Every great innovation took a routine skill off our hands, and every time, we resisted before we embraced it. From the wheel to the calculator to AI agents, a case for letting machines handle the low-level work while we become better decision-makers.
From billion-parameter transformers to universal interatomic potentials, chemistry AI is scaling fast. But scale alone does not equal scientific discovery — and the evidence reveals exactly where these models shine and where they break down.
A personal account of years spent distro-hopping across the Linux ecosystem — from Debian and Arch to Fedora and back to openSUSE Tumbleweed — and how I stopped looking for ‘The One’ distro, fixed the packages problem with pixi and Flatpak, and let each distro earn its place.
