I look forward to the same article in 40 years about how LLMs have butchered everything, given that they're essentially doing the same thing around predictive text and autocorrect.