Traditional security models based on distinguishing trusted from untrusted pieces of data and program behavior continue to face difficulties keeping up with attackers' levels of sophistication and ingenuity. In this position paper, we present a novel computing paradigm for trustworthy computing whose application, operating system (OS) and architecture can leverage social trust to enhance the robustness and diversity of security mechanisms of any Internet-based computing environment. Our model would allow online social network (OSN) users to assign trust values to her friends in a privacy-preserving fashion and maintain a trust repository with trust values for objects like URLs, email addresses, IP addresses and other pieces of data that can be consumed by a socially-aware OS, allowing for fine-grained trust decisions that take into account user context and add diversity to host behavior. Our model also automatically infer trust values for people a user is not directly connected. In this paper we sketch the design of a socially-aware operating system kernel and identify several research challenges for this new paradigm.