempt to solve development challenges. I reached a precipice with Copilot and Gemini reasonably quickly, and it did not know how to move forward. I was forced to come in and untangle the mess of code that lacked good design and sustainability.

I found the most success with these tools when I collaborated through the challenges of the design. When developing class structures and vocabulary around the design, describing objectives was apparent between the intent of the design and what resulted from the collaboration. My AI tools often apologized for suggesting poor architecture choices when I pointed them out. Here is the reality of having unskilled individuals suddenly “developing” software because they could do it themselves. They won’t have a clue about what is being created. “If it works, it must be right” will likely become commonplace.

My advice to developers is to learn how to collaborate with these tools. The real winners will likely be those who know how to leverage AI like any other development tool in their toolbox.

My advice to QA Engineers. Learn how to use these tools for analysis. Can it help you write a test case? Yes, but why stop at such a simple task? I suspect now will be the time when great QA engineers will be invaluable. You are about to encounter some of the most detrimental defects of your career. Become comfortable and well-skilled at wielding these tools. Craftsmanship will become more crucial than ever in our field. Be prepared for the barrage of AI QA gimmicks that will surface, and continue to be risk-minded.

Let me hear your two cents on this topic.

https://www.semafor.com/article/01/15/2025/replit-ceo-on-ai-breakthroughs-we-dont-care-about-professional-coders-anymore