Reading the comments here, I don't think the commenters are fully grasping Adams' point. Especially interesting is his desire that the hole-digging robots be self-replicating, so as to allow for ever greater numbers of replacements, lower costs, and so on. It reminds me of a post I read regarding the feasibility of space colonization written by an engineer who went on and on about the "long tail" of industrial production -- basically, there are so many thousands of parts in any sufficiently complicated piece of technology that there are whole billion-dollar industries employing thousands of people just to create a subset of the required components. Cars are a perfect example, although the idea has implications for many technologies we use (computers, anyone?). An interesting approach to this problem (I read about it in Make magazine) is the RepRap Project, at http://reprap.org. They are trying to solve the "long tail" manufacturing problem by creating an inexpensive machine that can be used to build almost anything, including copies of itself. Perhaps something like this will lead to the future Adams is hoping for...
When I saw this title, I expected the post to be about enterprise software. How, instead of buying and installing enterprise software that costs millions and never works, an entrepreneur could just offer to dig them into a hole for a third of the price.
But surprisingly, Scott Adams actually meant real, genuine, physical holes.
If I could license a hole digger, the first thing I would do is connect LA to New York the short way. If I'm really lucky they will have invented a room temperature superconductor by then so my now nearly frictionless train can fall across the country at break-neck speed with near zero energy expenditure.
