In this essay I discuss the political economy of data-driven platforms in terms of monopolies and monopsonies, arguing that the concentration of buying and selling power builds on and extends a pseudo-omniscient data architecture that feeds on an increasingly seamless data ecosystem. As the mathematical underpinnings of data-driven architectures are further extended into the hardware of smart homes, driverless cars and smart cities, they may at some point diminish or eradicate the semantic discontinuity that provides room for dissent and dissonance. Such discontinuity depends on a text-driven environment, whose affordances such as dissonance and adversariality cannot be taken for granted in the pervasive data ecosystem that drives the political economy of platforms. I contend that we need a small set of primitives of legal protectionto shoot holes in the semantic continuity that is generated by the backend systems of data-driven platforms; competition law, consumer law, and tort law cannot contribute to the effective contestability of algorithmic decision-making by themselves. They require effective and practical rights to data minimization and purpose limitation and an actionable right to not be subject to automated decisions without a contestable justification. Once these primitives of legal protection are operational both the pseudo-omniscient nature of data architectures and the seamless nature of data ecosystems will be ruptured, thus reinstating the interstices for innovation, and for meaningful individual and political self-determination.