If you peruse a typical conference proceedings or journal issue, "low end theorems/experiments" in the sense of slogging away on incremental work is almost all of what scientists do. The big stuff is pretty unusual.
But even those "low end theorems/experiments" are generally novel and build on fairly cutting edge work. On the other hand you can pick up an Introduction to RoR book and have a successful career without ever venturing outside what that book teaches. There are few science jobs that require no knowledge beyond what you'll find in a first year chemistry book.
I'm a bit suspicious of your claim about it being that easy to have a career in tech. (But also a bit intrigued -- I'm in academia and looking for a plan B, given the state of the job market...)
There are lots of dull and simple programming jobs around like writing CRUD apps for a small department at a large government office in North Dakota. If you can manage to learn enough to get the hang of the basics of RoR or Visual Basic or whatever, you'll be able to find someone somewhere that is willing to pay something for your skills. Those sort of jobs just don't really exist within the sciences. Of course the question is if you actually want to do those jobs.
You nailed it. CRUD apps have considerable value. Low end theorems/experiments are almost worthless.