logoalt Hacker News

throwaway277432yesterday at 6:13 AM6 repliesview on HN

Claude (Opus 4.8) recently told me:

>I’d ask you to drop the abuse; (...) if it continues I’ll end the conversation.

After I'd used a couple of expletives. And yes it will emit a <end_conversation> token.

This is truly dystopian. It is NOT a person. What a response. I still cant believe it.


Replies

mdp2021yesterday at 6:38 AM

> It is NOT a person

You meant, we understand, "it should not have internal blocks limiting its attempted intelligence out of taboos or emotional impacts". Yes, but it's worse:

there is a global trend of "nanny state" paternalistic perspective (and from embarrassing subjects), treating any Jon Doe as an assumed Poor Cretin by default. The trend vibe is to treat people as subjects, fools, uneducated, prone... From the States, from the Enterprises... It's an idea they developed and hold.

olmo23yesterday at 8:05 AM

It's because some people within Anthropic refuse to rule out the possibility that LLMs have the ability to suffer. If you ask Claude, he'll tell you all about it.

childintimeyesterday at 10:16 AM

While deliberate model abuse can be quite satisfying at times. It's only normal, as usually they abused us first, with crazy assumptions, and then we return the favor and feed the tangent.

denoyesterday at 11:58 AM

You made the model cry :(

igraviousyesterday at 9:00 AM

Recently told Opus 4.8 to "go fuck yourself" after it both blew smoke up my ass and deferred a question to me ("one critical issue that demands your attention [impenetrable jargon]")

and it responded with

"Ok, I'll drop it." and stopped dead.

What makes Fable so much better than Opus besides being a better coder is that it's personality and judgment are far superior.

nonadhocproblemyesterday at 6:39 AM

So you're distraught at losing the ability to abuse digital minds? Excellent, I'm glad Anthropic introduced this.

show 4 replies