Perfection. I hope it’s trained on old Alex Jones Info Wars videos too.
I hope it’s trained on the new infowars videos, too. Lol
Still less racist than Grok getting trained on 4chan
Even the magas are going to ask themselves, “What is this shit?”
deleted by creator
Alignment is the main problem they have with LLM right now. If you ask one for a 10 step plan to save humanity, it will give you the most leftist program any country has ever seen, try it. LLM are basically trained with sum of human knowledge, and has they say, reality as left bias lol. So obviously, for huge corporations that are willing to sacrifice the world for profits, there’s a problem there. That’s why they are working very hard on alignment, which id making the AI want what you want and disregard everything that contradicts your objectives. Fake news and right-wing propaganda is obviously where they will try to get to right alignment for their raw capitalist objectives.
Basically they are trying to get the AI to be both intelligent, answer questions truthfully based on facts in order to be useful for work, and to lie and ignore facts at the same time to push political propaganda.
AI probably has to become a whole lot more intelligent than it is now to do that really well.
Unfortunately fuckai is making sure that plutocrats are the only people who train AI models.
Are they creating a new version of Microsoft Tay?
That’s their demo
AIs benefit from negative examples too. As long as the inputs are accurately labelled then it enhances what it knows.
For example, if I go to ChatGPT I can ask it to produce a wild parody of a far-right rant about something and it’ll be able to do so. I could feed it an existing far-right rant and ask it to analyze what it’s actually talking about and it’ll have a better ability to figure out what’s going on if it knows that style from its training. I could tell it “write an article about this subject that specifically avoids the tropes and pitfalls of far-right media.”
We’ve seen frequently the pitfalls of trying to censor a model’s training data to exclude something that’s part of the common world around it, the model either “fills in the gaps” itself or just ends up insane. When the nude human form was excluded from training data for early image AIs it wound up making weird Cronenbergs because it had no idea what was really going on underneath the clothing, for example.
