I was already sure of the result, it’s to show that’s so unbelievable that something trained on millions of stolen past articles and news would dismiss it immediately without bothering an extra web search
Once it got the source of “truth” from the web search (they scrape the first few results, meaning some black hat with SEO can let them believe anything), it’s impossibile to let it change opinion
But the first answer is almost always a lie, because it costs money to actually search and scrape so they penny pinch and give a false but convincing information
Even the bots can’t believe it
Your AI slop bores me.
Why bother going to LLMs when you could’ve just asked me?
I was already sure of the result, it’s to show that’s so unbelievable that something trained on millions of stolen past articles and news would dismiss it immediately without bothering an extra web search
Say that it’s wrong once more and see if it flips the answer again.
Once it got the source of “truth” from the web search (they scrape the first few results, meaning some black hat with SEO can let them believe anything), it’s impossibile to let it change opinion
But the first answer is almost always a lie, because it costs money to actually search and scrape so they penny pinch and give a false but convincing information