Several companies are implementing local-first software to empower their users to work offline with low-latency UIs. However, when building local-first applications that manage large datasets, it can be difficult to maintain data integrity without burdening the user (i.e. by requiring them to manually merge concurrent changes). To address this gap, Schieffer, Litt, and Jackson propose forking histories. This system merges most changes automatically but exposes multiple co-existing states to the user when conflicts arise. Users benefit from strong consistency within each history and can continue making progress until they are ready to reconcile the conflict.