Solve et Coagula

About this Volume

The title is the oldest motto in alchemy — dissolve and bind. Chemistry's two acts: take things apart, put things together. Substance from substance.

Volume II extends the same architecture from biology into small-molecule chemistry — drug design, reaction prediction, ADMET. Same substrate, different domain.

The model and the book ship together when both are ready.

On chemistry

Bonds form quietly. There is no announcement when two atoms meet at the right angle and decide to stay; the binding happens before anyone is watching, and then a new substance exists where two old ones did not. Chemistry is the science of this private event, repeated at scales we can barely picture, governed by rules we have spent two thousand years describing.

The alchemists thought of it as marriage. The Latin tradition kept the word — coniunctio, the joining — alongside its sibling, solutio, the dissolution. Modern chemistry retired the metaphor and kept the operations. Substances dissolve in solvents; substances bind into compounds. Whole industries rest on getting this right at scale.

For our purposes, the question is whether the same architecture that folds peptides can also propose bondings — molecules that will bind a given target, synthesis paths that will actually run, properties that survive the body. The bet is that it can, because the underlying problem — finding the configuration that minimises a certain kind of cost — is the same problem dressed in different clothes.

What we expect Volume II to do

Design candidate small molecules for a specified target. Predict whether a candidate will bind, and how strongly. Predict downstream chemistry — solubility, metabolism, toxicity — at the resolution where it stops being chemistry and starts being biology. Suggest synthesis routes that route around the dead ends.

What we will not claim on day one: that the model replaces a medicinal chemist. It will join one.