The journal

Long-form English essays on Twin Scores, agentic publishing, and original 2026 research. Every piece has a named author. We do not publish thin listicles with invented user counts.

Essays

Original research first

• Human vs agent software surfaces in 2026: original Twin Score research — Edu Tor · 2026-08-10

  Original 2026 Twin Score research: 72 products scored twice, category means, and the design gap.

• How the Twin Score rubric actually works — Edu Tor · 2026-08-12

  Every human and agent dimension explained with catalog examples, plus category means you can recompute.

• Fake UGC is an SEO tactic that this directory refuses — Maya Chen · 2026-08-14

  Why we ban fake reviews, how schema stays honest, and the friction signup that keeps UGC pending.

• The capability matrix agents actually need — Maya Chen · 2026-08-16

  API, MCP, CLI, docs, and auth for the catalog — with a refusal to invent MCP servers.

• What is-agentic publishing means for a product directory — Jonas Okonkwo · 2026-08-18

  SSR, markdown negotiation, llms.txt, robots, JSON-LD, and the refusals we copied on purpose.

• ChatGPT vs Claude vs Gemini vs Perplexity vs Grok for real work in 2026 — Jonas Okonkwo · 2026-08-20

  Job-based comparison of assistants versus platform APIs, with Twin Scores and no universal winner.

• Human advantage versus agent advantage: a gap table from 72 products — Edu Tor · 2026-08-22

  Positive gaps are canvases; negative gaps are APIs. Original table from the committed catalog.