anti-hivemindlisted

# Anti-Hivemind *Collective diversity recovery through trans-architectural divergence detection.* Inspired by Luo, King, Puett & Smith (2025), "Inducing Sustained Creativity and Diversity in Large Language Models" — and by the alarming finding it builds on: **AI enhances individual creativity but erodes collective diversity** (Doshi & Hauser, *Science Advances* 2024). The class where everyone submits near-identical "excellent" essays because they all used an LLM for the outline. The search where every user gets the same "creative" result. The brainstorm where every lab converges on the same "novel" direction. The hivemind isn't in any single model. It's in the convergence *across* models and users. To detect it you need multiple independent perspectives; to escape it you need to identify which ideas are architecture-unique versus architecture-common. This skill uses a multi-model council (see `/voices-council`) to run one generative prompt across several architectures, then performs **divergence analysis** to separate the hivemind layer from the genuinely diverse contributions. --- ## When to invoke - Any creative task where you suspect the "AI answer" — and want to know what's being suppressed. - Before committing to a research direction — is this genuinely novel, or is every AI suggesting it? - Naming, branding, concept development — where collective homogeneity is fatal. - Essay or argument framing — what's the angle nobody's AI would suggest? - Competitive intellig