We should make the benchmarks we do collect for hyperlight more visible, by: - [ ] Making a public dashboard for them - [ ] Calling attention to how PRs change performance in both micro- and macro- benchmarks