The Symbiotic Mind

AI × HI — Artificial Intelligence × Human Intelligence is not a tool usage question. It is a relationship design problem.

The unit of value is the combined AI × HI system — and it belongs to neither alone.

AI × HI is not about using AI better. It is about designing how we think and act with it.

Published at symbiotic-mind.com.

---

Stack

Static site built with Eleventy 3.

• Markdown posts in src/posts/ with frontmatter (title, date, author slug, description, tags, optional cover and video retell).
• Layouts and partials in src/_includes/.
• Global config in src/_data/site.js; author profiles in src/_data/authors.js.
• CSS in src/css/ (the original style.css, unchanged).
• Images in src/images/.
• Build outputs to _site/.
• Pagefind indexes the built site for client-side, fully static search at /search/.

Local dev

nvm use            # Node 20
npm install
npm run serve      # http://localhost:8080

Build

npm run build
node scripts/validate-build.mjs   # also runs in CI

npm run build is intentionally side-effect-free. Use npm run publish:indexnow only after a production build that should notify IndexNow.

What this build produces

Output	Purpose
_site/index.html, /posts/, /about/	Pages, same look as before
/posts/<slug>/	Each post, with OG + Twitter + JSON-LD Article schema
/authors/vlad/, /authors/milena/	Author profile pages
/tags/<tag>/, /tags/	Tag pages and index
/search/	Static, JS-free crawl, Pagefind-powered search UI
/feed.xml	RSS 2.0 with full <content:encoded> for every item
/feed.json	JSON Feed 1.1 with full content_html
/feed/vlad.xml, /feed/milena.xml	Per-author feeds
/sitemap.xml	Real XML sitemap
/robots.txt	Plain text with Sitemap: directive
/llms.txt	llms.txt for AI crawlers
/_headers, /_redirects	Cloudflare Pages config

Writing a new post

Create src/posts/NNN-slug.md (NNN is the next sequence number):

---
title: "Post title"
date: 2026-06-01T07:30:00+02:00
author: vlad   # or milena
description: "One-sentence summary used in OG, RSS, and post lists."
tags:
  - relationship-design
cover:                     # optional
  src: "/images/NNN-cover.png"
  alt: "Alt text."
ogImage: "/images/NNN-cover.png"  # optional, defaults to /images/og-default.png
videoRetell:                # optional
  url: "https://www.canva.com/design/.../view?embed"
---

Body in markdown.

Open a PR → CI builds, validates, deploys on merge.

Newsletter

Two providers are wired in. Edit src/_data/site.js:

newsletter: {
  provider: "both", // "buttondown" | "listmonk" | "both"
  buttondown: { username: "symbioticmind", embedUrl: "..." },
  listmonk: { formActionUrl: "https://lists.example.com/subscription/form", listUuid: "..." }
}

• Buttondown (SaaS, free under 100 subs, RSS-to-email built in).
• listmonk (self-hosted, free, no ceiling). When you stand it up, fill in the formActionUrl and listUuid — both forms will then render on the home page and at the end of each post.

Comments

Existing Giscus config in src/_data/site.js. No change needed — the partial renders on every post.

Deploy

CI/CD: .github/workflows/deploy.yml builds and validates the site as a GitHub check. Cloudflare Pages owns production and preview deploys through its Git integration. Use npm run build:publish && node scripts/validate-build.mjs as the production Cloudflare build command when IndexNow pings should run.

Project name on Cloudflare: symbiotic-mind.