The hyperventillators are drafting proclamations on a discovery far bigger than AI this week. If even partially true, #LK-99 will be one of the greatest discoveries of all time. It does not matter whether it is bigger or smaller, but its symbiosis with the #GenAI revolution promises a new future.
In some ways, it is far less about what superconductivity does for GenAI, but the inverse that carries the real promise.
The LK-99 spotlight illuminates the multifaceted world of synthetic molecules and the potential turbo-boost GenAI provides to innovation across sectors and nations.
Synthetic molecules are mankind's version of nature's building blocks. Crafted not in the heart of stars but in the labs of chemists and material scientists, they are meticulously constructed to unlock applications from curing maladies to ushering superconductivity and everything in between.
The creation process is complex but trivial when juxtaposed with the colossal task of discovery. We find ourselves in an expansive playground so vast that the term "endless possibility" seems woefully inadequate. The theoretical "chemical space" harbors approximately 10^60 stable synthetic molecules obeying the rules of chemical bonding. This vast space, if visualized as our solar system, the volume of what humanity has investigated to date would barely fill a shipping container.
Suppose Pb9Cu(PO4)6O, aka LK-99, has some promising characteristics. In that case, it would be like stumbling upon an intriguing clue, driving scientists everywhere to want to explore zillions of similar molecules for more promise. As per the PubChem database, humanity has accumulated some information on about 150 million unique chemical structures from time immemorial. Theoretically, a high-octane GenAI model could screen billions of molecules in simulations for certain properties in a single day.
Screening billions, even if craftily designed through the latest GenAI tools, in simulation is a far cry from testing even one in real life, but GenAI tools' ability to speed up research is extraordinary. As GenAI processes improve, what scientists can screen for desired characteristics in any synthetic molecule space will expand exponentially in the coming decades.
We have a solar system to investigate, so to say, for potential solutions to Parkinson's to advanced fertilizers to superconducting material. With the tools so far, we have looked in a container-sized space. As we unlock the latent power of GenAI, we equip ourselves with a potent, ever-improving scientific telescope-cum-microscope.
So far in financial markets, we are getting excited about the tools of #AI, and not about what they will lead to in other innovations. As discussed before in posts, GenAI use cases revealed so far are trivial. And, it is not about GenAI itself but what it will lead to. LK-99 could provide another investigative space for GenAI to prove itself. #innovation
