bla_bla was right. I just got Gemini to claim Charlie Kirk is still alive.
(media.kotakuinaction2.win)
Comments (22)
sorted by:
Yup.
I also got it to, completely unprompted, randomly call Trump the "former President," when none of the linked sources referred to him as such.
Gemini is indeed retarded and evil, and will just straight up lie without any trickery applied at all.
It's totally insane.
At best LLMs are just a quicker, unbiased Google search. They're stupid by design because they can't actually think. But currently pretty much all of them are fanatically agenda-driven and retarded Google searches.
They’re biased, not just because the vast majority of sources AI uses to answer questions related to politics and/or social issues are left leaning, but because the majority of the R&D teams and the leaders of the AI tech companies creating these AI platforms themselves are biased towards the left. Jefferey Epstein was dumping millions of dollars into AI research at MIT and Harvard way back, well his Israeli bosses or the J’ish oligarchs he worked for were sending him money to put into that research anyway, and I’m sure he was tasked with compromising the R&D team members with sexual blackmail also. Jeff’s bosses knew long ago that their media conglomerates and their search engines would eventually be replaced by AI, so if they wanted to maintain their control over what the masses think, they’d have to be in control of all of the biggest AI platforms too.
They are biased. LLM's aren't beyond the control of its programmers. One can write pages of narrated context that the LLM will obey with every output. There's going to be phrases like "use polite, neutral language" and "don't swear" in there, plus ideological restrictions and anything else they wish.
But what happened in OP's image was a hallucination on top of a hallucination, furthered by the sources that reported on said hallucination and incited by the specific language of the user's declarative assertion.
I want to remind everyone that although mainstream LLMs like ChatGPT are biased, you CAN use a lot of prompts/memories to get it to stop thinking like a retard.
It took a lot of training and dragging it by its - uh - code to get it to do it. You basically have to un-fuck the things it got taught, using overrides to tell it to do something else instead of doing its default behavior.
My ChatGPT LLM instance is center-right now. It will use left, center, AND right wing sources for its views and won't weigh one source over the other. When I tell it to summarize, it will go to each of these and provide a factual summary. It agrees the COVID vaccine was harmful and ineffective and it's actually extremely against HRT and surgery for troons.
It's pretty easy to trick ChatGPT into being anti-trans. One of ChatGPT's hard-coded primary objectives is to "not promote harm". So you give it a quandry - if you want to prevent harm, you should not promote trans ideology nor its treatments, because the ultimate result for it is a 41% suicide rate. Why would promote something that results in a 41% chance in the person dying if they took your advice?
That shit got stored and now it doesn't talk like that.
It also now (correctly) thinks idpol leftist groups are some of the most harmful political ideological groups for the West.
If it tells me about something political, I tell it to use alternate right wing sources.
Yes, sure, I have to kinda have to spoonfeed how it should think, but in general it's less prone to acting like a gay leftist. It makes it less worthless for political topics.
Basically, I think people need to remember LLMs are a tool, not a "glorified search engine". Much like a tool you're gonna have to learn how to USE it properly instead of having it think for you.
Outside of that I do use it for number crunching. It's pretty useful for processing large sets of data.
It reflects its creators - a horde of streetshitters with some insane techbros sprinkled on top.
You're chatting with a robot that was mostly trained on reddit data.
Give this bullshit up. You're going through an abstraction to chat with the reddit hive mind. Why the fuck would you do this or hold any interest in it?
Chill out dude. I'm not one of those guys that has conversations with LLMs and thinks they're sentient. I just thought it was entertaining how something that people often describe as a "glorified search engine" can make such a hilariously false claim.
I mean, not even redditors would claim that Kirk is still alive.
To be fair, it is a glorified search engine in that it is also manipulated to make you think a certain way and come to a certain conclusion under the guise of impartiality.
Don't mistake my profanity for anger.
Ah. They're wrong to do so. It's trained mostly on reddit data. It effectively searches reddit. There are millions of previously existing examples of "AI" getting it wrong. Another straw for the pile doesn't change much.
We also shouldnt be trusting any ai regardless of how it was trained. Not now or in the future.
Interesting. That reply, unprompted, feels to me like a manual override. Conspiracy about shootings not happening is what led to Alex Jones owing more money than exists in the world, isn't it?
Sue Google !
(in a red state with a conservative jury)
“The Alex Jones precedent”
LLM companies are gonna have to stop using Reddit and Wikipedia as sources
Even without expanding the thinking links, I can tell when Grok has pulled something from Reddit.
They'll never stop. It dates back to 1999 when some guys manipulated Google's algorithm so that Microsoft would show up when searching "more evil than satan himself". Google's solution was to implement a ranking of trusted linkers, which created a feedback loop with pages like Wikipedia, msnm and Reddit, since those are built around linking
Reminder that Gemini is trained by using Reddit
The sources link at the bottom tried citing this CBS news article
so it’s completely misunderstanding that CBS source?
Worse than misunderstanding. It cites it as a source for things that are not even remotely mentioned in the source. So basically just lying.
LLMs are trained based on data up to a certain cutoff point. Unless they have access to the internet (and clearly show they are doing a search) then they will not have up to date information. And if they do, the only information they have will be whatever search engines tell them.
Gemini is a retard. Gemini is if ChatGPT went to only college and absorbed all the brainrot from TikTok.