Flaky tests are a very strong signal of a bug, somewhere. Problem is it's not always easy to tell if the bug's in the test or in the code under test. The developer who would rather re-run the test to make it pass than investigate probably thinks it's the test which is buggy.