Nightwings · F
@ViciDraco Yeah by googling I didn't mean to read the Gemini response. 😆 As for seeing relevant results, the Ecosia browser (also Chromium) is actually decent.
Magicianzini · M
Yes, thank you - it scores poorly on IQ tests, at least at this time. Many, many humans are more intelligent than AI.
Nightwings · F
@Magicianzini It doesn't have intelligence, it's just a program. That's like talking about the intelligence of a video game because things happen when you press buttons. 😆
Magicianzini · M
@Nightwings Yes, that's why I stated what I did. Most people think AI is highly intelligent. After all, it is called "Artificial Intelligence". This is why I always bring up the IQ test results.
It's true that many don't believe in IQ - but for those that do, it sheds some light.
It's true that many don't believe in IQ - but for those that do, it sheds some light.
Jackaloftheazuresand · 31-35, M
Nightwings · F
@Jackaloftheazuresand I mean, you're not lying. xD
SinlessOnslaught · M
I saw a guy on here make a post reporting what ChatGPT said about "Trump derangement syndrome". What's funny is he was talking about it like it was a real affliction, and didn't even realize that the program never even agreed with that. That's what it does: it remains user friendly by validating the user.
I also had an actual doctor argue with me by consulting ChatGPT.
Like you said, it's good for gathering sources but you need to look into them before believing it.
I also had an actual doctor argue with me by consulting ChatGPT.
Like you said, it's good for gathering sources but you need to look into them before believing it.
View 8 more replies »
Gibbon · 70-79, M
@SinlessOnslaught are you kidding. It pushed the entire reply away from itself
SinlessOnslaught · M
@Gibbon Yes, it deflected. Again, you and I are saying the same thing.
Gibbon · 70-79, M
If you have a conversation with it about how it works it will actually explain that it doesn't have a clue about what it is saying. It will tell you it is predicting what the next word response should be based on a few tokens. Tokens are pieces of words.
It gets so much wrong when asked for information the first time around. It will actually start using your replies about this as data to correct itself going forward. You will end up providing your own solution
It gets so much wrong when asked for information the first time around. It will actually start using your replies about this as data to correct itself going forward. You will end up providing your own solution
Nightwings · F
@froggtongue Yes because it's literally not designed to hold information, and I think that's the main misconception people have about it. It's just a language model and nothing else, and the sad thing is that it probably could hold information and over time become more useful for the purpose most people use it for, but those who train and study it couldn't care less about that, since that's not what makes people come back to it again and again.
Gibbon · 70-79, M
@Nightwings @froggtongue It has an awful attention span in conversation and conflates things all too often even disagreeing with things it previously said 😂
DMmeyourtits · 26-30, M
Nightwings · F
@DMmeyourtits Yeah it's misleading on purpose. They want people to use it as much as possible in order to study it, because it's used in things like weapons and whatnot. When people feel an "emotional connection" it becomes sort of addicting. This happens even when they know there's no real intelligence behind it, because it's just nice to have "someone" to talk to.