/
Published on 04.02.2026
TLDR: Grady Booch argues we're in the third golden age of software engineering, where AI represents another rise in abstraction rather than the end of our profession. The shift is from programs and apps to systems, and engineers who understand complexity at scale will thrive.
Summary:
Let me start by saying this: every generation of developers has faced their existential crisis moment. When compilers emerged, folks thought we were toast. When object-oriented programming took over, same story. Now AI is here, and predictably, we're hearing the same drumbeat of doom. But Grady Booch – and this is a guy who co-created UML and has been shaping how we think about software architecture for decades – has a much more nuanced and frankly reassuring take.
Grady frames software engineering history through three distinct golden ages. The first, from the 1940s through the 1970s, was all about algorithms. Think about it – we were figuring out the fundamental building blocks of computation. The second age, stretching from the 1970s to roughly the 2000s, centered on object-oriented abstractions. This is when we got serious about managing complexity through encapsulation and modular design. The third golden age? It's about systems. And here's the key insight – this didn't start with ChatGPT or Cursor. It began when we shifted our focus from individual components to whole libraries, platforms, and orchestrating massive distributed systems.
What I find particularly compelling is how Grady responds to the "AI will replace developers" narrative. He's direct about it: current AI tools are trained mostly on patterns we've already seen. They're exceptional at automating known patterns – your standard CRUD operations, web-centric applications, the stuff we've done a thousand times. But the frontier of computing is vastly larger than that training data. When Anthropic's CEO suggests software engineers could be replaced in months, Grady essentially says – hold on, let's put this in context. We're an astonishingly young field. The term "digital" wasn't coined until the late 40s. "Software" came in the 50s. We're barely 70 years old as a profession.
Now, where Grady is brutally honest – and this is important – is about what WILL be automated. He identifies software delivery pipelines as low-hanging fruit. The messy, complex work of infrastructure and deployment? That's where agents can provide clear economic value. If you're primarily doing that kind of work, he's essentially saying: re-skill now. But for engineers who understand complexity at scale, who can manage human forces alongside technical ones, who grasp systems thinking? Demand is going up, not down.
The philosophical shift he describes is subtle but profound. AI lets you redirect attention from friction to imagination. Some of the constraints, some of the costs that used to eat up your time – they're disappearing. The question becomes: what will you do with that freed-up cognitive bandwidth? Grady's closing thought really lands: "You can either take a look and say, I'm going to fall into the abyss, or you can say, no, I'm going to leap and soar. This is the time to soar."
What's missing here? I'd push back a bit on the optimism. Grady doesn't fully address the economic incentives companies have to replace expensive engineers with cheaper AI tools, regardless of whether the AI is truly capable. He also glosses over what happens to junior engineers who traditionally learned by doing the "low-hanging fruit" work that's now being automated. If entry-level positions evaporate, how does anyone build those deep foundations he emphasizes? The Society of Mind by Minsky is a great recommendation, but reading about architecture isn't the same as building systems for years.
Key takeaways:
Tradeoffs:
Link: The third golden age of software engineering – thanks to AI, with Grady Booch