Versions Updated - Ab Initio

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.

Here’s a draft for an interesting post about ab initio versions — tailored for a computational chemistry, materials science, or ML/physics audience. Why “Ab Initio” Versions Still Matter in an AI-Driven World ab initio versions

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

The first implementation of a theory — no experimental fitting, no empirical parameters, just fundamental constants and equations. Here’s a draft for an interesting post about

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.

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