← Back to context

Comment by 47282847

3 hours ago

> Instead of trying to keep clean, the intent is just to spam like crazy with anything so nobody understands the core.

I don’t think this is humanly possible against machine learning. After all, it is specifically designed to weed through noisy data and identify patterns. It may delay discovery, but will at some point easily fall apart, by something as simple as a “filter out shitposting and deliberate pollution” prompt. Even more so when you guide it towards specific attributes.