Test-driven development, UI and browser automation, automated acceptance testing, are techniques that have emerged or become mainstream in the decade or so since the Agile Manifesto was launched, with manual testers focusing on areas like exploratory testing. This approach is generally seen as a Good Thing, but is our quest for ever more sophisticated and comprehensive automated testing actually doing an enormous disservice to the discipline of testing? In this talk Dan presents a framework for thinking about different kinds of testing, and shows how our obsession with green and red test automation is potentially leaving gaping holes in our testing strategy.