Radical innovations prompt significant subsequent technological development and exhibit novelty and “architectural” innovation, i.e. rearranging the way design elements are put together in a system. Thus, radical innovations often serve as the foundation for new technological systems, industries or domains and are seen to involve significant conceptual breakthroughs, through either luck or genius. The much more common incremental innovations are perceived as mere improvements to existing technologies. Decreasing returns from incremental innovation are understood to motivate a search for a new radical innovation to provide a platform for more incremental innovation. However, deeper study shows that the conceptual “distance” a radical breakthrough travels is far shorter than would initially appear. On closer inspection, several innovations with undoubtedly radical effects comprise several small inventive steps that appear self-evident, even logical, to the developers. This conundrum appears to stem from conflating a radical effect with a radical development. What's more, this view of radical innovation views inventions as isolated from the broader currents of technological development. An alternative view sees innovations as embedded in a co-evolutionary socio-technical landscape, where inventions develop in a technological environment and become building blocks for further inventions. Although only inventions adopted for use can be called innovations, “inactive” inventions can also serve as building blocks. In this view, (almost) all steps to innovation are incremental, but the system's self-organized criticality (SOC) allows spontaneous radical effects. This work explores the importance of an evolutionary and SOC view of invention and innovation through agent based models. We develop a simple model capable of simulating the build-out of technologies in a series of simulation experiments. Although all inventions develop incrementally, the model behaviour is expected to exhibit SOC so that some inventions trigger much higher rates of subsequent development than others. Thus, the results are expected to support an evolutionary view of technological development where a radical effect is a consequence of the entire interdependent landscape rather than of the radical development of a specific innovation. The model should be of interest to several research streams concerned with simulating and studying R&D activities.