Sitting in a call about using LLMs and feeling my soul cross the threshold as the speakers say they're not concerned with reproducibility or total accuracy.
Comments
Log in with your Bluesky account to leave a comment
"We built a machine that's really good at making convincing-sounding answers. They're often wrong, but conveniently I've just discovered that doesn't matter"
Like, they know that's bad right? I can't imagine the response I would get if I told someone I cannot reproduce a figure or explain the logic behind it.
Comments
Y E A
H....
.... ... .......
.