From the perspective of a LLM's manifold, a factual statement and an hallucination look the exact same. It's all just training data interpolations. They cannot research a factoid the way a human would, identifying sources, gauging their reliability, cutting out obvious jokes,

Everything's fine. Gary Marcus

No, I'm not interested in developing a powerful brain. all I'm after is just a mediocre brain, something like a member of Congress.

Violence arms itself with the inventions of art and science in order to contend against violence.
-- Clausewitz

Everyone needs a break now and then.

We live in strange times indeed.

Because if it's on the internet, it must be true.

War in these modern times. finance.yahoo.com/news/asml-tsmc…

Add a few LED lights and this would look nice over my dining room table.

Follow me, lads!
— Sam Altman

Where did they hide the women?

And, were there only Caucasians available from which @mets could choose? investor.fb.com/Advisors/defau…

Confronting a friend about their bad behavior is not a condemnation of their entire humanity, but rather is an act of love, trying to help them grow.

This is true for individual relationships as well as it should be among nations.

There is indeed a considerable opportunity cost here, with so many brilliant minds chasing tiny, incremental improvements to large language models.

Talent that could be put to better use elsewhere.

I've been saying this long before Yann blocked me for his hyperbole regarding Meta's own LLM...

ft.com/content/23fab1…