AI and the Human in the Loop: What Actually Gets Replaced | IT Guy Show Ep. 022

Episode 22 of The IT Guy Show is a milestone for me, and not just because the number is round. Karl Abbott was my very first guest back on episode one, and he came back for what turned into one of the best conversations I have had on the show.

Karl is a product manager at Microsoft working on Linux vendor partnerships, a university regent, and the host of his own podcast on the human side of product management. He wrote a LinkedIn post called “AI Didn’t Replace My Work” that caught my attention and kicked off this whole episode. The short version of his argument: AI removes the mechanical friction, but it does not carry the consequence. Judgment, taste, and responsibility stay human. When the execution gets cheap, the quality of your thinking becomes the actual work.

We got into vibe coding, what it actually demands from engineers versus what people assume it removes, and why bad assumptions fed into an AI system come out looking clean and confident on the other side. We also talked about higher education, where Karl made a point that stuck with me: if students are not building judgment through struggle and repetition, schools need to be very intentional about where that formation is going to come from. I shared some of my own experience teaching Linux administration at JCCC and why I tell my students that using AI to fill out a lab is cheating themselves, not just the system.

There is also a good stretch near the end on book recommendations, including a few I had not heard of that I am now actively tracking down.

Watch the full episode on The IT Guy Show: A Human in the Middle | Episode 022

This one fits squarely in the middle of what I think about most: how people who work in technical fields stay sharp and relevant as the tools keep changing underneath them.