William Dembski claims that the fine-tuning supports the inference that the universe was designed. His ‘design inference’ is based on the identification of two features of the fine-tuning. Dembski claims that it is a ‘specified’ event of small (a priori) probability. Specification, in this context, is the ability to describe an event without using any knowledge of the actual event itself. I argue that we currently do not have the ability to describe accurately the fine-tuning of the universe without using any knowledge of the fine-tuning itself. If we cannot generate a specification, then the fine-tuning is not a specified event, so the ‘design inference’ is not justified.