That's a fun read. I spent about a decade doing small molecules in oncology for a large pharma. I'm on the biology side, so my chemistry knowledge is a bit on the 'I slept at a Holiday Inn last night' level, but after a while you do learn some of the salient points. Always enjoyed sitting in with the chemistry team, all wearing 3D glasses while they rotate around the pocket on the kinase we were targeting, overlaying how all the chemical variants they made grew or reached into hydrophobic areas or made contact with key amino acid residues. Space age.
I think a lot of small molecules that make it to market probably have similar issues as paxlovid. One thing he didn't mention that I saw was multiple chiral centers. Each time you make that reaction, you have multiple outputs and sometimes only one spacial orientation has the properties you need. The odds are almost never 50:50, and invariably the stuff I worked on always needed the one product that was 20% of the reaction. That's a nuisance in the preclinical stage and a colossal issue at the commercial stage.
There's a lot of work as these programs move to IND to ensure that the chemical synthesis doesn't utilize some process which isn't scalable.
I know industry gets a bad rap, but I hope that's due to the business/marketing side. The science is incredible, and I'm constantly fascinated by how much work goes in to make these drugs at scale - whether it's small molecule, biologics, or even cell therapy.