The post offers a brief explanation of homomorphic hashing:

> homomorphic hashing answers the question, “Given the hash of an input, along with a small update to that input, how can we compute the hash of the new input (with the update applied) without having to recompute the entire hash from scratch?” We use LtHash, a specific homomorphic hashing algorithm based on lattice cryptography, to create an efficiently updatable checksum of a database.

Imagine adding and subtracting hashes!

> For any two disjoint sets S and T, LtHash(S) + LtHash(T) = LtHash(S ∪ T).

This is cool because now you don't have to recompute from scratch a hash representing a large array; you can just compute the hash of the parts you update and perform addition and subtraction.