How do successive, forward-looking agents experiment with interdependent and endogenous technologies? In this paper, trying a radically new technology not only is informative of the value of similar technologies, but also reduces the cost of experimenting with them, in effect expanding the space of affordable technologies. Successful radical experimentation has mixed effects: it improves the immediate outlook for further experimentation but decreases the value and the marginal value of experimentation in a longer term, resulting in less ambitious `incremental' experimentation and in a reduced size of radical experimentation. Incremental experimentation lowers the option value of similar technologies, which may spur a new wave of radical experimentation. However, experimentation eventually stagnates for all parameters of the model.