Hey @jack or other #bluesky and #atprotocol folks, how are the personal data repositories going to be kept in sync? Presumably, they’d be used by multiple clients in the network, I could imagine potential “merge conflicts,” unless they’re accessing the same instance of the repository? That seems prone to race conditions and other inconsistencies.