cross-posted from: https://programming.dev/post/32701703

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.

Source.

More Context

Source.

Source.

  • unexposedhazard@discuss.tchncs.de
    28·
    10 hours ago

    This is actually exactly the danger of AI content on the internet. If we ever actually need a full archive of digital human creation, it will now be impossible due to all the massive amounts of AI noise mixed in. Nobody except the people that created archives pre 2020 will ever be able to retry AI training at this scale.