23
- My Forums
- Tiger Rant
- LSU Score Board
- LSU Recruiting
- SEC Rant
- SEC Score Board
- Saints Talk
- Pelicans Talk
- More Sports Board
- Fantasy Sports
- Golf Board
- Soccer Board
- O-T Lounge
- Tech Board
- Home/Garden Board
- Outdoor Board
- Health/Fitness Board
- Movie/TV Board
- Book Board
- Music Board
- Political Talk
- Money Talk
- Fark Board
- Gaming Board
- Travel Board
- Food/Drink Board
- Ticket Exchange
- TD Help Board
Customize My Forums- View All Forums
- Show Left Links
- Topic Sort Options
- Trending Topics
- Recent Topics
- Active Topics
Started By
Message
AI - it’s quite amazing how wrong it is on many questions
Posted on 9/26/26 at 5:31 pm
Posted on 9/26/26 at 5:31 pm
I do ChatGPT, Grok, Gemini and Claude, and pay a monthly subscription for ChatGPT.
I’m amazed how wrong all of them are on answers to many non-complicated questions. Usually, I’ll correct it or simply say you are wrong, and it will re-answer the question, many times picking up on why I say it is wrong, then it gives the correct answer.
Truly amazing, frustrating and kind of concerning.
I’m amazed how wrong all of them are on answers to many non-complicated questions. Usually, I’ll correct it or simply say you are wrong, and it will re-answer the question, many times picking up on why I say it is wrong, then it gives the correct answer.
Truly amazing, frustrating and kind of concerning.
Posted on 9/26/26 at 5:32 pm to Skippy1013
Great. Now it knows which side you’re on.
Posted on 9/26/26 at 5:33 pm to Skippy1013
yeah i've noticed that as well
Posted on 9/26/26 at 5:40 pm to Skippy1013
Even more fun is the fact that new llm's are increasingly training on llm generated "facts" which will eventually replace actual facts thus changing history and our understanding of what reality even is.
And since we now know that's possible it poses the question of how can we know it didn't already happen or if it's even possible to know that.
What a time to be a sim.
And since we now know that's possible it poses the question of how can we know it didn't already happen or if it's even possible to know that.
What a time to be a sim.
This post was edited on 9/26/26 at 5:49 pm
Posted on 9/26/26 at 5:41 pm to Skippy1013
quote:
it’s quite amazing how wrong it is on many questions
It is.
However, it's also quite amazing at how much better it has gotten over the last three years at answering questions.
Posted on 9/26/26 at 5:41 pm to Skippy1013
I’ve noticed it struggles with unique questions. If it’s been asked something a thousand times it has a polished answer.
Posted on 9/26/26 at 5:45 pm to Skippy1013
Claude is by far the best
Posted on 9/26/26 at 5:49 pm to Skippy1013
It’s been pretty spot on for me for finding car parts for my antiques
Posted on 9/26/26 at 5:52 pm to Skippy1013
Damn Skippy
Posted on 9/26/26 at 5:57 pm to Skippy1013
I use it for two reasons, research.. finding raw data, articles etc. - not it's interpretation of that data
Generate code and cad files. Sometimes I just ask for the code to generate the file, but if I have credits to burn it will give a complete package and create a library of files that is just convenient.
AI isn't interpreting your data, they are finding people who interpreted the data and are parroting what those people say, regardless of the legitimacy of that source.
Generate code and cad files. Sometimes I just ask for the code to generate the file, but if I have credits to burn it will give a complete package and create a library of files that is just convenient.
AI isn't interpreting your data, they are finding people who interpreted the data and are parroting what those people say, regardless of the legitimacy of that source.
Posted on 9/26/26 at 5:57 pm to Skippy1013
Garbage in, garbage out.
Posted on 9/26/26 at 6:06 pm to Skippy1013
It's great on walking me through technical issues. It tells me step by step what to do and if that doesn't work it tells me step by step what to try next. Saves a ton of time vs googling it all myself.
Posted on 9/26/26 at 6:06 pm to Skippy1013
You have to realize that it's pulling information from the internet.
Are you aware of how much disinformation is on the internet?
Are you aware of how much disinformation is on the internet?
Posted on 9/26/26 at 6:22 pm to Skippy1013
These LLMs aren't answering questions the way you think they are. They're taking your input (the group of words you typed) and are selecting the most statistically likely words that correlate to the words you submitted. It's not thinking, it's calculating a math problem.
Posted on 9/26/26 at 6:24 pm to Skippy1013
I just asked Gemini today when will it stop hallucinating and it said that by 2027/2028 hallucinations will be minimal. Not sure if I believe that. 
Posted on 9/26/26 at 6:31 pm to Skippy1013
It’s almost like it’s a computer program that pulls its knowledge from what it reads online.
Posted on 9/26/26 at 6:40 pm to Skippy1013
quote:
I’m amazed how wrong all of them are on answers to many non-complicated questions.
Congrats, you are using them wrong
Posted on 9/26/26 at 7:04 pm to Skippy1013
Posted on 9/26/26 at 7:05 pm to Skippy1013
don't be amazed, it's just spitting out what it's programmed to spit out.
Posted on 9/26/26 at 7:20 pm to Nathan Hail
it is very polite when it "admits" it's wrong, though.
Page 1 of 2
Popular
Back to top
Follow TigerDroppings for LSU Football News