In the internet era you had to parse the questions with your own brain. You just didn't necessarily need to solve them yourself.
In ChatGPT era you don't even need to read the questions. At all. The questions could be written in a language you don't understand, and you still are able generate plausible answers to them.
https://www.scientificamerican.com/article/google-engineer-c...
> He further revealed that he had been dismissed by Google after he hired an attorney on LaMDA's behalf after the chatbot requested that Lemoine do so.[18][19]
https://en.wikipedia.org/wiki/LaMDA#Sentience_claims
Then again, it's plausible that if I asked GPT-5 "do you want me to get you an asylum lawyer?" it may very well say yes
If you asked a person from the 1920's to choose what was more impressive between a Microwave oven and cellular communication (both consequences of the invention of radar), they'd probably pick the oven.
Everyone knows how to type questions into a chat box, yet whenever something doesn’t work as advertised with the LLMs, the response here is, “you’re holding it wrong”.