The Community Layer

Mathematical knowledge lives in communities as well as documents. A paper records definitions, theorems, and arguments. A community supplies the tacit layer: which proofs matter, which tricks transfer, which examples are misleading, which questions are exhausted, and which reformulations are alive.

AI can weaken this layer indirectly. If routine mathematical work is automated, fewer people may pass through the long apprenticeship that makes a field operationally usable. The archive can remain intact while the interpretive community contracts. In that state, the literature is formally accessible but practically harder to use.

Dead languages, obsolete engineering practices, and abandoned mathematical subfields can leave rich records that few people can operate. The AI version is more delicate because it may happen while the document layer is expanding. The visible signal is abundance. The hidden signal is loss of human navigability.

The issue is cumulative knowledge. A community that stores proofs but loses the capacity to recognize, compress, teach, and redirect them has preserved only part of the object. It has preserved the record while weakening the practices that make the record useful.

Blog Use

Use this after synthesis debt. Synthesis debt concerns the result layer; this file concerns the people who pay down that debt.

Source Trail

• Johan Commelin et al., “Shaping the Future of Mathematics in the Age of AI”, arXiv:2603.24914.
• Petra Schwer, “The meaning of doing mathematics”, arXiv:2509.15998.
• ../ai_views.md, “The community hollowing problem”.
• See reference_inventory.md for full source status.