Gregory's Serverone of the few “good endings” i can imagine for the future of LLMs is that they get people to see the value of information retrieval, indexing, and archival — i.e., library sciences — and we start funding that community to translate their work to similar applications, but with proper attribution and dataset transparency.
 8 comments
 @oblomov @chrisamaphone I wonder, though, will we ever get to see the armies of poorly paid humans who label the sources that the algorithms are trained on? Even in the libraries and archives and museums, cataloguers and palaeographers are so often kept invisible. I suppose that it doesn't help their cause that so many of them prefer it this way. I fear that the public doesn't see that actual work is needed, and that that work takes intelligence and creativity that no LLM has.  @libroraptor I read @chrisamaphone's message in a different way, that LLM may lead people to re-evaluate library sciences *by contrast*, rather than by becoming the interface through which people see it. @oblomov @chrisamaphone I'd like that to happen, too. Though I have worried for a long, long time that libraries and librarians, and museums and curators, and archives and archivists, work ever harder to make that work invisible. Does the AI-generated stuff need to get markedly worse to trigger a realisation, do you think? I mean, do we need to get people disappointed in contrast to being impressed?  If library science were invented today by a billionaire and marketed like an LLM, a "chat with someone who knows" button would appear on every web site. (And the billionaire make a cut on every knowledge worker salary.) But then again, what interest would a billionaire have to start hyping up well-funded and well-founded knowledge? He wouldn't really be able to capture the market, because these pesky humans might just teach *each other* library science 😱.  @chrisamaphone I think Larry Page and Sergey Brin closed the door on that when they transformed a breakthrough in information retrieval into an advertising channel so many years ago.  @chrisamaphone man, I hope you're right. I'm a librarian and a big portion of my work in the next year is going to be exactly this. @chrisamaphone this is one aspect of the world that Neal Stephenson explores in "Anathem". All people who work directly with "syntactic devices" (see: computer) are almost entirely dedicated to sorting bad information from good. |
@chrisamaphone it's a beautiful thought. Hopefully it won't also mean that those will be relegated to inaccessibly expensive tools.