Comment by jonas21
I think the main reason for sunsetting the project is hinted at near the bottom:
> The field I know as "natural language processing" is hard to find these days. It's all being devoured by generative AI. Other techniques still exist but generative AI sucks up all the air in the room and gets all the money.
Traditional NLP has been surpassed by transformers, making this project obsolete. The rest of the post reads like rationalization and sour grapes.
I think the reason to sunset the project is actually near the top.
> I don't think anyone has reliable information about post-2021 language usage by humans.
It's information about language usage by humans. We know the rate at which generated text has increased after 2021. How do we filter this to only have data from humans?
The bottom is just lamenting what's happening in the field (which is pretty much what everyone that's been doing anything with NLP research is also complaining about behind closed doors).