Computational Chemistry Using PyTorch

> Democratizing Deep-Learning for Drug Discovery

I have this pet peeve about startup culture coining awkward words to describe concepts that already have words (like "impactful"). This language is then of course cargo-culted and never considered.

But sometimes the coined words aren't just superfluous, they're completely wrong. Every person that writes "democratizing" actually means liberalizing because what they're actually doing is commdifying some skill/product/process. As far as I know, none of the current users of the word "democratizing" is increasing voting rights or representation to anything. Case in point - using this software will not magically grant me access to the Academy of Sciences or a university faculty board or anything like that. It might help me produce some research more easily and then with that research in hand I'm admitted access but that's not the same thing.

Why am I ranting about this? If you cargo cult language, most people won't notice, but you'll still be in a cult. And cults aren't fun for people inside or outside. Just my 2c.

Would you recommend RDkit for use with pyodide? I.e. vscode.dev (w/ the Jupyter and pyodide extensions) and JupyterLite run in a browser tab, which is slower but maybe ideal for students and instructors without time to get mambaforge installed for everyone first

Neat, thanks for these links. What kind of stuff have you been doing with it?

(Asking because I have a kid who's hardcore on the chem side and wants to get sharper on the comp side).