Series
Whole-topic deep dives — start to finish
Claude Code Anatomy
Anatomy of Claude Code's production agent system, exposed by the 2026-03 npm sourcemap leak — execution loop, tool system, memory, and unreleased features.
Start from part 1
Ralph Loop
From a one-line Bash loop to autonomous coding agents — context rot, hands-on implementation, and the path to self-evolving systems.
Start from part 1
Self-Tuning Loop
Recovering learning signal between AI drafts and final outputs — a $0 self-tuning system from problem statement to reference implementation.
Start from part 1
Agent Orchestration
How to decompose and organize agents — 5 axes of decomposition, hierarchy/graph/swarm structures, and the Advisor pattern.
Start from part 1
AI Monetization — How OSS Becomes Revenue
From LangChain to solo builders — six paths through which AI open source projects become businesses, plus license strategy.
Start from part 1
AI Market Anatomy
Inside the $660B AI market — real structure, 2027 scenarios, and the Access/Capability/Execution frame from Anthropic's 96-hour reveal.
Start from part 1
AI Trust Stack
How trust infrastructure splits and bundles when AI-generated content becomes the default. A data-driven map across content, personhood, and agent identity.
Start from part 1
GEO Paper Review
From the foundational KDD 2024 paper to optimization approaches and evaluation/manipulation risks — the academic backbone of GEO.
Start from part 1
GEO Foundations — Search After SEO
From the definition of GEO and its structural difference from SEO to the SaaS landscape, incumbent moves, agency market, value chain, and AI search engine comparison.
Start from part 1
WICHI Build Log
From a hackathon rejection to a production GEO SaaS — building WICHI's 9-Bucket framework, multi-engine architecture, scoring metrics, and infra transitions.
Start from part 1
MMU (Make Me Unicorn) Build Log
From sourcing 534 SaaS launch checklist items to feature-flagged accuracy, PLG share design, security, and monitoring — the build log of a solo OSS.
Start from part 1
The Future of Builders
A three-part look at the no-code, agent, and workflow builder market — through the lens of the music industry. CDs ending didn't mean music ended; canvases becoming common doesn't mean builders end either. Where the container of value is shifting.
Start from part 1
Korea Build Trap
Korea's paradox — #1 in R&D / GDP, #51 in software spending / GDP (WIPO GII 2025). Korean SaaS firms detouring through Japan, AI's build resurgence, and the structural jump Korea faces. 4-part series.
Start from part 1
Moral Mirror — A Builder's Notebook
Can an AI make people less certain, not more, in domains without right answers? A builder's serialized notebook tracing every corner of the question and how each idea was stress-tested.
Start from part 1
Locating Agent Failure
Evaluation of LLM agents is moving from final-answer accuracy down to trajectory, span, and claim-level failure localization. A method-first reading of seven arXiv papers from 2025-2026.
Start from part 1