I'm talking statistically in terms of number of issues that reference segmentation faults in the Bun vs Deno GitHub repositories. Anecdata doesn't mean much.