Growth In The Age Of AI: One Computer Scientist’s Reckoning

As a computer scientist, I have a big dilemma with AI. It has sharply increased my productivity and accelerated the pace at which I can innovate, design, and ship my abstract ideas into real-world products. But using it every day for the last two years has raised an uncomfortable question. For twenty-plus years, my skillset has run the whole gamut, ideation, research, prototyping, algorithm design, system engineering, software implementation, testing and evaluation, and my growth came from doing all of it myself. Figuring things out, getting them working, running the analyses, chasing the bugs: doing that hard work myself is what made me a better scientist. So what happens to that growth when the hard part becomes optional?

Reasoning through minute details of my algorithms to ensure they work correctly under all conditions. Chasing a null pointer dereference for days through the memory module of my file system in C, until at some ungodly hour it finally made sense. Digging through the API documentation of one ML library after another to be sure my model was doing what I thought. Thinking through every test case to cover all the ways my software could fail. That daily practice, and the friction that came with it, is how my skills sharpened and my understanding of computing deepened.

To be clear, my fear is not that AI is taking my job. On the contrary, I am now doing more than ever, exploring multiple ideas simultaneously, and converting ideas to products in hours or days instead of months. I could not be happier to stop poring over stack traces chasing memory bugs or reading confusing API documentation. It is closer and more personal than that. It is that AI slowly erodes my daily practice, the friction that made me good in the first place. So, the question for me isn’t whether I keep my job. It’s whether I keep growing.

My goal here is to explore this honestly for myself first and hopefully generate enough insights that may help others in my situation. I am not here to tell you AI will make us all obsolete, and I am not here to tell you it’s fine and we will all just move up the value chain. Both of those seem too simplistic in my opinion. I keep circling something more uncomfortable: the same tool that is making me more productive may be quietly eroding the very skills and understanding that made me good in the first place, and I can already feel it happening.

There are several angles to explore here. In this series I want to start with the question as it applies to me: a professional with two decades of experience watching the ground beneath my feet shift quickly. In a later article I will explore the harder version of it, i.e. what this means for the younger generation (my kids, 10 and 14) growing up with AI as a given, who never had to build the skills before the tool arrived to replace them.

My initial attempt at exploring this question quickly filled about 15 pages of a Word document, which would have created one boring long post. Instead, I decided to break it down into four posts, each exploring a topic connected to my main question. Each stands on its own but builds on the last.

Part 1: How AI is different. Why doesn’t AI feel like every other tool I have adopted in my 20+ years? I think it’s the first tool to automate the end-to-end reasoning itself, not just execution, and that’s what makes it different.

Part 2: Have we been here before? As it turns out this is not a new phenomenon. The evidence that automation erodes expert skill is measurable and comes from brain scans, cockpit studies, and medicine, and the timescale is collapsing from decades to months. (coming soon)

Part 3: What I’m Willing to Lose to AI, and What I’m Not. Not all atrophy is bad. The real task is telling the skills I can safely hand over from the load-bearing ones whose loss hollows out my judgment, often without my noticing. (coming soon)

Part 4: What I’m Doing About It. The practical stance I am taking to keep growing, and an honest accounting of how I have incorporated AI into my workflows, including to write this very series, and whether I stayed on the right side of my own argument. (coming soon)

My goal with this series is to understand, in real time, how AI is reshaping the way I grow. I have a clear view of the problem and some sense of what to do about it, but I don’t pretend to have it all figured out. If you are somewhere in the same boat, I would genuinely like to hear where your experience matches mine and where it doesn’t.