This chapter takes leave of the idea that lawyers can remain immersed in legal text. It takes a stand for a careful reflection on what data-driven architectures do to some of the assumptions of modern law that are mistakenly taken for granted. Merely enacting the presumption of innocence by means of legal code will not do in the present future. If the defaults of Big Data analytics all point in the direction of precrime punishment or the pre-emption of inferred criminal intent, we need to reconfigure the smart decision systems that progressively mediate the perception and cognition of law enforcement and intelligence. Architecture is politics and code is law. This chapter starts with an analysis of different conceptions of law and technology, followed by a discussion of technology and neutrality in the light of the Rule of Law. After these explorations a relational conception of the criminal law is developed, based on Radbruch’s antinomian conception of law, highlighting justice, legal certainty and the instrumentality of the law. This is aligned with a pluralist understanding of technology to flesh out the implications of data-driven intelligence for the meaning of the criminal law. Special attention is given to the presumption of innocence that seems to be overruled by the affordances of data-driven law enforcement. Finally, the chapter explains the need for a ‘presumption of innocence by design’, thus translating some of the crucial affordances of the written law into the critical infrastructures of data-driven society.