
I have a lot to say about the hype currently surrounding AI in total synthesis, but I will limit myself to this: I dare those who promote AI-driven total synthesis in industry and to funding agencies to bet their savings on the success of the synthetic routes proposed by AI. In my thirty years as a practicing synthetic chemist, having published more than 125 total syntheses as a PI, I have encountered only one case in which each step worked as planned. Everything can look good on paper (Papier ist geduldig). But, as Mike Tyson famously put it, “Everybody has a plan until they get punched in the mouth.” How you respond to the inevitable setbacks, surprises, and insults that chemistry throws at you is precisely what makes total synthesis such a challenging and rewarding endeavor. And it is the reason why total synthesis is such a good training ground for problem solvers in academia and industry. I do wonder, however, what all this means for writing proposals on total synthesis, which seem to be increasingly reviewed by colleagues who have little understanding of what total synthesis actually entails. Writing such proposals has always been a somewhat silly exercise, for the reasons outlined above. Evaluating them will become even more challenging in the age of Claude, when anyone can generate a seemingly convincing synthetic route on paper. The real question, of course, is what happens when that route encounters reality.

















