The core thesis: AI won't kill software engineering — it will expand it, driven by something called the Jevons Paradox.
The role was never about typing code. The job has always been about solving problems — reducing complexity, minimizing maintenance burden, and delivering something useful. Code was the medium, not the purpose. AI can handle the medium now, but not the judgment behind it.
Engineers become orchestrators. The era of the specialist coder — fluent in one language, deeply embedded in one stack — is giving way to the generalist orchestrator. Engineers will increasingly supervise fleets of agents that write business logic, generate tests, analyze logs, and suggest architectural changes.
This raises the bar, not lowers it. To manage agents effectively across large codebases and complex systems, you need to understand the underlying technology deeply — what good architecture looks like, how systems fail, where performance bottlenecks emerge, and when an agent's output is subtly wrong in ways that won't surface until production.
Three real hazards the author identifies:
- Replenishment — if agents absorb the work traditionally handled by junior engineers, the profession's apprenticeship layer is at risk, and eventually the pipeline dries up.
- Atrophy — many engineers using AI agents extensively describe a sense of skill erosion; when the agent handles implementation, you stop building the same intuition and muscle memory.
- Exhaustion — constantly switching between agent sessions, reviewing parallel workstreams, and maintaining coherence across semiautonomous systems creates a new kind of exhaustion that is productive but intense.
The economic argument (Jevons Paradox): As AI makes engineers more productive, the cost of building software drops — but demand won't stay flat, it will explode. Companies that once couldn't afford custom tools will build them, and problems that weren't worth solving will suddenly become economically viable. More efficiency → more demand → more engineers needed, not fewer.
The bigger picture: When the cost of implementation drops dramatically, engineers can think more broadly — prototyping physical systems, modeling complex processes, and solving problems in industries they've never worked in before.
It's a thoughtful, optimistic take from someone with genuine industry experience, though the hazards section is the most honest and arguably the most important part — particularly the concern about killing off the junior engineer pipeline that the whole profession depends on for its future talent.