About

About FLOW SEEKER LAB

Frameworks, not forecasts. AI, Crypto, and the long game.

I’m seonjae — a Korean office worker without a computer science degree, documenting my transition into AI systems, agents, and vibe coding in public.

This site is the notebook. Every post here is something I’ve actually built, refused to build, measured, or broken. The parts that did not work stay on the page.

Why this site exists

Most writing about AI right now is either breathless hype or vendor copy with a productivity costume on. Neither one helps a working professional who pays for ChatGPT Plus, opens Notion every morning, and quietly wonders where to start with automation that lasts longer than a weekend.

FLOW SEEKER LAB is built for that reader.

The promise is narrow on purpose:

  • No predictions. No “where AI will be in three years.” Frameworks instead — the kind that survive whichever tool is hot this month.
  • No income claims. I do not quote dollar figures for builds I have not actually run. If a number sounds suspiciously precise, it is.
  • First-person only. Every framework is something I shipped against, or deliberately did not ship and wrote down why.
  • The long game. Posts here aim to be useful six months from now, not six hours.

What you’ll find here

The site is organized into five clusters. The AI Agent Lab is the spine; the others are adjacent investigations.

  • AI Agent Lab — no-code agents, vibe-coded scripts, the ten-step framework I use before opening any builder. Build logs include the URLs that broke, the timezones I got wrong, and the default permission scopes I forgot to tighten.
  • On-chain Digest — calm literacy for non-traders. How to read on-chain data, how to spot a rug pull, the five-layer mental model. No price predictions.
  • Protocol Deep Dives — slower technical walk-throughs of specific systems.
  • Asia Pulse — the Korean and broader East Asian view of AI adoption, individual workers versus institutions, ground-level observation.
  • The Compounding — meta-frameworks. Why depth beats breadth. Why constraints enable focus. The long-game thinking underneath the build logs.

Who I am, honestly

  • I am not a developer. I am learning to ship working software using AI tools and a stubborn habit of writing things down.
  • I work a full-time office job in Korea. The publication cadence here matches a working professional’s life — roughly one or two posts a month, not one a day.
  • I pay for the AI tools I write about with my own money. I do not get free seats from vendors.
  • The Korean office worker angle is real, not a marketing position. I write from inside the gap between what individuals can do with AI today and what companies still take quarters to approve.

What this site is not

  • Not investment advice. Nothing on the On-chain Digest or anywhere else is a recommendation to buy or sell anything.
  • Not a signals service. Not a course funnel.
  • Not generated and posted without editing. Every post is shipped by hand and credited to me. AI is used as a drafting and research assistant the same way a writer in 2010 used a search engine — never as a substitute for judgment.
  • Not a place for hot takes. New tools take weeks to be worth a workflow here.

How to follow along

The most reliable way to keep up is to bookmark the AI Agent Lab and check in once a month. New build logs land there with the same series numbering.

If something here saves you time or changes how you scope your next agent, that is the point. If something here is wrong, please tell me — corrections are how the framework gets sharper.

The long game beats the short sprint. Thanks for reading.

Last updated: 2026-05-28. Site author: seonjae. Site identity: Korean office worker documenting his transition into AI systems, agents, and vibe coding — without a CS background. Shipping in public.