PatronView owner Nick Gray says his 1.5 million-page website spent a year fighting scrapers, according to a Techmeme item summarizing Gray’s PatronView post. The reported numbers are stark: Gray says bot page loads outnumbered human page loads by 214 to 1. He also reported 35,000 Claude crawls per referred user, while an Amazon bot referred no users, according to the Techmeme summary. Those figures should be treated as Gray’s site-level account, not as an independently verified benchmark for the broader web. The cluster contains only one item, and it does not include outside measurement, responses from the named companies, or enough detail to compare crawler behavior across sites. Still, the account captures a concrete operator problem: large websites can face heavy automated access from AI-related crawlers while receiving limited or no measurable referral traffic in return. For publishers, directory sites, archives, and user-generated databases, the operational question is not only whether AI crawlers visit, but whether their visits impose load without sending users back. The details available here do not establish what countermeasures PatronView used, how much infrastructure cost the traffic created, or whether any bot activity complied with the site’s rules. They do, however, put numbers around the imbalance one site owner says he observed over a full year. Who benefits: Companies building crawlers benefit when large public websites remain broadly accessible. Site owners benefit from better measurement of crawler load, referral value, and bot behavior. Who's exposed: Large content-heavy websites are exposed if automated traffic increases infrastructure burden without clear user acquisition. The named crawler operators are exposed to scrutiny over how much value they return to the sites they access.