Your goal at this stage is to conduct an experiment, maximize learning, and minimize the time it takes to complete one turn of the build-measure-learn cycle. With the feedback loop as your bull's eye, you will find yourself engaging in activities that may seem superficially illogical, but which greatly accelerate your learning.
For instance, it probably seems extraordinarily inefficient for CEOs or founders to interact directly with customers. If a startup wants to shorten the feedback loop and understand its customers' needs on a deeper level, though, the purpose of this tactic becomes clearer.
Similarly, it may seem wasteful to hand users an interactive wireframe or to present a prototype made up of coding shortcuts and workarounds. When viewed as a stage in an MVP test, however, such an approach makes more sense...