Interesting article with connections to mathematical #logic But there is a mistake and an important missing part.
The mistake is the implication that "easy to verify, hard to work out" is the same as NP. The missing part is in the description of the graph coloring procedure: you need some way for the prover to show that she didn't change her colors in some way inconsistent with her chosen coloring. This is called a commitment scheme.
https://www.quantamagazine.org/how-unknowable-math-can-help-hide-secrets-20260511/
