Editing Openai/686aedb2-60b0-8003-98a8-e1766cf496a6 (section)

==== What the authors actually show ====
* They collate work (much of it from Sutherland’s group) in which simple feed-stock molecules such as HCN, formaldehyde and hydrogen sulfide give, in one reaction network, activated pyrimidine nucleotides, 12 proteinogenic amino-acid precursors, and glycerol phosphate—all without enzymes. They call this “cyanosulfidic protometabolism.” Prebiotic Systems Chemistry by …
* They highlight cases where molecules sort or purify themselves by crystallisation, sublimation, phase separation or volatility; e.g. ribose-configured amino-oxazoline (a nucleotide precursor) crystallises out of a messy mixture, removing unwanted isomers. Prebiotic Systems Chemistry by …

My take
* The chemistry really does “assemble itself” in the sense that no cell or polymeric catalyst is involved—only small molecules plus UV light, minerals, pH shifts, etc.
* But the phrase still oversells things: yields remain modest, conditions are hand-tuned (pH 7 buffer here, UV there), and no experiment yet produces a self-replicating macromolecule. So “self-assembly” means spontaneous under lab-imposed conditions, not spontaneous anywhere, anytime.