"I cannot speak for your experience because I am not a woman, but I experience various forms of discrimination and sexism... directed at me because I'm a man..."
This guy is going to be single before I'm done drinking my chocolate.
Guy is talking with his girlfriend. I caught "OK, but except for the parts that were sexist, what about the rest?" and later him asking over and over "Do you think I am sexist?"
This date is going great!
Campaign slogans should all strike this tone:
Instead of "Yes We Can" use "We'll give it a shot"
Instead of "I'm With Her" use "I think she's with my friend..."
Instead of "Make America Great Again" use "We've been better..."
Instead of "Build Back Better" use "Just like use duct tape or something..."
Instead of "Feel The Bern" use "I'm kinda itchy."
The numbers from this story are based on a laughable "fake paper detector" that literally consists of the following ONLY. Do the authors:
1) use private (non-institutional) email addresses and/or have a hospital affiliation,
and
2) have no international coauthors.
That's it.
If these criteria are met, the paper is deemed a "potential red-flag fake publication" and counted toward that 30% tally.
I think using LLMs for communication is bad.
Mostly that's because what you are trying to communicate is contained in the prompt you fed to the model. You're making the recipient effectively reverse-engineer your prompt from the output.
Just send me the prompt. I could always feed it into an LLM myself if I wanted to. But most likely, I won't have to because the prompt tells me what you are trying to say.
Carl T. Bergstrom
Greg Egan