Strongly held belief that I think is backed up by academic consensus: we should not treat LLMs as stores of knowledge. As my mantra goes, “language models, not knowledge models”. So the future of search might involve LLMs at a very fundamental level (and I think it will!), but they’ll never be the central component. Humans will never ever ever invent a better knowledge system than a database / a piece of paper, I guarantee.
Agreed but there's nothing about this discussion that requires LLMs are the store for anything. That's an implementation detail. One that's not that controversial imo. The current tech is clearly too destructive to "store" anything.
