link to archived Reddit thread; original post removed/deleted
most of the proofs ChatGPT 5 has given me are plain incorrect, despite OpenSlop's claims that it is vastly superior to previous models
Both of those can be true.