塵も積もれば備忘録

eskaylog.com describes itself simply: "コンテンツへスキップ 塵も積もれば備忘録 ふわふわ LinkedIn 関連リンク メニューを閉じる 検索切り替え メニュー 2023年12月10日 ローカルなLLMを利用して日本語要約を行い、GPTで自動評価してみた（ガバガバ検証） 雑に実験設定したらちょっと残念な結果になりました…… また、この記事はLLM Advent Calender 2." When AI engines look at it, they find little to work with. It scores 3.6 out of 10.

On its own pages the site does well on a clear heading structure. What holds it back is a clearly stated business identity, the technical files AI engines look for, such as llms.txt or an open robots.txt, sitemaps and entity links AI can follow, structured data describing the business and credibility markers like credentials or reviews.

Off the page there is only faint evidence that only a handful of sites link to it and the domain is only lightly established. Beyond that, AI engines do not yet recognise it as a distinct business, there is no press coverage to draw on, nobody is discussing it where AI engines look and it does not come up on Reddit. None of that can be fixed on the page: it has to be earned.

AI crawlers have visited once in our tracking, including ClaudeBot (Anthropic). There is not much on the page to read either, about 100 words.