Schirp's dartboard mental model: every contribution is a dart whose landing position is classified by three vertical thresholds. Darts below the ecosystem threshold are auto-rejected by the language itself (syntax errors, boot failures). Darts between the ecosystem and automation thresholds fail CI-style quality gates — easy, cheap to throw again. Darts between the automation and contribution thresholds are the red zone: they pass all automated checks but still damage the company on merge, and only discipline stands between them and production. Darts above the contribution threshold are the green outcomes the organization wants. Overconfident developers assume their personal distribution skips the red zone; integrated over time (bad nights, fights, low caffeine) everyone regresses to the same Gaussian-ish curve. Minimize time spent with red; maximize the cycle of ecosystem/automation rejection; make the distribution irrelevant by narrowing the automation–contribution gap. LLMs multiply the dart count, not the distribution — they don't know where their darts land; statistics does the rest.