By André van der Walt, Solutions & Software Engineer at Insight Consulting. Blog post 3 of 4 in the Prompted series.
I would be caught lying if I said I don’t have some level of concern for the future of development. It’s not because I think developers will be replaced. Our responsibilities have already started changing and will continue to do so. My worry stems from understanding how I built the foundational knowledge I now rely on when I use AI to accelerate my work. I was handed hospital-pass problems that I wrestled with to solve and, in doing so, I built a fundamental set of skills and understanding of my craft. It taught me code, yes, but it also taught me a certain type of thinking, inquisitiveness and intrinsic understanding that now permits me to have the word “senior” in my title.
So, now that we have these AI tools and they have largely replaced those exercises (except in the most extreme cases but even those are getting fewer), when do we, and more specifically juniors entering the workforce, learn, practice, and hone these skills? Perhaps it is even prudent to ask if these skills will still be applicable in 10 years’ time. I believe they will be, but my confidence, even in that statement, is not rock-solid. Should we artificially hamstring juniors to teach them these lessons to ensure the pipeline doesn’t dry up? No business can afford to sacrifice productivity for that long.
Some people have said that the skills that now need to be taught and learned have changed – those that revolve around managing AI workflows while developing. I agree, but I do not think that can replace the judgement that gets slowly distilled in a senior developer over years of problem-solving. In my own case, inside a consulting business, I need to be close to clients to extract requirements, identify solutions and solve problems that often are not solvable with just code. Losing the path to that judgement and skillset that make me effective is simply not an option.
What is the answer? I don’t know. It’s very easy to get caught up in the “here and now” of the situation, focussing on how AI improves our work and output. Thinking ahead to when we need the next round of devs to start taking our place is uncomfortable; thinking about our own obsolescence always is. But it is necessary as that was exactly what was done for us.
I think of my mentor when I started my current role. The lessons I learned from him are exactly what I now rely on, and I don’t know how different my training would have been if AI was around when I started. This concern is not novel; many people are flagging it. Figuring out how to teach judgement to the next generation is our problem to solve, and we can’t code that away.
How do you teach something when the opportunities to learn it are disappearing?
This article first appeared in Insight Edge, our newsletter on LinkedIn.