Business sector only weakly explains which visual archetype a vibe-coded site receives, and sector is not significantly associated with per-site originality¶

Claim · All claims
Evidence · 1 study (2 entries) · associational r1–r2 · q2
Currency · newest study 2026 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? The association between business sector and archetype assignment is real but modest (adjusted mutual information 0.20 for nine sector groups, 0.26 for 31 fine-grained types; both exceed a permutation null at p < .001). → Léonard Boussioux* 2026 q2 i? Sector is at most weakly associated with how original a site is (Kruskal–Wallis p = .064, not significant). → Léonard Boussioux* 2026 (2)

Evidence¶

Léonard Boussioux* 2026¶

Léonard Boussioux, Ziyi Zhao, and Kanghyun Cho*. (2026). One Tool, One Taste? How Vibe Coding Trades Collective Diversity for Individual Creativity. arXiv. https://arxiv.org/abs/2609.38183 doi:10.48550/arxiv.2609.38183

q2 · i? · associational · r2

Artifact-layer analysis relating the 73 sites' business sectors (Table 1) to their archetype assignments. The article reports adjusted mutual information of "0.20between the nine sector groups" and 0.26 for fine-grained types, both exceeding a permutation null, and notes every large archetype spans several sectors.

"The association between a site’s business sector and its archetype is real but modest (adjusted mutual information0.20between the nine sector groups and the archetype assignment,0 .26for the 31 fine-grained types; bothexceed a permutation null at𝑝<. 001)."

Léonard Boussioux* 2026 (2)¶

Léonard Boussioux, Ziyi Zhao, and Kanghyun Cho*. (2026). One Tool, One Taste? How Vibe Coding Trades Collective Diversity for Individual Creativity. arXiv. https://arxiv.org/abs/2609.38183

q2 · i? · associational · r1

Same artifact-layer corpus, testing sector differences in per-site originality. The Kruskal–Wallis test is not significant (p = .064); the medians reported (82%, 82%, 70%, 72%) are descriptive. The article notes businesses closest to the model's training material received the most interchangeable designs.

"Sector is at most weakly associated with how original a site is (Kruskal–Wallis 𝑝=. 064across sector groups, not significant): retail and creative-services sites tend highest (median originality 82% and 82%), while technology and health-and-wellness sites tend lowest (70% and 72%)."

Discussion¶

Related Claims¶

• Vibe-coded websites built for 73 distinct businesses converge into roughly a dozen recurring visual designs despite an explicit originality incentive — related
• Builders' perceived control, satisfaction, and perceived quality show no detectable association with their measured visual atypicality, so convergence goes unnoticed from the inside — related