Machine 'Eyeball Kicks' Flatten Literature

Machine‑generated fiction often optimizes for immediate stylistic 'wow' moments — what this author calls 'eyeball kicks' — by recycling a small set of flashy metaphors and structural ticks. That optimization is driven by model limits and reward signals (RLHF), producing output that reads literary at first glance but collapses on reflection and repeats across stories. — If mass AI writing adopts and normalizes these cheap attention heuristics, it can shift public taste, lower incentives for human literary craft, and alter what publishers and readers reward.