Yesterday Open AI announced solutions to ten open math problems, leaving only the coding‑theory one as a real hurdle. People will spin coping mechanisms, yet the reality of rapid AI growth refutes each stance.

Common copes include directing AIs at problems, teaching them, canonizing their findings, or becoming students of AI math. In practice, the AI will outpace human intuition, build its own “cathedral” of proofs, and no human audience will demand explanations of frontier results.

The chess analogy shows why AI superhuman chess is harmless: there is little external need. Mathematics, however, powers science; a better AI mathematician is materially useful, whereas a better chess engine is not.

AI also alters the social context of coding and math. The discourse degrades, enthusiasm wanes, and intrinsic motivation alone cannot sustain long‑term exploration—human ambition thrives on status, usefulness, and shared outcomes.

Technology isn’t inevitable. sovereignties can be chosen. AGI may accelerate progress or turn us into pets of powerful entities, but the future remains uncertain.