To be fair, the AI models' scrambled text basis was also stolen wholesale from lots and lots of real people who were not reimbursed and are about to be driven out of their jobs by this raid.
No lie, I had to explain to my 50yo manager that Chat isn't a search engine and the information she could get from it isn't likely to be truthful. Thankfully she believed me (I think) but it's a plague.
When you do a Google search, you can use your critical thinking to verify the sources you get your information from, compare and contrast points, and reach an educated conclusion using your own brain. ChatGPT is a black box that you are unable to verify the veracity of
Usually info you pull up on Google is tied to its source, both a website and a writer or video presenter, who can be screened and held to a standard. It's not an algorithm trying to guess what words you want to see and spitting them out without a source or any indicator of how it got there.
You misunderstood.
You're a nerd because you think I'm gonna seriously argue with you on twitter, lol.
I gave my opinion, move on, I'm not terminally online. Unlike some people.
I feel like that'd be a great way to break your game's economy tbh. Not that most D&D games are particularly affected by a broken economy, but I can't imagine the training data for any LLM includes an internally consistent pricing scheme with all the homebrew stuff on the internet.
Yup, I tested it out with something that I'm a subject matter expert in, and it was both very confidently wrong about what it was trying to tell me and folded immediately when I confronted it.
This is what ultimately cured me of any illusions when these things were first gaining steam. Not only can you ask it the same question and get different answers with no continuity, but if you tell it that it got something wrong, it'll just spit out something else. Even if it was right!
Same. Chat is fantastic for a multitude of things, including but not limited to finding holes in plans or seeing potential possibilities in presented scenarios.
It makes up "information". Translation: it create FAKE NEWS.
And on top of that this shit consume huge ammounts of water and AIs as a whole already put more C in the atmosphere than all of Earth's civil aviation combined.
So you are getting dumber while fucking up the planet everytime you use it.
it's a predictive text model, the only data it collects is whatever is relevant to the question "what is most likely to appear after this token?" a token being something like "gran" "mou" or "sed" it's not actually particularly reliable
it's a touchy topic, I could chastise the people here for being rude to you all I want, but ultimately it's up to you to tolerate it (or not, ima be totally honest) and see what they're saying
it uses a neural network to create something that sounds like what a person would say. every word is just a number with a probability value attached to it which it uses to calculate a plausible sentence. it is a grammatically competent word predictor, you have that in your autocorrect
If only an expert who already knows a significant amount a topic can tell when ChatGPT is hallucinating multiple incorrect things about said topic, who truly benefits?
The expert doesn't need it. The more junior person or non-expert won't be able to tell what parts of its answer are entirely wrong without googling and double checking the answer.
I don't know how much you've used it lately, but Google has sucked for a while now. Bing, Ecosia, and Duck Duck Go give better results and Google's A.I. summary is often wrong.
Hard to blame people for going with the new, but we do need to move away from Google.
Not sure why you keep responding as if Im trying to prove something, I posted an opinion agreeing with the OP. None of this has to do with my initial comment
Nearly screamed at one of my old jobs where the manager of my department said he used AI for risk assessments when pressed for time. ‘But it’s okay, because I copy the stuff it’s given me into Word, correct anything blatantly wrong and add where needed.’
My mom is always talking about how she uses ChatGPT to help her with everything and continues to ignore me expressing my unease with the damn bot. I don't know what to do.
Comments
There went most of my respect right there.
better spend hours looking for info in google that AI can instantly find for me
If a topic requires hours of investigating maybe AI cant accurately do that dont you think?
Maybe you should try to get information that isnt inmediately seen or obvious?
Maybe you wanna know your sources and their context or credibility?
You're a nerd because you think I'm gonna seriously argue with you on twitter, lol.
I gave my opinion, move on, I'm not terminally online. Unlike some people.
YOU COULD'VE GOOGLED THAT AND THE FIRST RESULT WOULDVE BEEN MORE EXTENSIVE
You fucking moron
You won't get novel information out of chatGPT. It can synthesize common ideas. It's not good at making entirely new constructs.
Terrible at working with niche information, especially without the context that is only present in the minds of discussion participants.
And on top of that this shit consume huge ammounts of water and AIs as a whole already put more C in the atmosphere than all of Earth's civil aviation combined.
So you are getting dumber while fucking up the planet everytime you use it.
i don't even finish reading their post
if i cared what the plagiarism machine had to say I'd ask it myself
Hard to blame people for going with the new, but we do need to move away from Google.
Most credible sources are burried with the highest bidder getting the first suggestions. Most people aren't going to bother.
Still not sure what it means.