Is it crappy or perfect?

The fact that you focus on the features that are most important is because they contribute to the hypotheses that you want to prove. However, this does not mean that your product has to be crappy. Using a timeboxed approach will help you deliver often (daily or weekly) and on time.

The features that you will make available are unlikely to be perfect, but with each new iteration you can improve them. Of course, you will never have a second chance for a first impression, but aiming for perfection is not going to help you prove your hypotheses. Instead, it will prevent you from getting feedback early. Still, it is important that you choose your first users carefully. Early adopters are very different from mainstream users, having different expectations. Managing expectations is therefore very important when asking early adopters to test your solution. Be honest about the phase your startup is in and tell them that the solution has been built to maximize learning and that you would love to get their feedback. It may sound a bit harsh, but ultimate perfection does not exist anyway. The opinion of your early adopters is important and your opinion does not really matter.