In electronic structure calculations, the electron-electron interaction is the hard part. Two electrons interact through the Coulomb potential, which depends on the distance between them. In a basis set of N functions, this creates N⁴ interaction terms — every pair of basis functions with every other pair. For large systems, this quartic scaling dominates the computational cost.
Reader note (2026-11-17): Checked against the full text of arXiv:2603.22646 (White; March 2026), which I had earlier held as matching. The reduction from O(N^4) to O(N^2) terms is in the paper. Two points differ from the essay. First, the diagonal form is an approximation, the integral diagonal approximation (IDA): the paper says it is not variational, and that in practice its errors are much smaller than the error from the basis not resolving the electron-electron cusp. Second, the essay says each basis function interacts with itself and its neighbors; in the paper the diagonal form is a density-density interaction V_ik between pairs of basis functions i and k, and the paper says the point is not that the off-diagonal integrals are numerically small but that replacing the full four-index integral by the diagonal form is accurate as a whole. For the radial gausslets the interaction is a set of two-index radial matrices, one for each multipole order L, combined with angular (Gaunt) factors. The paper's test is restricted Hartree-Fock for helium, with microhartree accuracy for fewer than 20 functions and about 1e-9 for about 30, plus exact diagonalization on atomic systems.
White (arXiv:2603.22646) extends gausslets — a rare class of basis functions that produce diagonal (two-index) electron-electron interaction terms instead of the usual four-index terms — from Cartesian coordinates to radial coordinates in three dimensions. The N⁴ problem becomes N². The diagonal property means each basis function interacts with itself and its neighbors, not with all other functions.
The construction is technically demanding: radial gausslets must handle the singularity at the origin and the boundary conditions of atomic wavefunctions. The resulting basis is compact — a modest number of functions captures the relevant physics — and the diagonal interaction terms make Hartree-Fock and exact diagonalization on atoms significantly cheaper.
The through-claim: the interaction bottleneck in quantum chemistry isn't fundamental — it's a basis set choice. Standard Gaussian basis sets spread the Coulomb interaction across all pairs of functions because they weren't designed to localize it. Gausslets are designed around the interaction, not the wavefunction, and the scaling drops from quartic to quadratic.