Joseph Ruscio's piece Write-Only Code landed on something I've been thinking about for a while: we're moving toward production code that no human ever reads. Not because we want to, but because the volume and pace make human review the bottleneck.
His argument is sharp. The engineer's role shifts from author and reviewer to constraint writer and systems designer. You spend less time shaping implementation, more time defining what must hold true.
But that shift assumes something I'm not sure we've fully reckoned with: that we know how to write those constraints.
A Different Aptitude
Designing systems that remain correct without human inspection requires people who think differently. Who can operate at higher abstraction. Who understand what AI can and can't do today — not six months ago. Who reason about failure modes without seeing implementation. Who stay close-hauled to the wind as capabilities shift monthly.
That's not everyone who could review code. That's a different aptitude, and it requires constant recalibration.
Which means the write-only code future isn't just a technical transition. It's an organisational DNA change. You need people who can make that jump and keep making it as the landscape shifts. And you have to rebuild how teams think and operate while AI evolves faster than traditional organisational change happens.
The Real Bottleneck
The bottleneck might not be trusting the code. It might be building the human capacity to design trustworthy systems — and maintaining that capacity as everything continues to accelerate.
Building cross-functional pods, deploying agentic workflows, restructuring around new ways of working: what becomes clear is that the organisational transformation is harder than the technical one. The humans are the constraint, not the models.
Read Write-Only Code by Joseph Ruscio →
Originally published on LinkedIn · View discussion →