Software Engineering Is Becoming Civil Engineering

PMs as code “welders”? Internet splits over bridge analogy

TLDR: A tech lecturer says AI lets product managers “weld” features while platform engineers design the safe structure—just like bridge building. Comments erupted with safety fears, AI fatigue, and analogy wars, as devs debated whether this split is a smart evolution or a shaky shortcut that could snap under pressure.

A guest lecture turned hot take: Christopher Meiklejohn says software is doing a 1700s glow‑up—split the roles like building bridges. In his view, AI helpers can let product managers (folks who decide what to build) “weld” features, while platform engineers design the safe, sturdy framework. He claims the split is happening now, not later—and the comments lit up.

The most jaw‑drop moment came from a darkly funny confession: one commenter wrote they’re a project manager writing code to keep airplanes from hitting things—“It’s been a rough couple years.” Cue nervous laughter and a thousand side‑eyes. Others were not laughing. One critic snapped that the post reads like an AI‑generated essay, adding to the growing fatigue with anything AI‑flavored. Another took issue with the framing itself, firing back that Computer Science—the academic study of computing—is the foundation, not “the welding.”

Then the analogy wars began. A commenter with degrees in both Computer Science and Civil Engineering argued comparing feature work to welding downplays the skill and domain knowledge needed to build real features. A more measured voice appreciated the thought but noted a key gap: civil engineering is all about site‑specific quirks, while software tries to avoid those. Meanwhile, the thread devolved into bridge memes, “who’s the real engineer?” debates, and a chorus wondering if letting non‑coders ship with AI is genius…or a disaster waiting to happen. Buckle up.

Key Points

  • The author argues software engineering is splitting into platform/system design and feature implementation roles, analogous to civil engineering’s separation of structural design from craft.
  • Product managers and non-technical collaborators can now ship features by describing requirements to AI tools (e.g., Claude Code), with human verification.
  • Platform engineers are responsible for foundational design: database schema decisions (e.g., multi-tenancy), safe deployment with automatic rollbacks, and abstractions to prevent feature changes from breaking critical flows.
  • The article maps civil engineering tasks to software: site selection/domain isolation, material specification (stack choices), load analysis (capacity planning/rate limiting), and inspection regimes (observability/code review).
  • Emphasis is placed on designing for capacity beyond expectations and on semantic observability that ties failures to recent feature deployments and affected users.

Hottest takes

"I'm the project manager who is responsible for writing the code stopping airplanes from hitting other objects" — the_real_cher
"I'm tired of reading AI blog-posts. Write in your own words, please." — saxelsen
"No. Computer Science (CS) is the basis." — slindsey
Made with <3 by @siedrix and @shesho from CDMX. Powered by Forge&Hive.