Tech writers now judge AI prose instead of writing it

What happened

Tom Johnson interviewed Fabrizio Ferri-Benedetti about how technical writers now spend their time reviewing and judging AI output rather than writing from scratch. The conversation applied Italo Calvino's literary principles — lightness, quickness, exactitude — to explain what separates beautiful documentation from functional but graceless work.

Why it matters

As LLMs can generate competent documentation quickly, the human skill that distinguishes a seasoned tech writer is the ability to tell whether prose is light and clear or heavy and wasteful. An engineer named Siddhant Khare described 'AI fatigue' — the psychological toll of hundreds of small judgment calls per day replacing the creative flow of building something yourself. Not all writers experience this the same way; some find reviewing less exhausting than writing from scratch.

What to watch

The profession may be splitting into two paths: writers who create foundational content based on real experience (like DevRel), and writers who tend automated content production systems. Building agent skills that automate repeatable tasks like release notes creates compounding value as models improve.