The Productivity Paradox: How AI Slop Trained One Engineer to Tune It Out

An engineer describes a new cognitive reflex: when a document at work carries the telltale fingerprints of a large language model, his brain simply refuses to engage with it. He recounts several examples — a design doc padded with Claude-specific phrasing and self-aware ‘reasoning,’ a marketing deck that pitches mundane features as revolutions, a requirements spec that buries a simple idea in verbose hedging. The common thread is content that is long on words and short on meaning, and the result is real friction: he ends up re-asking questions already answered in material he couldn’t bring himself to read.

The author pushes back on research suggesting humans can’t reliably detect machine-written text, arguing that low-effort output is easy to spot once you know the tells — inflated word choice, breathless sentence flow, and the compulsion to frame every trivial detail as a breakthrough. He likens his response to ‘banner blindness’: after being saturated with AI-generated LinkedIn posts, emails, and websites, his attention has learned to filter the pattern out automatically as a survival mechanism against information overload.

The sharper point is the irony. The same technology sold as a productivity multiplier has become a drag on his focus, because the flood of generated filler has trained him to disengage — a habit that now bleeds into legitimate work. He closes with a fitting image from a vacation: instinctively skipping a restaurant because its advertising photo, on closer inspection, appeared to show a moldy quiche, another piece of low-effort content his brain flagged and dismissed before he consciously noticed.