Extra Quality — Ab Initio Versions

The first implementation of a theory — no experimental fitting, no empirical parameters, just fundamental constants and equations.

So next time you run a fast, production-level calculation, thank the awkward, unoptimized, 1990s-era ab initio code that proved the physics first. ab initio versions

And if you’re building something new — start with the ab initio version. Even if it only runs on 10 atoms. The first implementation of a theory — no

We talk a lot about machine learning potentials, DFT surrogates, and foundation models for materials. But here’s a quiet truth: every new, truly predictive method still starts with an ab initio version. Even if it only runs on 10 atoms

Real insight emerges when you know exactly what you’re approximating. Would you like this adapted for LinkedIn, Twitter, or a blog format?

ML potentials are getting shockingly good. But they depend on training data — and that data comes from the expensive, “ab initio version” codes. When the ab initio version changes (e.g., higher accuracy functional, core-valence correlation), the ML model’s ceiling moves too.