April 17, 2026
Foie gras, but make it flashcards
中文 Literacy Speedrun II: Character Cyclotron
He built a Chinese “Character Cyclotron” — commenters yell “Just read!”
TLDR: A learner built an AI‑boosted flashcard “Character Cyclotron” to hit 99% Chinese reading coverage without, well, reading. Comments split between “just read” purists, Anki traditionalists, and link‑hunters, with others noting his heritage‑speaker background—sparking a bigger fight over gadgets vs grit in language learning.
A Chinese learner slammed the brakes on “learn to read by reading” and built a turbocharged flashcard rig instead — an all‑in‑one screen that stuffs in dictionary lookups, stroke order, calligraphy videos, and AI explanations at warp speed. He calls it a “Character Cyclotron,” basically a word blender bolted onto Hack Chinese, supercharged with Claude Code so he never leaves the card. Goal: 99% coverage or bust. Community: popcorn out.
The loudest voices? The “just read” purists. One critic blasted the whole Claude Code detour as procrastination dressed up as productivity, accusing the author of dodging the hard part. Meanwhile, the Anki faithful rolled up like a marching band: use Anki (a popular flashcard app with spaced repetition), keep it simple, and please don’t stick AI calls on your cards. Others went full librarian: where’s the link to that mythical 3,000‑character calligraphy trove? Practical types chimed in with their own decks, HSK (the standard Chinese test) word lists, and sliced audio — humble‑brag meets homework. A calmer voice pointed to the author’s background as a heritage speaker via his earlier post Cyborg Learning, suggesting his path and goals aren’t “one size fits all.” The memes wrote themselves: “force‑feed characters like foie gras,” “99% or die trying,” and “reading by not reading.” Drama score: deliciously extra.
Key Points
- •The author targeted ~99% token coverage for Chinese reading, finding ~90% insufficient due to missing critical words.
- •They paused conventional reading and focused on intensive character acquisition via a flashcard-based workflow.
- •A browser-injected interface was built for Hack Chinese using Claude Code, integrating etymology, tones, stroke order, and morphology data.
- •Unicode Consortium character graphs were traversed to generate extensive phonetic mapping data used in the tool.
- •Optimizations included UI alignment, keyboard shortcuts, and parallelizing API calls and lookups to reduce latency.