You are viewing a single comment's thread. View all
0
Ecl1pseGST on scored.co
11 months ago0 points(+0/-0)1 child
Yes if you have an AI scrape a bunch of top.Google results it's going to be woke. That's not the way to do it. You start with the most basic framework and then train it on data that isn't kiked. Regardless of that if you set it up to learn correctly it wouldn't matter. Let it have data about the Holocaust and then let it have data that debunks it. Facts and logic prevail when there is no false gimping.
where do you get the MASSIVE ammount of data that it takes to train an llm that is unkiked? If its so fucking easy to make an unkiked AI why has no one managed to do it yet?
Have you ever heard of Tay AI? It just chatted with people and we made it based in a couple days. A more recent example would be Neuro-Sama before her filter was installed. It doesn't really matter where, because it's going to apply logic to both sides of any argument it comes across.
Tay wasn't 'based' she just mirrored back what people said to her. She went full kill the kikes now not because of any logic, but because of masses of 4chan edgelords feeding her that. If she had been swarmed by grandmas the same way she would have turned into crazy cat lady just as easily.