AI Mathematician

An AI mathematician is the source’s broader target for AI For Math: a system that can prove, conjecture, formalize, use mathematical knowledge, and eventually help expand research frontiers. In 137. 对洪乐潼的4小时访谈：AI for Math、把数学变成Lean、数学天书中的证明、直觉、被创造与被发现的, Hong Letong / 洪乐潼 treats Axiom Prover as a step toward this target rather than the whole target.

Key Claims

• Proof is necessary but incomplete: mathematical creativity also involves asking natural questions, finding useful definitions, and generating conjectures.
• Human mathematicians may shift toward higher abstraction, taste, direction-setting, and benchmark design as AI handles more formal proof labor.
• Machine proof can be correct without being elegant; the system may first become useful through brute-force formal reasoning before matching human intuition.
• A true AI mathematician needs Auto-Formalization so it can read and convert ordinary mathematical literature, not only solve already-formalized statements.
• The source connects the idea to Mathematical Abundance, where math supply expands dramatically and mathematicians allocate attention and compute toward the best problems.

Connections

• Axiom, Axiom Prover, Hong Letong / 洪乐潼, and Ken Ono — main source cases.
• AI For Math, Interactive Theorem Proving, Lean Theorem Prover, and Mathlib — proof and knowledge substrate.
• Research Taste, Discovery Model, and AI For Science — adjacent discovery and problem-choice themes.