Slop*Detector
Verdict for anthropic.com
INCONCLUSIVE — slop score 26/100
The evidence here is thin, and that's the point: this reads like a site built by editors and designers, not funneled through a single AI prompt. The strongest tell is a short run of clipped questions in the navigation ('How does AI work? How should I use AI? Who should govern AI?'), which brushes up against the stacked-fragment rhythm the detector flags, but it's a nav menu, not prose, so the weight is limited. The three-card release grids and the link-list block under 'At Anthropic, we build AI' both hit familiar templated layouts, yet they're carrying real editorial content (product announcements, policy docs) rather than generic marketing filler. A repeated 'hard questions' phrase and one three-noun triad round out the findings, but neither is frequent enough to build a pattern. At 26 out of 100, the file stays inconclusive: there's a handful of structural tells common to any modern corporate site, but nothing here points specifically to end-to-end AI generation.
Findings
- Feature-tile grid — Equal-width card grids of 3–4 identical tiles (icon, subheader, blurb) are the signature layout block of generated landing pages.
- Uniform 3-card grid for latest releases — The 'Latest releases' section shows three identical cards (Redeploying Fable 5, Introducing Sonnet 5, Announcing Claude Science) each with the same structure: headline, blurb, date, category, and a black CTA button. This is close to the templated feature-tile pattern, though it's editorial content rather than marketing features, so it's a weaker instance of the tell.
- List section resembling stock template row — The 'At Anthropic, we build AI...' block pairs a short mission statement with a plain list of links with category tags on the right, a common generated-content pattern for filling space with rows of links.
- Stacked short-sentence drumbeat — Three or more clipped sentences in a row ("Important emails buried. Drafts that sound wrong. Labels everywhere.") is the manufactured-rhythm tell.
- Mild rule-of-three — The phrase groups three abstract nouns in the classic triad pattern, though this is a single, low-frequency instance rather than a repeated tic.
- Repeated phrase echo — The words 'hard questions' / 'hardest questions' are reused back-to-back in adjacent sentences, a minor redundancy but not a strong AI signature on its own.