Interesting article. I note that what is actually being used to sift through the possibilities is a neural net. Those have been around a long time. IIRC, they were first studied before digital computers arrived on the scene, and are not to be confused with things like chatGPT, even though the latter has neural net layers in its architecture, as I understand it.
I remember when the neural net boat came in for the field of remote sensing. My department had a few people working on them in the late ‘80’s. By the early ‘90’s, they were well on the way to transforming the field. At a conference I saw a presentation on extracting geological information from airborne imaging spectroscopy of a mountain valley with vertical relief of a full kilometer. Suffice it to say that traditional methods of remote spectroscopy would have been brought to their knees when confronted with that problem.
That example seems like a piker compared to the string theory application. I dusted off one of my math physics texts and looked up Calabi-Yau manifolds, and sort of see where the loofah analogy comes from. I’m reminded that I really need to finish that text someday.
I had a colleague who knew both Calabi and Yau. He described Calabi as a true gentleman. His description of Yau was…not complimentary.