This is the core thesis of the essay. I have two problems with this statement. I think only a small minority of people reads books or original sources in the first place. The general audience relies on simplified and often wrong generalizations aimed at, well, the general audience. I also don't think researchers are at risk of "rewiring themselves" to rely entirely on LLMs. Science is suffering from an influx of AI slop papers, but it's not that different from all the weak or dishonest research that was being published before. Ultimately, our core academic institutions are designed around minimizing this and incentivizing high-quality intellectual output.