FPGAs just allow you to remove a layer or two of abstraction and corresponding latency (vs a CPU) where you don't want/can't have it and gives you control, speed and/or efficiency improvements in exchange. The price you pay is you're dealing with things at a much lower level (i.e. bits and wires)... if you don't need the trade-off, don't make it. They are for the situations where there is no off-the-shelf solution that quite solves your particular problem adequately.

I can't speak for the OP, but for myself, my motivation for learning FPGA's is to find out what they can do, that I can't already do on a microcontroller. I'm already pushing a screaming fast MCU to its limits. The engineers look at my designs and say: "You're trying to do stuff in an MCU that's easy in a FPGA."

I believe one big advantage is that most FPGAs can have multiple clock domains. So while everything on the Propeller is tied to the system clock (and the system clock is hopefully fast enough that you have the resolution that you need to do the right thing at the right time), FPGAs can simply use clocks running at different frequencies.

FPGAs offer unlimited parallelism.

Up to some limit that varies from device to device.