he Most Insidious Trick Of AI Language Models

Epoch Times ^ | 08/14/2025 | Jeffrey Tucker

[mmichaels1970]

But will it be “good enough”?

For lazy creators, maybe. For content consumers, likely not. To me, it’s kind of like if you had a Vulcan around to help you out. The logical stuff would be awesome. The emotional stuff would be a fake, contrived, shallow shell that could be spotted pretty easily.

[cgbg]

What percent of humans are truly creative?

Most just spew what they have been taught or read somewhere—and start name calling when seriously challenged.

[mmichaels1970]

If AI wants to emulate humans it needs to remember to add the occasional insult—call the user “stupid” and a “moron” at least once every fifty words.

I angrily posted something in ALL CAPS once (i think it was in copilot), and it told me it “wasn’t interested in continuing this conversation any further”. True story.

[cgbg]

That is hilarious.

I don’t use AI but if I did I would go straight into Don Rickles mode and go after it with an endless string on insults and ad hominem attacks—just to see the reaction.