> Has theoretical physics advanced enough now that such pen and paper discoveries are all but over, and the only way to continue making progress is to dedicate an ever larger share of the global economy's productive capacity to building larger and more expensive experiments?
That's a bit disingenuous. At the time, GR was an unconfirmed theory not unlike, say, String Theory is today. Except it only took a couple of years to confirm by experiment.
Particle theorists and cosmologists have plenty of theories. But deciding which one describes reality best can only be done by data, no two ways about it. And yes, since most low hanging fruits have been found, experiments become harder and harder. Not to say more and more expensive.
Your conclusion is correct though, that at some point a society has to decide whether they can afford further progress.
Perhaps we also haven't found a theory as convincing as Einstein's GR because the math isn't there yet. GR was discovered shortly after differential geometry was formulated, and without it it would have been impossible. Similarly with Newton's theory and calculus.
So maybe what we need is the right breakthrough in math?