Why ProfileScribe keeps fact-finding and claim-writing separate
Abstracts
Abstracts
ProfileScribe works from a source graph that tracks professional links, repositories, and writing, then normalizes what a crawler agent finds into timestamped facts before anything reaches a profile. A review queue sits between that evidence and any published update.
The reason for the separation is simple: a system that drafts about someone's work should not blur what a source actually shows with what sounds good. Keeping extraction and claim-writing apart makes it possible to check a draft against its evidence before it goes out, rather than trusting a summary on faith. That structure is the part worth making visible, more than any single update it produces.