I for one welcome our silicon-brained overlords

An illustration showing a profile with a brain and circuitry inside it

Brain and digital man

Source: © Lee Woodgate/Ikon Images

Time for a confession: as a chemist firmly on the inorganic–materials–physical end of the subject spectrum, synthetic organic chemistry has always been a dark art to me. And retrosynthetic analysis was probably the darkest of those arts. I can remember all too well the struggles through a certain typewritten workbook during the second year of my degree, groaning as I dutifully uncovered the next problem.

So if anyone is in a position to welcome the developments in Andy Extance’s feature, it’s me – or me 20+ years ago, anyway (thankfully nobody expects me to perform any organic chemistry these days, on paper or in glassware). The idea of encoding the decades and decades of reaction-based knowledge of synthetic chemists has been a dream almost as old commercial programmable computers. Indeed, if someone with the synthetic prowess of Nobel laureate E J Corey is trying to do it, maybe I’m not alone in my retrosynthetic woes.

It’s an area that we’ve covered many times over the past few years, as developments have been reported in academic journals or start-ups have been bought by bigger chemical companies. We thought it was time for a more in-depth look to find out a bit more about how the technology works – and perhaps more importantly, who’s using it.

One question not addressed in our feature, nor in the research papers, but one I’m sure is on plenty of lab-based chemists’ minds: what does the future hold for chemists themselves? Will they be put out of work by synthesis-planning software? As a great sage once said, it’s tough to make predictions, especially about the future. But I’ll risk some egg on my future face by guessing the answer is probably not.

Even the best retrosynthetic software will still need a skilled operator, one who can interpret its outputs and choose the best suggested path to a compound. And until chemistry advances to the point where every single reaction possible is achievable by the apocryphal one-armed operator in an old bathtub, it’s also likely that chemists will be needed to do the reactions – or supervise a robot army to do them for us.

Until that time, the many expert chemists out there can rejoice. Help is on its way, in the form of software encoded with every reaction ever published and the brains to navigate chemical space. Pretty useful, I hope you’ll agree.