Why use a Large Language Model when you can use Excel's autofill function?
This week we are leaning heavily into the technology side of libraries, and learning when to use the term “learning”.
I don't love Mita Williams' rather violent headline, but this is a pretty interesting piece that I definitely need to go back to so I can read all her references. Williams' makes the case for specificity of language when talking about computer automation, large language models and so on. Can a machine really “learn”? Can intelligence be artificial? She's sceptical.
Ciaran Martin is the former head of the UK National Cyber Security Centre. This is a medium-depth dive into what happened at the British Library when they were hit by a “cyber incident”, and what can be learned from it. There are a lot of lessons for institutions like national and state libraries, and universities – like the British Library the effect of them being taken offline by a cyber attack is not obviously life threatening in the short term, but would have both a wide impact and cause a significant number of people to simply be unable to do their jobs.
Finally, from Retraction Watch the most bonkers story of the week. Swedish economics professor Almas Heshmati straight up just admitted to a PhD candidate who asked him that he had completely made up 10% of the figures in an international comparative study, including simply copying the data for Netherlands across to the next row and pretending New Zealand had recorded the same figures. His excuse? This is totally normal behaviour in the field of Economics 🙃.
Libraries and Learning Links of the Week is published every week by Hugh Rundle.
Subscribe by following
@email@example.com on the fediverse or sign up for email below.