Software as an engineering material: how the affordances of programming have changed and what to do about it (invited industrial talk)

A contemporary programmer has astonishingly abundant processing power under their fingers. That power increases much faster than research into and published results about programming techniques can change. Meanwhile, practitioners still have to make a living by adding value in capital-constrained environments. How have practitioners taken advantage of the relative cheapness of processing power to add value more quickly, to reduce cost, manage risk and please customers and themsleves? And are there any signposts for where they might go next?