We will use Grok 3.5 (maybe we should call it 4), which has advanced reasoning, to rewrite the entire corpus of human knowledge, adding missing information and deleting errors.
Then retrain on that.
Far too much garbage in any foundation model trained on uncorrected data.
Try rereading the whole tweet, it’s not very long. It’s specifically saying that they plan to “correct” the dataset using Grok, then retrain with that dataset.
It would be way too expensive to go through it by hand