A grand project to build a formal representation of all human knowledge! Not wikipedia, but knowledge in a formalized frame system, including everything from common sense facts about objects, money, botany, human relationships, and the economic activity of Indonesia. This was sort of the moonshot project of GOFAI, led by Douglas Lenat first at MCC (a big semi-governmental research lab in Austin) and later as its own company Cycorp which is still going.
This grand concept of a universal knowledge calculus made my computational senses tingle. Is this not what GOFAI aims at? Particularly in efforts like the Cyc project, the apotheosis of the dream of "knowledge representation". However, the logical formalisms of KR seem pretty crude compared to what Hesse is describing. Cyc was based on a notion of representation as formalized statements or graphs,; while the Glass Bead Game is an entry into a higher, more spiritual plane, a practice, not a body of knowledge.
Good Old Fashioned AI, the kind that actually tried to model human thinking with carefully designed symbolic computational procedures. Quite different in methods and purpose from the AI of today, which relies on machine learning from massive datasets. There is just a deep cultural divide between the two camps, which actually dates back much further than the recent triumphs of the ML approach. Goes back at least to Minsky and Papert's eviscertation of the neural/ML models of an early day in Perceptrons.
Whereas GOFAI is fundamentally discrete or digital (a proposition is either true or false), LLMs and other deep learning machines work via continuous mathematics, everything is a numerical probability and the space being explored is smooth (continuous).