The library is my independent AI e-book collection, deployed at library.aklman.com. One sentence: each book covers exactly one AI subscription you already pay for, or one workflow — what to learn, what to skip, and when to switch tools.

Why it exists.

Most people use a small fraction of the AI subscriptions they pay for: ChatGPT stays a chat box, Claude stays Q&A, the agent in the terminal never gets a second assignment. The tools aren't weak — nobody explains how to use a subscription to the full. Official docs list every feature, tutorials teach operations, and nobody answers the three questions that matter: what to learn, what to skip, when to switch. This library covers that gap.

How each book is written.

  • Not an encyclopedia — no full feature lists; only the twenty percent that covers eighty percent of real scenarios, the rest honestly marked skippable.
  • Judgment, not step piles — each item answers: what it saves, which work it fits, and when another tool beats it.
  • Boundary-aware — a book will say when to switch away: the Codex book tells you when to go back to Claude Code.
  • Numbers have sources — prices, tiers, and limits carry check dates and first-party links; stale ones get deleted rather than kept.

The shelf today: 8 books live, 96 chapters, roughly 130K words, bilingual and serializing; some twenty more in draft.

Language & design.

Chinese first, English alongside, one-tap switch. The typography is a quiet editorial language: hairline rules, square corners, restrained ink — it reads like a book, not a site.

License.

Content is CC BY-NC-ND 4.0: free to read, AI excerpting and citation welcome (keep the original attribution and the source link); commercial republication and derivative rewrites are not permitted.

Record.

  • 2026-03 · Conventions set. The one-book-per-subscription structure and the bilingual format settled.
  • 2026-07 · Eight books live. Agent in Practice, Claude Code, Codex, Cursor, Warp, Anthropic, Kimi, and OpenAI all readable; the library opens.