Fable Took My Job

The piece is a first-person account of losing a job after leadership embraced large language models marketed as "Fable/Mythos" intelligence. The layoff did not come because the model outperformed the engineer’s craft but because the model was cheaper, giving managers confidence to cut costs and coast toward M&A. The author recounts prior experiences with budget-driven cuts, the immediate physical shock - ringing ears, drained color, a racing heart and sleeplessness - and the wider ripple effects: multiple peers in a DevOps group and employees at major tech firms receiving notices, awkward severance rituals, and managers who themselves were later cut after declaring AI would transform development.

The core argument is that frontier models are not yet replacements for seasoned engineers because they lack taste and long-term foresight required for sound product decisions, accessibility, and operational trade-offs. Yet these models excel at repeatable, measurable tasks: copying across domains, porting code, and reproducing bugs, which lowers hiring costs and enables offshoring or wholesale technical changes - Shopify’s push from React Native to machine-driven native apps is cited. Specific examples include a Typescript-to-Rust port that replicated bugs with Opus 5.5 and a rival model, Astra, failing a familiar task. The result: companies spend differently on human labor, the joy of programming erodes, but product delivery persists. Contact details and feeds are offered for follow-up.