One of my classes that I remember most fondly when I was getting my degree was called "Logic in programming". This was at a Dallas County Community College and, frankly, was an absolute gem of a class that was worthy of an ivy league school.
The very first day our teacher gave us only task: solve the Tower of Hanoi puzzle and break it into precise instructions. I've always been someone who enjoyed puzzles and took to the task with enthusiasm. All semester, it was mostly those types of tasks as I slowly learned Visual Basic and C#. He was teaching us how to solve a problem; not how to code. I treasure that experience as one of my most formative as an engineer.
Continuing from my previous post Being a JavaScript Dev vs. a Framework Dev, we now get into even deeper pondering. I worry that what is happening with Javascript and frameworks, is also happening with engineering as a whole, thanks to the prevalence of AI. Many newer developers are no longer learning how to solve problems, but simply how to use a specific tool, resulting in an AI-first approach that gives false confidence about the difficulty of programming. Learning how to swing a hammer well doesn’t make you skilled at building a wooden shelf.
I recently heard a story from a friend. They have been occasionally mentoring someone trying to break into her first role in tech. She was lamenting that she seemed to reach a plateau in her coding skills and was struggling to debug something. With questioning from my friend, she admitted the root of the problem: she couldn’t debug without the use of AI and AI wasn’t helping her with her niche issue. Thus, she was absolutely stuck. This person was a full stack dev and had touched front and back end systems, yet she lacked the most important core skill as an engineer: the principles of problem solving.
This was shocking to me. It seemed inconceivable to be relying on AI so much. Yet, I don’t blame the aforementioned coder because technology as a whole seems to push AI as a panacea, especially to newer developers. That is where my problem lies. I do not disagree with using AI as a tool for occasional help, advice, organization or output amplification. I do disagree with using it to replace your own problem solving skills entirely.
AI isn't going anywhere, and in many ways, it will become an essential part of the workflow. However, our value as engineers will never lie in the ability to simply copy-paste a solution. It lies in our judgement. In our ability to critically assess an AI's output, debug it when it inevitably breaks, and architect a system from vague business requirements. If you don't know why a solution works, you can't be the one who fixes it.
If you are a newer developer, I urge you to treat your problem-solving muscle as essential. Don't outsource the struggle. When you encounter a puzzle, step away from the AI prompts at first. Take the time to sketch out the logic on a piece of paper, trace the data flow, and deliberately work through the problem without a safety net.
That ability to solve puzzles independently is the foundational skill that was taught to me in that Dallas classroom. In the current muddy waters of AI usage, it is still a skill that separates a durable, adaptable, and genuinely skilled developer from one who just passes the output of a prompt.
Invest in your logic first. The code will follow.
Top comments (0)