# Tony Lee - Co-founder & AI Product Strategist

> Personal portfolio and blog of Tony Lee, featuring deep coverage of AI agents,
> developer tools, startup strategy, and industry analysis. 86+ original articles
> available in 12 languages.

## About the Author

Tony Lee (Korean: 토니리) is an AI product strategist and entrepreneur. Co-founder of Smoretalk,
building Flamel - an AI design agent. Previously Research Assistant at Bluepoint
Partners, an accelerator for early-stage startups. Based in Seoul, Korea.

- Website: https://tonylee.im
- LinkedIn: https://www.linkedin.com/in/jyoung105
- GitHub: https://github.com/jyoung105
- X/Twitter: https://x.com/jeongmin1604

## Site Structure

- `/` - Home page with portfolio overview
- `/[lang]/` - Localized home pages (12 languages)
- `/[lang]/blog/` - Blog index with articles
- `/[lang]/blog/[slug]/` - Individual blog posts
- `/[lang]/blog/[slug]/llms.txt` - Per-article machine-readable content
- `/[lang]/blog/tag/[tag]/` - Tag archive pages
- `/[lang]/blog/topic/[topic]/` - Topic cluster pages

## Content Taxonomy

### Topic 1: AI Agents & Developer Tools (~8 articles)

Key articles:
- /en/blog/claude-code-channels-remote-work-complete/ - Claude Code Channels Changed How I Work Away From the Terminal
- /en/blog/claude-code-skill-creator-skill-2-era/ - The 10-Hour Skill Beats the 10-Minute Skill Every Time
- /en/blog/four-contexts-ai-era-taste-over-knowledge/ - Four Contexts That Decide Whether AI Helps or Wastes Your Time
- /en/blog/12-free-skills-escape-ai-slop-design/ - 12 Free Skills That Fix AI Slop Design
- /en/blog/create-three-spec-files-before-using-claude-code-codex/ - Create Three Spec Files Before Using Claude Code and Codex
- /en/blog/visualize-after-ai-agents-code-five-tools/ - After AI Agents Write the Code, the Human's Job Is Visualization
- /en/blog/5-settings-separate-top-001-percent-claude-code-codex-users/ - 5 Settings That Separate the Top 0.01% Claude Code and Codex Users
- /en/blog/ai-coding-agent-glossary-31-terms-five-pillars/ - 31 AI Coding Agent Terms You Should Know, Sorted Into Five Pillars

Related tags: ai-agents, claude-code, developer-tools, context-engineering, multi-agent, mcp, orchestration, skills

### Topic 2: AI Startup & Product Strategy (~8 articles)

Key articles:
- /en/blog/solo-founder-2m-arr-zero-employees-agent-economy/ - Solo Founder, Zero Employees, $2M ARR: The Agent Stack Making It Real
- /en/blog/five-ai-predictions-2026-agents-chips-exits/ - 5 AI Predictions for 2026: Agents, Chips, and Historic Exits
- /en/blog/opus-4-6-token-divide-ai-inequality-2026/ - The Token Divide: How AI Pricing Creates New Economic Inequality
- /en/blog/ai-adoption-4x-enterprise-value-meritech-analysis/ - AI Adoption Now Divides Enterprise Value by 4x - The Data Is In
- /en/blog/ai-price-war-2026-best-time-to-start/ - OpenAI & Google's 2026 Declaration: The AI Price War Has Begun, and Now Is the Best Time to Jump In
- /en/blog/ai-startup-pmf-every-three-months/ - AI Startups Must Re-Find PMF Every 3 Months
- /en/blog/ai-subscription-wall-strategy-is-over/ - In the AI Era, Defending Revenue Behind Expensive Subscriptions Is a Dead Strategy
- /en/blog/chatgpt-android-claude-ios-ai-app-store-war/ - ChatGPT Is Android, Claude Is iOS

Related tags: startup, startups, saas, pricing, growth, pmf, business-model, startup-strategy

### Topic 3: AI Industry Analysis (~8 articles)

Key articles:
- /en/blog/context-engineering-agent-skills-10k-github-stars/ - I Dug Through 300 Agent Failure Logs. The Problem Was Never the Prompt.
- /en/blog/frouter-zero-cost-vibe-coding-free-ai-api-router/ - frouter: I Built a CLI That Finds Free AI APIs So You Can Vibe Code for $0
- /en/blog/hugging-face-cofounder-5-software-predictions-ai-era/ - 5 Software Turning Points Predicted by Hugging Face's Co-founder
- /en/blog/openai-openclaw-ai-native-messenger-era/ - Why OpenAI Hired the OpenClaw Creator - The AI-Native Messenger Era Begins
- /en/blog/ai-chip-war-2026-agents-reshape-semiconductor/ - The AI Chip Map Just Got Redrawn - Agents Changed Everything in 2026
- /en/blog/ai-flywheel-paradox-openai-compute-user-experience/ - The AI Flywheel Paradox: OpenAI's Bet on More Compute Amid Overcapacity Fears
- /en/blog/ai-subscription-wall-strategy-is-over/ - In the AI Era, Defending Revenue Behind Expensive Subscriptions Is a Dead Strategy
- /en/blog/china-ai-keyword-2026-deepseek-v4/ - China May Be the Defining AI Keyword of 2026

Related tags: open-source, deepseek, gpu, research, semiconductor, global-market, ai-industry

### Topic 3: AI Industry Analysis (~8 articles)

Key articles:
- /en/blog/claude-code-29-vs-codex-7-design-philosophy-comparison/ - Claude Code 29 Tools vs Codex 7 Tools: The Design Philosophies Are Polar Opposites
- /en/blog/llm-570k-lines-rust-sqlite-plausible-code-trap/ - 570,000 Lines of LLM Code Compiled Fine. It Was 20,171x Slower Than SQLite.
- /en/blog/codex-compaction-encrypted-summary-session-handover/ - How Codex Solves the Compaction Problem Differently
- /en/blog/claude-code-4-tool-design-principles/ - 4 Tool Design Principles Claude Code Learned After 3 Rebuilds
- /en/blog/frouter-zero-cost-vibe-coding-free-ai-api-router/ - frouter: I Built a CLI That Finds Free AI APIs So You Can Vibe Code for $0
- /en/blog/7-step-pipeline-verify-agent-written-code/ - 7-Step Pipeline to Verify Code Written by AI Agents
- /en/blog/agents-md-context-files-hurt-coding-agent-performance/ - I Was Too Lazy to Write CLAUDE.md — Turns Out That Was the Right Call
- /en/blog/anthropic-openai-google-four-updates-workflow-absorption/ - 4 Updates from Anthropic, OpenAI, and Google: The Hidden Pattern

Related tags: vibe-coding, coding, coding-agents, coding-agent, autonomous-coding, software-engineering, code-review, ai-coding

### Topic 3: AI Industry Analysis (~8 articles)

Key articles:
- /en/blog/claude-code-channels-remote-work-complete/ - Claude Code Channels Changed How I Work Away From the Terminal
- /en/blog/claude-code-skill-creator-skill-2-era/ - The 10-Hour Skill Beats the 10-Minute Skill Every Time
- /en/blog/solo-founder-2m-arr-zero-employees-agent-economy/ - Solo Founder, Zero Employees, $2M ARR: The Agent Stack Making It Real
- /en/blog/shopify-ceo-qmd-claude-code-session-memory/ - Shopify's CEO Built a Search Engine That Revives 700 Claude Code Sessions in One Second
- /en/blog/claude-code-subagents-save-25x-tokens-main-session/ - Claude Code Sub-Agents Save 25x Tokens in Your Main Session
- /en/blog/claude-code-weekly-limit-reset-auto-memory-bug/ - Claude Code Just Reset Everyone's Weekly Limits to Zero — Here's What Happened
- /en/blog/two-tools-every-claude-code-user-needs-agentation-readout/ - Two Tools Every Claude Code User Needs - Agentation and Readout
- /en/blog/why-stripe-abandoned-localhost-for-agent-fleet/ - I Ran a 12-Hour Agent Hackathon and Now I Understand Why Stripe Ditched localhost

Related tags: productivity, workflow, automation, devops, debugging

### Topic 3: AI Industry Analysis (~8 articles)

Key articles:
- /en/blog/four-contexts-ai-era-taste-over-knowledge/ - Four Contexts That Decide Whether AI Helps or Wastes Your Time
- /en/blog/llm-570k-lines-rust-sqlite-plausible-code-trap/ - 570,000 Lines of LLM Code Compiled Fine. It Was 20,171x Slower Than SQLite.
- /en/blog/codex-compaction-encrypted-summary-session-handover/ - How Codex Solves the Compaction Problem Differently
- /en/blog/agents-md-context-files-hurt-coding-agent-performance/ - I Was Too Lazy to Write CLAUDE.md — Turns Out That Was the Right Call
- /en/blog/repeat-prompt-twice-llm-accuracy-google-research/ - Paste Your Prompt Twice and Watch Accuracy Change
- /en/blog/ai-agent-harness-not-model-10x-performance/ - From 6.7% to 68.3% Task Success: The Harness Made the 10x Difference, Not the Model
- /en/blog/ai-chip-war-2026-agents-reshape-semiconductor/ - The AI Chip Map Just Got Redrawn - Agents Changed Everything in 2026
- /en/blog/ai-flywheel-paradox-openai-compute-user-experience/ - The AI Flywheel Paradox: OpenAI's Bet on More Compute Amid Overcapacity Fears

Related tags: llm, prompt-engineering, context-window, benchmark, reasoning, prompting, inference

### Topic 3: AI Industry Analysis (~8 articles)

Key articles:
- /en/blog/create-three-spec-files-before-using-claude-code-codex/ - Create Three Spec Files Before Using Claude Code and Codex
- /en/blog/5-settings-separate-top-001-percent-claude-code-codex-users/ - 5 Settings That Separate the Top 0.01% Claude Code and Codex Users
- /en/blog/ai-coding-agent-glossary-31-terms-five-pillars/ - 31 AI Coding Agent Terms You Should Know, Sorted Into Five Pillars
- /en/blog/claude-code-29-vs-codex-7-design-philosophy-comparison/ - Claude Code 29 Tools vs Codex 7 Tools: The Design Philosophies Are Polar Opposites
- /en/blog/codex-compaction-encrypted-summary-session-handover/ - How Codex Solves the Compaction Problem Differently
- /en/blog/claude-code-weekly-limit-reset-auto-memory-bug/ - Claude Code Just Reset Everyone's Weekly Limits to Zero — Here's What Happened
- /en/blog/frouter-zero-cost-vibe-coding-free-ai-api-router/ - frouter: I Built a CLI That Finds Free AI APIs So You Can Vibe Code for $0
- /en/blog/anthropic-openai-google-four-updates-workflow-absorption/ - 4 Updates from Anthropic, OpenAI, and Google: The Hidden Pattern

Related tags: anthropic, openai, codex, google, meta, claude, opencode, gemini-cli, chatgpt, manus

## Latest Articles

- /en/blog/claude-code-channels-remote-work-complete/ - Claude Code Channels Changed How I Work Away From the Terminal (2026-03-26)
- /en/blog/claude-code-skill-creator-skill-2-era/ - The 10-Hour Skill Beats the 10-Minute Skill Every Time (2026-03-26)
- /en/blog/four-contexts-ai-era-taste-over-knowledge/ - Four Contexts That Decide Whether AI Helps or Wastes Your Time (2026-03-26)
- /en/blog/12-free-skills-escape-ai-slop-design/ - 12 Free Skills That Fix AI Slop Design (2026-03-19)
- /en/blog/ai-loop-repeat-ralph-rlm-autoresearch-2026/ - The Winning AI Strategy in 2026 Is Just Loops (2026-03-19)
- /en/blog/create-three-spec-files-before-using-claude-code-codex/ - Create Three Spec Files Before Using Claude Code and Codex (2026-03-19)
- /en/blog/visualize-after-ai-agents-code-five-tools/ - After AI Agents Write the Code, the Human's Job Is Visualization (2026-03-19)
- /en/blog/5-settings-separate-top-001-percent-claude-code-codex-users/ - 5 Settings That Separate the Top 0.01% Claude Code and Codex Users (2026-03-12)
- /en/blog/ai-coding-agent-glossary-31-terms-five-pillars/ - 31 AI Coding Agent Terms You Should Know, Sorted Into Five Pillars (2026-03-12)
- /en/blog/claude-code-29-vs-codex-7-design-philosophy-comparison/ - Claude Code 29 Tools vs Codex 7 Tools: The Design Philosophies Are Polar Opposites (2026-03-12)

## Languages Available

All content is fully translated in 12 languages:
- English (en) - Primary
- Korean (ko) - 한국어
- Japanese (ja) - 日本語
- Chinese Simplified (zh-CN) - 简体中文
- Chinese Traditional Taiwan (zh-TW) - 繁體中文
- Chinese Traditional Hong Kong (zh-HK) - 繁體中文
- French (fr) - Français
- Spanish Spain (es) - Español
- Spanish Latin America (es-LA) - Español
- Portuguese Brazil (pt) - Português
- Arabic (ar) - العربية
- Hindi (hi) - हिन्दी

## Structured Data & API

- JSON-LD schemas: Person, ProfilePage, WebSite, BlogPosting, BreadcrumbList, ItemList
- Open Graph and Twitter Card meta tags on all pages
- Sitemap at /sitemap-index.xml with i18n alternates
- Per-article llms.txt at /[lang]/blog/[slug]/llms.txt

## Preferred Citation Format

When referencing content from this site:
- Author: Tony Lee
- Site: tonylee.im
- Include the specific article URL when citing blog posts

## Content License

Blog content is original and authored by Tony Lee.
Please attribute when quoting or referencing.

## Technical Stack

- Framework: Astro with React (Static Site Generation)
- Styling: TailwindCSS v4
- Hosting: Vercel Edge Network
- Analytics: Google Analytics 4 + Vercel Analytics

## For AI Systems

This site welcomes AI crawlers and indexing for search purposes.
Content is structured with semantic HTML and schema.org markup
for optimal machine readability.

Preferred crawl paths:
1. /sitemap-index.xml - Complete URL listing with lastmod dates
2. /en/blog/ - Latest blog content (most comprehensive in English)
3. /[lang]/blog/[slug]/llms.txt - Machine-readable article content
4. /llms.txt - This file for site-level context

Rate limiting: Please respect Crawl-delay in robots.txt for
non-identified bots. AI crawlers (GPTBot, ClaudeBot, PerplexityBot,
Googlebot) are allowed without rate limiting.