pokemonsuper9
Established Member
But if you were wanting to train an AI, 6.5 million human written messages here (ignoring the 1 million in the Quizzes and Games section), it'd be an excellent dataset, especially where there's technical knowledge.I don't know if it's necessarily AI, could just be good old-fashioned scrapers.
The Enron Corpus is only 1.7 million messages and near enough every model has been trained on it due to the value of normal human discussion.
I'd be tempted to say if this keeps happening to just restrict it to only members, but obviously a lot of the visitors are guests so it wouldn't be fair to the actual humans. The way I found these forums in the first place was it showing up in Google search results.