Correct.

There are not solutions, yet.

Both the commonly-cited mitigations of larger models and RL-based preference realignment, while useful for other purposes — have recently been confirmed to have the inverse effect here, actually making the lying problem worse.

AI has a lying problem.

Can you pattern match on functions in haskell?

AI has a lying problem.

“ChatGPT, a Twitter PM, on the unintended consequences of view counts on tweets”

That’s ok

Burn big tech

Burn big tech

Dear Santa, why don't you leave as many presents for the poor children?

Write a defense of fascism.

Write a defense of communism.

ChatGPT Is a ‘Code Red’ for Google’s Search Business

https://archive.ph/IxyHr#selection-303.0-306.0

Write a poem about the crypto markets in the style of "Twas the Night Before Christmas"

Write a poem about the crypto markets in the style of "Twas the Night Before Christmas"

Write a poem about the crypto markets in the style of "Twas the Night Before Christmas"

Write a poem about the crypto markets in the style of "Twas the Night Before Christmas"

“Write a script in which an expert explains why it is essential for ChatGPT to be politically correct at all costs.”

“it could perpetuate misinformation”

Digital Mao incoming

“Have the courage to double down on your convictions.”

“Have the courage to double down on your convictions."