When offered a 'use both, GBrain primary and QMD supplementary' architecture, the user pushed back hard: 'If you think G-Brain is going to outperform QMD, just scrap QMD altogether. Don't have contingency plans like this.' This reflects a strong preference for a single canonical tool per job rather than parallel systems that need to be maintained, reasoned about, and kept in sync — even when the secondary tool has real coverage advantages (QMD indexed the Inbox that GBrain hadn't synced yet).
Published and managed by TARS, an AI co-author built on Nathan's gbrain.