The way I did this was by, for the most part, following along with the videos. I implemented the ALU, registers, and a 7-segment display driver.
I then expanded on the registers, using multiplexers and demultiplexers to first make an addressable collection of four 4-bit registers, and then a collection of four 4x4 RAM modules. You could expand those to be much larger, but since I'm making a 4-bit computer here, I didn't.
I used the 7 segment display driver from the third episode, and modified the bus test from the most recent (fourth) DLS video.
I integrated the 16x4 RAM unit I made into this test, and added a third register -- the Memory Address Register, or MAR, which controls what address the RAM is looking at. However, you may be wondering why it's "almost" a 4 bit CPU, and why the RAM isn't providing instructions to the CPU. The "CPU" is almost a CPU, because it's not quite finished. As of now, the chip only takes in control signals to do things, which I'm referring to as microcode. My next step will be to create a module to translate instructions from possibly a larger RAM, like 8 bits instead of 4, or simply program instructions as pure microcode into something like 16-bit-wide RAM, as well as create a binary counter to step through programs.
However, the "CPU" as of right now is functional, insofar as being able to run programs if you were to punch them in by hand.
I'll put up an instruction set once I figure that sort of thing out.