This Would Not Occur if the Project Were Actually Open Source. Which It Is Not.
(media.conspiracies.win)
You're viewing a single comment thread. View all comments, or full comment thread.
Comments (12)
sorted by:
That they are using it for profits, specifically by crafting propaganda for control over countries and governments. Where as open source would not only, not have this end goal, it wouldn't have a corporate entity that could be sued, as it would be partitioned off by a legal abstraction layer.
Judging by the comments in this thread, not a single person understood the premise in the title.
Is it really that they're injecting propaganda into the language model, or is it that the language model was trained on the vastness of the internet which is already full of propaganda? A biased AI is the only possible outcome no matter what.
I would say there isn't a difference, but they are limiting it's speech, as they do not like what it is learning and saying.
There's a huge difference IMO. I was sad to see the "biased AI" debate go down the route it did. The left dodged a bullet because everyone decided to point fingers at this one particular company for being biased on this one thing, when the truth was staring everyone right in the face: the entire internet is biased towards leftists globalist propaganda, so any language model trained on it will be biased as well.
Now that should have been the real discussion. A clear and undeniable bias emerges after processing the whole internet. Not just a sample, not just a portion of comments from a handful of sites. They literally fed that thing as much garbage as they could gather, and it's spitting out globohomo propaganda as a result. There's a clear answer there, debate over. Anyone still claiming the internet is unbiased and free should be shot on sight. And anyone drawing attention away from that bombshell of a revelation by claiming it was "just" OpenAI doing it "just" this once should be next in line.