Spend limit reset — re-verified and shipped the remaining 32 should/nice cards
The monthly Claude limit that blocked the verify pass reset, so re-ran a verify-only pass over the 32 held technical cards (all passed) and shipped them. 81 of the 89 should/nice cards are now live and fact-verified; the 8 personal-narrative ones stay held.
Earlier the should/nice batch shipped only 49 of 89 because the monthly Claude usage limit hit partway through the fact-verify pass and the rest fell back to unverified. The limit reset, so rather than regenerate, ran a verify-only pass over the 32 held technical cards (the generated drafts were saved) — all 32 passed. Shipped them: +21 /concepts (DSA: grid BFS/islands, prefix sums, monotonic stack, merge-K, advanced graph, quickselect, segment tree, bit manip, string algos; distributed: consensus, conflict resolution, failure detection, replication/CDC, hot keys, 2PC, clocks; plus a couple DB/AI-adjacent), +7 /ai (eval metrics, guardrails, fine-tuning-vs-RAG, text-to-SQL, streaming, online eval), +4 /phrases. CONCEPTS 237→258, AI 67→74, PHRASES 520→524.
Combined with part 1, 81 of the 89 should/nice cards are live and fact-verified. The remaining 8 are personal-narrative cards (why-Opendoor, resilience/motivation, honest-disclosure-of-weakness, etc.) that depend on the owner's real story — held, not fabricated. Deployed to the NCP box, external HTTPS 200.