I'd assembled Sol-20 computers for my local college while I was in high school (it's a long story), and my intent was to become an electrical engineer. I got seduced by the software side, and by the time I graduated from high school I'd switched my plans to computer science. But I remained an electronics hobbyist.
I came across the Intel 4004 design documents a few years ago and found them fascinating. My first thought was to implement the design in Verilog, a hardware definition language I was teaching myself, with the idea of building a Busicom 141-PF replica around a FPGA experimenter's board. As I learned more about the 4004 and FPGAs, I came to realize that a direct implementation would be impossible.
The other problem was that Verilog is a programming language. It's unlike the typical programming languages I use daily, but once the thrill of learning something new goes stale, it's still writing lines of code, loading it into a chip, then recoding to fix the inevitable mistakes. I've done this for 40+ hours a week for about 30 years, and it's not what the distraction from work a hobby should be. I wanted to make printed circuit boards, solder components, and play with oscilloscopes and logic analyzers.
In short, I wanted to do something different, something where I'd be learning something new at every turn.
Thus the idea of building bits of a 4004 took shape.