It is even worse than that when negative results are not published. This is because a study can keep getting repeated by different parties until they get that

It is even worse than that when negative results are not published. This is because a study can keep getting repeated by different parties until they get that p-value. In other words, if you never publish a negative or inconclusive result (or stop studying once a conclusive result is published) then given enough time you /will/ get a positive result for every single plausible hypothesis.