Current discussions about technoscience are mostly centered around reflections on the characteristics of technoscience in comparison to the traditional realms of scientific knowledge production. One of the characteristics is the yet unresolved status of technical artifacts as a potential representation of knowledge. Advanced products from rapidly evolving fields such as e.g. nanotechnology or computer science may significantly deviate from what is known in traditional engineering. They are thus setting up empirical basis that seems far from being intelligible. Technical artifacts with such autonomous standing could be viewed as paradigmatic for technoscience. In order to discuss this emerging «epistemic gap» we will present some advanced engineering scenarios that fully rely on computer-based automated design procedures using numerical structural optimization. Though such design approaches yield artifacts with potentially unparalleled performance these outcomes are often barely understood. Besides the proper synthesis aspect of computer-guided design the underlying optimization also offers highly formalized modes of observation for tackling this «novel» empirical basis in a framework of what we like to call «postprocessing».