I’m with you on that one, I mean it really wouldn’t be that hard to just.. do the math on this, especially if it’s your job to write articles like this.
Maybe the phrase “we asked chatGPT” is the new hotness in clickbait but fuck me if it isn’t depressing.
ChatGPT was cited as source. That’s not terrible. To not double-check those numbers is not great, but we aren’t doomed because it was cited. We would be doomed if this became the norm.
Big brain move, asking an LLM to do economics math. They’re large LANGUAGE models, not large MATHEMATICAL models. They forgo rigorous methodologies in favor of stochastically predicting correlations based on intractably large datasets, which can lead to such impressive mathematical feats as failing to count to three. Great job, Yahoo.
Seriously. My sister was the one who believed ChatGPT can think and do calculus or math. I gave a rather simple but definitely unique math task. At first I put it into google search calculator and showed the result, which was correct. Then I gave the same exact thing to ChatGPT and the result was completely different and false.
My favorite example was when they gave all the top LLMs a task to figure out how many circles in an image were touching. Every single one failed miserably at this task that would have been trivial for a 3 year old.
Unless the answer was five. Because of the Olympics logo.
Edit: I posted this because I would imagine LLMs to have readily accessible figures that random reporters may not.
LLMs have the same information that a reporter can find online or other public records, but will more often than not conflate separate things into new, inaccurate info or straight up make shit up that sounds accurate to someone who doesn't know anything about the subject.
Kronusdark
in reply to Maeve • • •So, we are citing ChatGPT for news now?
And not even trying to hide the fact?
Jesus. We’re cooked.
like this
ElcaineVolta likes this.
CLOTHESPlN
in reply to Kronusdark • • •like this
Maeve likes this.
Keilik
in reply to Kronusdark • • •I’m with you on that one, I mean it really wouldn’t be that hard to just.. do the math on this, especially if it’s your job to write articles like this.
Maybe the phrase “we asked chatGPT” is the new hotness in clickbait but fuck me if it isn’t depressing.
Kronusdark
in reply to Keilik • • •We asked chatGPT for the top ten fashion trends of 2025. You won’t believe number 6.
🙃
crandlecan
in reply to Kronusdark • • •aarch0x40
in reply to Keilik • • •santa
in reply to Kronusdark • • •Kronusdark
in reply to santa • • •If I can’t cite Wikipedia as a source in a school paper, journalists shouldn’t be allowed to cite ChatGPT for an article.
Research is part of their job.
EvilBit
in reply to Maeve • • •REDACTED
in reply to EvilBit • • •EvilBit
in reply to REDACTED • • •My favorite example was when they gave all the top LLMs a task to figure out how many circles in an image were touching. Every single one failed miserably at this task that would have been trivial for a 3 year old.
Unless the answer was five. Because of the Olympics logo.
Pfeffy
in reply to Maeve • • •🇰 🌀 🇱 🇦 🇳 🇦 🇰 🇮
in reply to Maeve • • •LLMs have the same information that a reporter can find online or other public records, but will more often than not conflate separate things into new, inaccurate info or straight up make shit up that sounds accurate to someone who doesn't know anything about the subject.
like this
fonix232 likes this.