Grady Booch, co-creator of the Unified Modeling Language and Chief Scientist for Software Engineering at IBM, argues we are already inside the third golden age of software engineering, not approaching the end of it. The first golden age centered on algorithms, roughly 1940s to 1970s. The second centered on object-oriented abstractions, 1970s to 2000s. The third is about systems, and it predates the AI boom by years, rooted in the shift from individual components to whole libraries, platforms, and packages.
The Pragmatic Engineer podcast host Gergely Orosz sits down with Booch to put current AI automation claims into direct historical context. Booch's core argument is that systems thinking, human judgment, and accountability cannot be automated away, and that every prior period of rapid tooling change produced both real progress and badly inflated expectations. The episode is worth reading in full because Booch provides 10 specific observations, not just a general rebuttal to AI doom narratives.
The open question Booch leaves on the table is whether the constraints shaping this third golden age, technical and human, will produce the same pattern of consolidation and new abstraction layers that defined the first two. That question has direct consequences for how engineering roles, responsibilities, and org structures evolve over the next decade. The full transcript and timestamped episode are available on YouTube, Spotify, and Apple Podcasts.
[READ ORIGINAL →]