I took this approach once and enjoyed the result. I filled out 10 copies of the template of a handwriting font generator and generated all 10 fonts. Then I wrote a python script to process a libreoffice document. If it saw the 'handwriting1' style anywhere in the document it would pseudorandomly alternate between fonts. Since uncanny resemblance of two adjacent letter is the biggest giveaway that a handwriting font is at play, I made sure my script would change the font within a word if there were two adjacent 'T' or 'S' characters.

I've since lost the code (it wasn't something I needed to often use) but with LLMs these days I'm sure I'd be inclined to build something better -- for instance, performing the randomization within a single font file, and using custom glyphs for adjacent 't' characters that might have a common crossbar, improved support for other languages I use, or rendering a particular case of my legal name as a signature.

And learning to write in 'fonts' (hands) like block-print is still a form of calligraphy.