I think you're looking at this the wrong way.
Today, only about 20 million people[1] are building software for over 3 billion internet connected people[2] (and even more who can get access to software in other ways). That's only about 0.6% of users contributing to the rest. Tools like bubble.is are not replacing the very scarce resource of software engineers, they're just vastly expanding the bubble (har har) of potential developers.
The world will always need people (or at least creative robotics) capable of designing and building hardware as well as the software that interfaces with that hardware. But the amount of resources spent on those critical systems will be minuscule compared to the resources spent building very custom things on top of it.
It's sort of like how you may be able to build your own computer at home with pre-made components and plug it into the wall and have it start up, without having to know how to generate the electricity required to run it, or how to make the hardware components.
In 2020 farmers may very well be building their own application, and if that is the case then most likely the number of people using these "WYSIWYG" tools will outnumber the kind of software engineers we see today. I personally think 2020 is a bit optimistic, but it's clearly the direction in which we're going.
[1] http://www.techrepublic.com/blog/european-technology/there-a...
[2] http://www.internetworldstats.com/stats.htm