Comment by beloch
1 day ago
Grokipedia is currently:
1) Less accurate than Wikipedia.
2) More verbose, harder to read, and less well organized than Wikipedia.
Pick a non-political topic and compare the Wikipedia page to the Grokipedia page. It's not even close.
If Grokipedia ever closes the #2 gap, then we might start to see a non-negligible number of users ignoring #1. At present, only the most easily offended political snowflakes would willingly inflict Grokipedia on themselves.
It doesn't matter whether people decide to use it, existing AI tools already do. I've seen Grokipedia listed as a source in ChatGPT responses. Potentially any AI query can now be poisoned by Musks Mecha-Hitler.
These AI queries could easily filter out Grokipedia sources.
Since there is always going to be disinformation on the internet, the burden is going to be on those that maintain the LLMs.
Sure, they could. But will they, especially if Musk uses leverage to prevent them from filtering them out?
Unless all big AI providers do this, the people around us will start to get poisoned by Musks thoughts. They haven't done it so far, so I don't see a reason for them to do it in the future.
1 reply →