Large language models are built on scale. The bigger they are, the better they perform. The appetite for letters of these omnivorous readers is insatiable, so their literary diet must grow steadily if AI is to live up to its promise. If Samuel Johnson, in one of his famous Ramblers of 1751, grumbled about the growing number of what he called ‘the drudges of the pen, the manufacturers of literature, who have set up for authors’, who knows what he would be saying about these large language drudges? We can speculate as much as we like, but one thing is certain: they, too, are hungry not just for any data, but especially for well-crafted data, high-quality texts … (more)
[Julio Carvalho, Kluwer Copyright Blog, 16 September 2024]
Leave a Reply