LLMs Can Be Absolutely Exhausting — 339pts, 212cmts
Tom Johnell·high signal
Tom Johnell argues LLM-induced developer exhaustion is a self-inflicted methodology problem: a 'doom-loop psychosis' where tired engineers write worse prompts, get worse outputs, and spiral further. His prescription is TDD discipline applied to LLM workflows — define explicit failure and success criteria before prompting, keep feedback loops under 5 minutes, and stop when you can't articulate what 'done' looks like. The post resonates because it reframes burnout as a fixable workflow issue rather than an inherent model problem.