I've never been sure of the use case of hashids. They're not really a security measure since they can be easily defeated; you might as well just expose the sequential ID at that point. If you do need something with better properties, you should probably be using UUIDv6.
If an attacker can generate sequential hashids, they can decipher the alphabet order that hashids use without needing to know the seed, and then use the seed to invert other hashids.
I wrote a library that generates short IDs with the goal of making the similarity between two codes have nothing to do with sequence order.
