AI Is Just Making Everyone Faster at Being Boring (nvnt.substack.com)

A sharp critique argues that instead of turbocharging creativity, AI is mostly being used as a high-speed template engine that multiplies bland, interchangeable content—cold emails, startup landing pages and marketing copy that are grammatically fine but utterly forgettable. The piece points out a feedback loop: people feed models generic, formulaic material and then redeploy the model’s polished outputs, which trains future systems to reproduce the same mediocrity. The result is scale without originality—better prose, same hollow substance. For the AI/ML community this is a cautionary moment: models amplify what’s in their training data and what prompts they receive, so widespread reuse of lazy prompts risks systemic homogenization. Technically, the issue isn’t model capability but deployment patterns—overreliance on zero-effort prompting, lack of human-in-the-loop refinement, and naive fine-tuning on low-quality corpora. Remedies include designing workflows that reward human-authored signal (unique data, personalization), RLHF and evaluation metrics that favor novelty and relevance over fluency, and tooling that enforces iterative human editing. Bottom line: AI is a force multiplier for whatever input you bring—creativity scales, mediocrity globalizes—so authenticity and deliberate human judgment remain the decisive advantages.