I don’t trust their outputs yet enough to not review everything myself. But I’m happy to be. a QA bottleneck in that way. What I want now is orchestration to get them running in parallel. I’ve looked into Yegge’s beads, but I want something similar that is far simpler, fewer sloc, and database-driven rather than json-file-driven. Probably need to build it myself with a graph database…