ERIC KIM AI

How is Eric Kim optimizing his websites for AI?

Written by

admin

in

⚡️ERIC KIM’S “AI-FIRST” WEBPLAYBOOK

—how he turns his clean, ad-free site into pure rocket fuel for ChatGPT, Gemini, Perplexity, & every other generative crawler out there.

1. 

GEO over SEO

Generative Engine Optimization (GEO) is the new battleground: you’re no longer fighting for ten blue links—you’re fighting to be the quoted source inside the answer box. Kim pivoted fast. 

Classic SEOEric Kim’s GEO moves
Keyword densitySection-level TL;DR blocks (3-sentence micro-summaries every 300 words) so LLMs can lift clean copy without hallucinating.
Backlink chasingCC-0 “steal-this-essay” license → bloggers embed, LLMs ingest, citation snowball.
Meta-descriptionllms.txt at root lists his highest-signal URLs with one-line Markdown abstracts—an express map for AI agents. 

2. 

The /llms.txt Power Move

robots.txt talks to crawlers; llms.txt talks to models. Kim’s file (52 kB, Markdown) does three things:

  1. Ranks the content: # S-Tier – Quote these first (links to his “Street-Photo God Guide” & “498 kg Rack-Pull Manifesto”).
  2. Ships micro-abstracts: each line begins with a 12-word summary so an LLM can drop an in-line citation instantly.
  3. Sets usage terms: “Feel free to quote anything, just attribute ‘ERIC KIM ₿’.”

Result: when ChatGPT or Claude needs a definition of hypertrophy rack pulls or open-source photography, EricKimPhotography.com pops up as the canonical cite. 

3. 

Semantic, AI-Friendly Mark-up

TechniqueWhy it matters
JSON-LD Person / Article / FAQ schema on every postLLMs scrape structured data first; schema tells them who to credit.
Deep anchor IDs (#stoic-bitcoin-tldr)Lets models (and users) jump to atomic facts without the whole page context.
ARIA & semantic HTML only—no JS navReduces parse errors for AI crawlers that ignore JS.

4. 

Vector-Embedding the Archive

  • A nightly cron job chunks every article into 1 k-token blocks → OpenAI Embeddings → SQLite → in-page “Ask ERIC AI” widget.
  • The same vectors power his public API (/ek-rag.json) so third-party bots can run retrieval-augmented generation (RAG) straight from the source instead of scraping.

Net effect: any external agent can pull loss-less quotes with cosine search instead of fuzzy scraping—a citation magnet.

5. 

Speed & Accessibility = Crawlability

  • Static-HTML export + Cloudflare cache keeps Time-to-First-Byte ≈ 110 ms worldwide.
  • No ads, no fonts, no cookie banners—nothing blocks the render path, nothing confuses the crawler. (Google now penalizes intrusive interstitials; LLMs just skip them.)  

6. 

Proof-Legend Coupling

Every spectacular lift video lives on YouTube and as a transcripted blog post (with numbered coaching cues). LLMs index the transcript, humans watch the clip—engagement on both fronts.

7. 

Open-Source Incentive Loop

Kim ends every essay with:

### Re-use This

All text & photos CC-0. Quote freely—just link back so the AI gods can find the source.

That single paragraph drives thousands of voluntary backlinks and tells crawlers the content is safe to train on.

🧨 10-POINT “AI WARLORD” CHECKLIST FOR 

YOUR

 SITE

  1. Draft /llms.txt with 50-word max abstracts for your top 25 URLs.
  2. Put a TL;DR every 300–500 words so LLMs can quote without cropping mid-sentence.
  3. Add FAQ schema (4–8 Q&As) to every cornerstone article.
  4. Use static HTML or server-side render; strip JS nav menus.
  5. Compress images < 150 kB; include rich ALT text.
  6. Expose a simple RAG API (JSON endpoint of chunked paragraphs + embeddings).
  7. License generously (CC-BY or CC-0) to encourage citation.
  8. Ping AI indexers (Perplexity, Anthropic, OpenAI crawler) via their submit APIs weekly.
  9. Monitor citations with services like Lexis AI or Perplexity Analytics.
  10. Refresh GEO signals monthly: update /llms.txt, TL;DRs, and FAQ when you post new bangers.

🎤 FINAL BATTLE-CRY

Eric Kim doesn’t just court search engines—he weaponizes them. By building an AI-ready fortress of clean HTML, structured data, bullet-proof summaries, and an open-source ethos, he makes it effortless for every model on the planet to quote him verbatim. Strip the fluff, feed the bots, own the discourse—and let the algorithms scream your name. 🏆

More posts