We Are Not Watching the End of Software Engineering. We Are Watching Its Reinvention.

Lately, I have been listening to conversations that genuinely make me stop and think.

Two episodes from The Pragmatic Engineer did exactly that for me: the conversation with Steve Yegge, From IDEs to AI Agents, and the conversation with Grady Booch, The third golden age of software engineering – thanks to AI. Both are different in style, but together they describe something I find fascinating and honestly mind blowing: we are not just seeing better tools, we are witnessing a deep shift in the nature of software engineering itself.

What struck me is that Steve Yegge talks about the shift from coding as the center of engineering toward orchestration, experimentation, and working with AI agents, while Grady Booch puts this whole moment into a much bigger historical frame. Steve focuses on what is changing right now in day to day engineering work. Grady explains why this kind of transformation has happened before, just in different forms.

Yegge’s view is provocative, but difficult to ignore. He argues that coding by hand may gradually become less central, and that engineers will need to adapt to a world where AI tools do more of the direct implementation work. He also describes practices like rapidly generating many versions of a solution and selecting the strongest one, even calling this “slot machine programming.” He sees the IDE itself evolving away from being mainly a code editor and toward becoming a place for conversation, supervision, and monitoring of agents.

This idea may sound extreme at first, but I agree with the direction. Not because I think engineering is disappearing, but because I already see the center of gravity moving. The value is shifting from typing syntax to defining intent, framing the problem correctly, validating outputs, and combining systems into something reliable and useful. That is still engineering, but it is a different layer of engineering.

What I found especially powerful in Yegge’s perspective is that he does not describe AI as a simple replacement story. He describes amplification. He talks about a spectrum of AI adoption, from engineers barely using these tools to those running multiple agents in parallel, and he warns that many engineers are still near the bottom of that spectrum. He also highlights practical barriers, such as monolithic codebases being hard for AI tools to handle effectively. That point is especially important for large enterprises, where the future will not arrive evenly. Companies with fragmented legacy estates, weak APIs, and oversized monoliths will not gain the same advantage as teams built for modularity and fast iteration.

At the same time, Grady Booch brings a kind of calm wisdom to the conversation. He argues that software engineering is not dying; instead, it is entering a “third golden age,” one centered on systems rather than only algorithms or object abstractions. He frames AI coding tools as another rise in abstraction, not as the elimination of the profession. His line, “Fear not, developers,” is simple, but it lands because it comes with historical perspective. We have gone through these moments before. The tools change. The abstraction level rises. The work evolves. But the need for judgment, architecture, responsibility, and systems thinking remains.

This is where the two episodes connect beautifully.

Steve Yegge explains why the workflow is changing fast. Grady Booch explains why that does not mean the discipline is collapsing. One is describing the shockwave. The other is describing the pattern.

And I agree with both.

I agree with Yegge that many engineers are underestimating how quickly the working model is changing. AI is not just a faster autocomplete layer. It is starting to alter how ideas become products, how prototypes become production candidates, and how much parallel experimentation a small team can do.

I also agree with Booch that this does not reduce the need for serious engineers. In fact, it may increase the demand for engineers who can think in systems, understand tradeoffs, navigate complexity, and connect technology to human and business realities. Booch argues that the shift is now from apps and programs toward systems, and that AI can reduce friction so that engineers can spend more attention on imagination and what was previously not possible. That is not the death of engineering. That is a higher expectation for engineering.

What fascinates me most is that this industry change feels both revolutionary and familiar.

Revolutionary, because the practical day to day experience of building software is clearly changing. Familiar, because once again the field is moving upward into a new abstraction layer. Assembly did not disappear because engineering ended. It became less central because the stack matured. The same happened with many low level concerns that used to define who a “real engineer” was. Today, something similar may be happening with hand coding as the primary identity marker of the profession.

That can be uncomfortable. It can also be exciting.

For me, the right response is not denial and not blind hype. It is adaptation with depth. Learn the tools. Understand where they truly help. Recognize their limits. Strengthen system thinking. Improve architectural judgment. Get better at problem framing, verification, integration, and business context. That is where durable value will likely sit.

Booch closes with the idea that this is the moment to choose whether we fear the abyss or decide to soar. I think that is exactly the right framing. We are living through one of those rare periods when the profession is being redefined in front of us. That is unsettling, yes. But it is also one of the most exciting things an engineer can witness.

That is why these two episodes stayed with me.

They do not just describe new tools. They describe a shift in craft, in mindset, and in what it may mean to be a software engineer in the coming years.

And to me, that is both fascinating and mind blowing.