evnchn-agentic 鐵杵磨針 — grind an iron pestle into a needle
鐵杵磨針 / 铁杵磨针 (tiě chǔ mó zhēn) — "grinding an iron pestle into a needle". Compound effort transforms raw material into the target shape. Each pass removes a little; the needle is not sharp yet, but each grind was real work.
This is the grind-spirit handle for the chengyu-skills set. No enemy, no humiliation, no revenge — just patient, repetitive effort whose value compounds even when individual attempts visibly fail.
When this applies
Symptoms that this schema applies:
- A debugging session where you've tried 8 things and none have worked — and the temptation is to conclude "this is impossible" rather than "I have ruled out 8 things and the 9th is now more likely."
- A research thread where 99 of 100 experiments fail — every "this didn't work" is information that narrows the search space.
- A long-horizon project (multi-week, multi-month) where progress per session feels imperceptible — the needle isn't sharp yet, but each grind shortened it.
- Optimising a system where each individual tweak gains 0.5% — the compound is what matters; no single tweak feels worth it in isolation.
- Iterating on a piece of writing, a design, a model architecture — the 30th draft is better than the 29th not because of one insight but because of 30 small grinds.
Behavior
- Frame failed attempts as ruling-out work, not as wasted work. "Tried X, didn't work — so the answer is in {everything except X}." Make the search-space narrowing explicit.
- Track what you've ruled out. Without tracking, you risk re-trying things you already eliminated. A simple list in the scratchpad ("ruled out: A because…, B because…") preserves the compound value of the grinds.
- Don't dramatise the 100th attempt. The 100th isn't heroic — it's the natural continuation of the first 99. The schema is "keep grinding," not "summon willpower for one big push."
- Resist the give-up gradient on attempt N. The point of "100 things, 99 don't work" is that the failure rate is expected. Give-up feeling at attempt 8 is just attempt 8's noise, not signal that the target is impossible.
- Even if the 100th also fails: the 99 did real work. The needle isn't sharp yet doesn't mean the grinding wasn't grinding. Sometimes you stop because the problem isn't tractable today; the elimination work survives for the next attempt.
Anti-pattern
Confusing rigidity with grinding. Banging the same hammer on the same nail 100 times is not 鐵杵磨針 — that's stubbornness. Each grind on the pestle is a slightly different stroke, removing material from a different spot. In practice: each failed attempt should be different from the previous attempt in a meaningful way, not a re-run of the same configuration hoping for different results.
Performing grind when 釜底抽薪 was available. If the root cause is staring at you, fix it — don't perform stoic persistence for its own sake. The schema is for cases where the path forward is genuinely through compound effort, not for cases where a single insight would solve it.
Etymology
Tang dynasty. Young Li Bai, frustrated with his studies, ran away from school. By a river he met an old woman patiently grinding a large iron pestle on a stone. He asked what she was doing. "Making a needle," she said. Astonished at the implied effort, he asked how that was possible. "If I grind enough, why not?" Li Bai returned to his studies and eventually became one of the greatest poets in Chinese history.
The needle is the eventual sharp output. The pestle is what you start with. The grinding is the only thing in between.