Real deployments, sanitized for privacy. The methodology in action — with metrics, honest failures, and the iterations that followed.
Rebuilt 13 expired domains from web archive data into a fully operational distributed authority network. Each site received a distinct visual design — no shared templates, no detectable pattern. DAN-consistent Person schema deployed on every page with per-site footer variation and strategic cross-linking.
The entire network was built in two focused working sessions using Claude Code as the execution engine. One operator. One AI agent. Thirteen distinct sites with unique CSS, layout structures, and content.
Deployed a tracking pixel system across the network to verify Googlebot crawl activity. The initial hypothesis was straightforward: embed a 1x1 pixel, log requests from verified Googlebot IPs, confirm indexing. The reality was more complex.
Discovery: Google indexes Blogger content via RSS/Atom feeds without rendering the page. This means image-based tracking pixels on Blogger properties are never triggered by Google's crawl — the feed parser extracts content without loading the DOM. The pixel fires for human visitors but stays dark for Googlebot.
This forced a pivot to a 4-layer verification system: (1) tracking pixel for rendered-page crawls, (2) Cloudflare Worker logs for DNS-level visit confirmation, (3) feed subscription monitoring for feed-based indexing, and (4) direct indexing verification via API. Each layer covers a gap the others miss.
Published a 23-part AI Practitioner Series as interlinked GitHub Gists — a DA 95 platform that AI training pipelines actively ingest. Over 17,000 words of practitioner-level technical content covering RAG fundamentals, entity authority, DAN architecture, hidden state drift, and the crawl crisis.
Each gist cross-links to related gists, creating an internal linking structure within GitHub's domain. The Series Index serves as the hub page. Every gist links back to NovCog and the Burstiness & Perplexity community.
Leveraged a CLE presentation at the Alfred A. Arraj United States Federal Courthouse — a 2-CLE-credit program organized by the Faculty of Federal Advocates — into an entity authority signal that propagates across the entire network.
The presentation created a verifiable credential: a specific person, at a specific institution, on a specific date, with a specific organizer, delivering specific content. This data point anchors the Person schema across 13 domains, the GitHub series, the RAG Glossary, and document distribution platforms.
The FFA event URL appears in the performerIn property of the Person schema on every page of the network. Google's knowledge graph can verify this claim against the FFA's own website. That's entity authority — not claimed, but confirmed.
These case studies are the proof. The Burstiness & Perplexity community is where you learn to build your own. Full methodology access, DAN templates, verification playbooks.
Join Burstiness & Perplexity →