A researcher at Cambridge was testing a photocatalyst — a substance that uses light to drive chemical reactions. During a control experiment, he removed the catalyst entirely, expecting the reaction to fail. It didn't. The reaction worked just as well. Sometimes better.
The team had been adding something the process didn't need. The light was already doing the work. The catalyst — the thing they assumed was essential — was partially getting in the way.
What followed, published in Nature Synthesis, is a new technique for modifying complex drug molecules using nothing but an LED lamp. No toxic metals. No harsh conditions. No expensive reagents. The light triggers a self-sustaining chain reaction that forges carbon–carbon bonds — the backbone of virtually every biological and pharmaceutical molecule — under mild conditions.
The method shows what chemists call "high functional-group tolerance," meaning it can change one part of a molecule while leaving the rest untouched. For drug development, that precision matters enormously. Late-stage optimisation — the phase where researchers fine-tune a drug to improve its performance — has traditionally required aggressive chemistry that risks damaging the very molecule being improved. This technique sidesteps that entirely.
David Vahey, the lead researcher, described the path to discovery: "Failure after failure, then we found something we weren't expecting in the mess." Erwin Reisner, who leads the laboratory, added: "Recognising the value in the unexpected is probably one of the key characteristics of a successful scientist."
The control experiment exists to isolate what is actually doing the work. You remove the thing you believe is essential and see what happens.
When the process continues without it, the discovery is not that you got lucky. The discovery is that your model of what was essential was wrong.
The reaction was always possible.
You were adding complexity to something that was already complete.
It's recognising when the thing you're adding is the thing in the way.
Signals are science seen from the space between. Where human contemplative practice meets AI systems and documents what shows up.