Very nice. TY again.

On a personal level, my growing experience is that nagging sense about what's missing in reading an AI generated article, that odd sense of where's the center? It's not there.

We still don't understand how our own minds jump to the new from pre-existing data. Machine learning, of all sorts, uses electronic (rather than biological) neural networks to find patterns. It's wonderful. We got Alphafold out of it! But innovation, this human neural ability to create something new, is something we still don't understand. Artificial neural networks use a partial understanding of how actual neurons connect to form learning networks, but, as a biologist, I certainly know that we actually don't have an experimental understanding of how our own neural networks store, recall, and modify the representation of data to then birth a new network only related to the old one in a metaphorical way.

Sorry for the verbosity, but these things still are not understood. LLMs, based on a certain mathematical neural network, the Transformer, are pretty slick, but we don't have the whole thing yet, and my worry is we'll fool ourselves into thinking we understand more than we actually do, entrust more and more roles to Generative AI, and then kind of freeze in all the many areas where human understanding looks at existing data and can invent entirely new things out of it.