Shopify reports that AI-driven traffic and orders to its stores tripled year over year, directly contradicting the narrative that AI search cannibalizes referral traffic. The same week, the Bushwick Daily shuttered, a small but symptomatic casualty in a year of editorial collapse. The divergence is not a paradox. It is a sorting mechanism, and it is sorting by intent.
Why Commerce Gets the Clicks and Editorial Gets the Void
AI search engines are extraordinarily good at transactional queries: find me this product, compare these options, where do I buy this. They are structurally disincentivized to send users to editorial sources because doing so creates a secondary loop the AI cannot control. Shopify merchants benefit because a user saying 'best running shoes under $100' ends at a product page, not a review. A user saying 'what's happening in Gaza' ends at a synthesized answer that cites no one and links nowhere. The commercial web and the informational web are experiencing opposite fates from the same technological shift.
The Misinformation Gap Widens for Small Businesses Too
A 2026 arXiv paper by Islam et al., Enhancing LLMs with Context-Specific Knowledge for Mitigating Misinformation in SMEs, finds that small and medium enterprises are especially vulnerable to LLM-generated misinformation in B2B contexts, where AI confidently surfaces outdated or fabricated vendor data. Shopify's aggregate traffic numbers look healthy, but the quality of AI-driven commerce queries is uneven. The stores benefiting most are those with strong product data and review density, i.e., already-established merchants, not the long tail. Kyle Raymond Fitzpatrick's framing of enshittification applies here with unusual precision: AI search is accelerating the Matthew effect, more traffic to those who already have it, less to those trying to build it.