A performance claim is easier to judge when you can run the code. The string construction benchmark compares +=, strings.Builder, and a pre-grown buffer. The case study project links the source and tests. Treat the reported numbers as measurements of that setup, then measure your own workload.