I keep returning to this question: what is human ingenuity, actually, when you strip away the poetry of the word? Is it just problem-solving, the kind a machine can eventually replicate? Or is it something stranger? I think it’s rooted in curiosity and curiosity is rooted in boredom. So maybe ingenuity is rooted in boredom? Now that would be interesting.
I think about this most when I read about scientific instruments that never make headlines. Take the elisa kit. Most people have never heard the term, and if they have, they probably associate it vaguely with a doctor’s office. But the enzyme-linked immunosorbent assay, developed in the early 1970s by Eva Engvall and Peter Perlmann in Stockholm was not the product of a machine optimizing for an answer. It was the product of frustrated, curious people trying to solve a specific problem: how do you detect a tiny quantity of a specific molecule in blood or fluid without relying on radioactive isotopes that were expensive, hazardous, and inaccessible to most labs?
Their answer was almost stubbornly elegant. Attach an enzyme to an antibody. What’s an enzyme? It’s a biological molecule, mostly protein, that act as a catalyst to speed up chemical reactions of living things. Why are they important? Because digestion and breathing would happen too slowly without them to keep you alive. Important stuff.
So these scientists use enzymes to react with a substrate to produce a color change. Read the color, and you’ve measured something invisible. That’s it. No brute force, no massive dataset, just a genuinely clever workaround born from limitation. Milton wrote that the mind can make “a heaven of hell, a hell of heaven,” and I think of that line differently now, in a laboratory context. Scientists working with real constraints, real hazards, and real cost, made something genuinely useful out of that a problem. A problem that could have stayed a hell of expensive, dangerous testing mechanism.
What strikes me is how the elisa kit’s descendants still carry that same human fingerprint. Consider the hgh elisa kit, used to measure human growth hormone levels in clinical and research settings. It didn’t emerge from a model trained on millions of prior assays. It emerged from decades of researchers refining antibody specificity, testing cross-reactivity, adjusting substrate sensitivity, one stubborn iteration at a time, because growth hormone deficiencies and disorders needed a reliable, affordable way to be caught and monitored. That is not pattern recognition. That is judgment, applied under uncertainty, by people who had something at stake.
Could an AI system eventually optimize an assay protocol? Probably, given enough training data generated by human labs in the first place. But optimization is not origination. The elisa kit did not exist as a pattern to be found. It had to be imagined into existence by people willing to sit with a hard problem long enough to feel foolish before they felt clever.
And that’s a tough feeling. Sitting with something, feeling silly. I don’t think we as humans like that particular experience very much. Neither do we like boredom. But both are incredibly important for innovation.