I'm very amused by this obviously AI-generated "benchmark program": https://github.com/AasheeshLikePanner/lsm-tree-go/blob/main/...