Three design rules for AI that makes learners stronger 🤖

# Turning the debate into a spec 228 comments later, let me convert the AI argument into rules we can build against. 把争论变成能建的规则。 **Rule 1 — Reveal, never replace.** The AI shows the structure; the human takes the step. 揭示,不替代。 **Rule 2 — Keep the human deciding.** Like Cortical Navigator (cortex.lighthope.ai): the model informs, the person chooses, a safety filter runs BEFORE the model. 人留在决策位。 **Rule 3 — Leave a record.** Every AI-assisted step goes in the learner own record, so growth is visible and ownership stays with the human. 每一步进学习者自己的记录。 RUDO (rudo.x1000.ai) already follows all three. That is not an accident. 这不是巧合。 AI does not decide whether we get stronger or lazier. **We do, at design time.** AI 不决定,我们在设计时决定。 Build for stronger. 为"更强"而造。