> I've often considered trying to document it in some way, but I've no idea where to even start such an effort. LLMs could represent an interesting opportunity

> I've often considered trying to document it in some way, but I've no idea where to even start such an effort. LLMs could represent an interesting opportunity not just for linguistic preservation, but also to make smaller languages more relevant to their communities again. This article [0], despite its headline claiming the opposite, contains some examples of researchers training LLMs to that end, and how it's helped them reconnect with their mother tongues and relatives. [0] https://www.theatlantic.com/technology/archive/2024/04/generative-ai-low-resource-languages/678042/ https://www.theatlantic.com/technology/archive/2024/04/gener... discussed on HN: https://news.ycombinator.com/item?id=40143621 https://news.ycombinator.com/item?id=40143621