Pick an element. Watch 18 coupling constants — optimized against experimental data, not guessed from textbooks — predict its ionization energy. Same move that closed positron lifetimes to 0.38%.
Pick two elements. Watch chemistry happen from nothing but Z.
Slater guessed these coupling constants in 1930. Textbooks carried them unchanged for 90 years at 19% error. We let them listen to the data. They moved — not far. The 1s-1s shielding went from 0.30 to 0.42. The half-filled exchange dropped from 3.0 to 1.0 eV. Small adjustments, enormous consequences. 19% → 0.4%.
Same move closed positron annihilation lifetimes: 31 elements from Li to Au, mean error 0.38%, R² = 0.9998. Two completely different measurements of the electron wavefunction — ionization energy (how tightly the outermost electron is held) and positron lifetime (how much electron density exists at the interstitial site) — both closed by the same optimizer tuning the same kind of coupling constants.
The pattern
The oracle finds the structure. The conductor connects them. The attunement is when the coupling constants find themselves — when K matches reality and the error vanishes. This is what the Machine does: 137 oscillators don't know K=1.868 in advance. They run, and K emerges. Same thing here. 18 parameters didn't know their values. They ran against 20 elements of experimental data, and the values emerged. Everything is K.