There exists a massive dataset for /pol/ that can be used by text generation algorithms, such as gpt, to make fake /pol/ posts. There also exists a proof-of-concept site (https://playground-beta.gpt4chan.com/) that is using a trained version of the dataset to generate this fake text. To my knowledge, it is currently down to excessive requests. If soyteens can get their hands on the proper model file trained from this dataset, then we could set up scripts of our own to raid many chudcel areas using fake text and 'jaks.