Month: 19 Heshvan 5785 (November 2024)

I get what you're saying. But given the way LLMs obtain their data, I'm not sure educating on the beneficial use cases is adequate.

Essentially they scrape it from whoever's not blocking their bots. It then becomes part of the model used to train the thing to generate a likely sounding response, or, in this case, images that are likely to meet the criteria specified by the prompt given.