I suspect "radically change the economics of software development" means "drastically reduce the cost to achieve any of a large set of goals we believe our technology covers". Reduce by how much? Maybe they have proposed answers to that question that they are willing to share publicly, but such things can be difficult to quantify today, even with over 40 years of world-wide experience of teams using different programming languages, tools, processes, workflows, etc.
How do they propose to achieve this? I would not be surprised if the linked articles on the red-planet-labs web site might be the extent of the public answers to that question. Lots of startups remain in 'stealth' mode for a while, being very very careful to let out details of what they are doing, lest some other company use the ideas themselves and run with them.
Andy