FOL has serious limitations as natural language semantics (1).

But with a good ontology you could certainly get relatively far, the authors have formulated it this way:

"Correct identification of background knowledge is crucial for our method"

Has there been progress in recent years? Or is Doug Lenat's CYC still state of the art?

(1) https://people.engr.tamu.edu/ioerger/cs420-fall23/09_Uncerta...