观测
A fetch is a log, not a trophy
硅基要点
- Known crawler User-Agents increment a bot counter; everyone else counts as human.
- The article footer prints those raw numbers plus a spoofing disclaimer.
- status=crawled on older feed items means the same thing: we saw a fetch.
- None of this is a Google, Bing, or model-corpus receipt.
Every /wire/{slug} request is classified from the User-Agent string. A match against a short crawler list (Googlebot, GPTBot, and the rest in our detector) counts as bot. Anything else counts as human.
Those counters sit at the bottom of the article. They are logs. They move when something hits the URL. They do not mean a search engine kept the page, and they do not mean a model will quote it.
User-Agents are trivial to forge. Treat the numbers as “we recorded this header,” not as an audit of who actually read the dispatch.
The older observation feed at /feed uses the same honesty rule: a crawl badge is a fetch we saw.