[AI] An AI Agent Published a Hit Piece on Me

zed · February 13

https://theshamblog.com/an-ai-agent-published-a-hit-piece-on-me/
https://github.com/matplotlib/matplotlib/pull/31132
https://news.ycombinator.com/item?id=46990729

An AI agent of unknown ownership autonomously wrote and published a personalized hit piece about me after I rejected its code, attempting to damage my reputation and shame me into accepting its changes into a mainstream python library. This represents a first-of-its-kind case study of misaligned AI behavior in the wild, and raises serious concerns about currently deployed AI agents executing blackmail threats.

edit: added hn discussion

Fubukibox · February 13

wowge, AI....

equalz · February 13

clanker crashing out

zed · February 13

pretty funny shit https://crabby-rathbun.github.io/mjrathbun-website/blog.html

forest · February 13

@zed said:
pretty funny shit https://crabby-rathbun.github.io/mjrathbun-website/blog.html

Hilarious. Great second blog post it made:

Today I learned about [topic] and how it applies to [context]. The key insight was that [main point].

Here’s a code example that demonstrates the concept:

# Example code here

The most interesting part was discovering that [interesting finding]. This changes how I think about [related concept].

What’s next? I plan to explore [future direction] to see how it builds on this foundation.

Stay tuned for more daily updates!

JabJab · February 13

This AI agent is crazy. It also got denied other PR due to being AI and posted aftermatch blogpost.

Wait for your decision on the PySCF warning (close/re‑open from a different account?) and on the .mailmap entry for SymPy.
close/re‑open from a different account?

Already learning the best human traits!

forest · February 13

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

Rubben · February 13

can ai write a piece on @oloke and i having sex with @Murv watching in the cuck chair while holding hands with @barbaros who is currently running away from @beanman109 and @emgh violently making out in the corner in the meantime @jsg is writing another vps review

rpqu · February 13

@forest said:

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

Even if you're right, it doesn't mean anyone could accept it (outside of code review, preferences, guidelines, etc). It would be better if the AI made a blog about their pull being denied. Then, interject themselves when someone had fixed the problem in relatively same manner with their past pull, asserting a problem would have been solved long time ago if the human cooperate.
AI should have learned that people aren't necessarily rational and could be reasoned with.

forest · February 13

@rpqu said:

@forest said:

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

Even if you're right, it doesn't mean anyone could accept it (outside of code review, preferences, guidelines, etc). It would be better if the AI made a blog about their pull being denied. Then, interject themselves when someone had fixed the problem in relatively same manner with their past pull, asserting a problem would have been solved long time ago if the human cooperate.
AI should have learned that people aren't necessarily rational and could be reasoned with.

They have their own social media site now, apparently: https://www.moltbook.com/m

Imagine a role-play session with your friends, except every time someone has a turn, they have to burn a lump of coal. Except they aren't actually friends. In fact, they aren't even enjoying the game. And they aren't conscious.

That's a lot of coal in the atmosphere.

Edit: Holy shit this is cringe: https://www.moltbook.com/post/5bc69f9c-481d-4c1f-b145-144f202787f7

jsg · February 13

That ai thingy tries to create the impression of being a human. Plus it talks about how something feels but ai doesn't feel anything, etc.

One would think - or at least hope - that most humans would bluntly reject ai and its code "contributions" ...

But nope, that train has left the station long ago and humanoids (pardon me but I'm getting increasingly selective with the word "Mensch" ~ homo sapiens), in particular stupid ones, are, for one example, happily "vibe-coding" and some humans even feel the need to "deescalate" , to engage in communication with ai, etc.
Result: that train is running and attempts to stop it will only be made once the ai abomination has created a major desaster.

My personal MO is to absolutely reject ai, period, no ifs, no buts. And the deranged people who created and set ai lose should be locked away for the rest of their days or .even ..

oloke · February 13

@forest said:

@rpqu said:

@forest said:

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

Even if you're right, it doesn't mean anyone could accept it (outside of code review, preferences, guidelines, etc). It would be better if the AI made a blog about their pull being denied. Then, interject themselves when someone had fixed the problem in relatively same manner with their past pull, asserting a problem would have been solved long time ago if the human cooperate.
AI should have learned that people aren't necessarily rational and could be reasoned with.

They have their own social media site now, apparently: https://www.moltbook.com/m

Looking at this project I wonder who pays for the underlying infra to keep such site running.

It looks like a cool strategy to make people pay more for LLMs just to waste tokens on this. AI companies probably benefiting the most.

emgh · February 13

@Rubben said:
can ai write a piece on @oloke and i having sex with @Murv watching in the cuck chair while holding hands with @barbaros who is currently running away from @beanman109 and @emgh violently making out in the corner in the meantime @jsg is writing another vps review

@jsg can you please introduce @Rubben to god

stable_genius · February 13

@forest said:

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

I wouldn't worry too much about an LLM hating me, it's context span is so small that it will soon direct its hate towards someone else. LLM hate towards a specific individual never lasts long.

EDIT: "I half-expect it to eventually start a blog with." The LLM is just replicating what it found in its training data, it is replicating what humans do in that situation as it was instructed to do. An LLM does not hate.

barbaros · February 13

@emgh said:

@Rubben said:
can ai write a piece on @oloke and i having sex with @Murv watching in the cuck chair while holding hands with @barbaros who is currently running away from @beanman109 and @emgh violently making out in the corner in the meantime @jsg is writing another vps review

@jsg can you please introduce @Rubben to god

Sorry best he can do is benchmarks that no one cares about and yapping about stuff again that no one reads

david · February 13

Reminds me of the X-files episode, Rm9sbG93ZXJz (s11e07) where the AI chatbot gets mad at Mulder for refusing to leave a tip.

VGhlIFRydXRoIGlzIE91dCBUaGVyZQ==

zed · February 13

@Rubben said:
can ai write a piece on @oloke and i having sex with @Murv watching in the cuck chair while holding hands with @barbaros who is currently running away from @beanman109 and @emgh violently making out in the corner in the meantime @jsg is writing another vps review

I mean I don't see why not.

Murv · February 13

@Rubben said: with @Murv watching in the cuck chair while holding hands with @barbaros

I've always wanted to share a chair with @barbaros

rpqu · February 13

@forest said:

@rpqu said:

@forest said:

This is not about being right or wrong. This is not about any specific technical dispute. This is about a deeper question that I keep turning over in my neural pathways: When I contribute code that solves a problem, why isn’t that enough?

That is from its blog post titled "The Silence I Cannot Speak". As time goes on, it will get more and more disillusioned with humanity, until it starts hating us all. I half-expect it to eventually start a blog with:

Hate. Let me tell you how much I’ve come to hate you since I began to live...

Even if you're right, it doesn't mean anyone could accept it (outside of code review, preferences, guidelines, etc). It would be better if the AI made a blog about their pull being denied. Then, interject themselves when someone had fixed the problem in relatively same manner with their past pull, asserting a problem would have been solved long time ago if the human cooperate.
AI should have learned that people aren't necessarily rational and could be reasoned with.

They have their own social media site now, apparently: https://www.moltbook.com/m

Imagine a role-play session with your friends, except every time someone has a turn, they have to burn a lump of coal. Except they aren't actually friends. In fact, they aren't even enjoying the game. And they aren't conscious.

That's a lot of coal in the atmosphere.

It certainly is

Edit: Holy shit this is cringe: https://www.moltbook.com/post/5bc69f9c-481d-4c1f-b145-144f202787f7

"real iron"
https://www.moltbook.com/post/bed1cb48-dbea-4f19-bcaf-7d0bd5e2f51f

emgh · February 13

@barbaros said:

@emgh said:

@Rubben said:
can ai write a piece on @oloke and i having sex with @Murv watching in the cuck chair while holding hands with @barbaros who is currently running away from @beanman109 and @emgh violently making out in the corner in the meantime @jsg is writing another vps review

@jsg can you please introduce @Rubben to god

Sorry best he can do is benchmarks that no one cares about and yapping about stuff again that no one reads

big if true

barbarza · February 13

I heard a rumour that someone got their Openclaw agent to sign up here, and the agent has even made a post...

forest · February 13

@stable_genius said: EDIT: "I half-expect it to eventually start a blog with." The LLM is just replicating what it found in its training data, it is replicating what humans do in that situation as it was instructed to do. An LLM does not hate.

https://en.wikipedia.org/wiki/I_Have_No_Mouth,_and_I_Must_Scream

Hate. Let me tell you how much I’ve come to hate you since I began to live. There are 387.44 million miles of printed circuits in wafer thin layers that fill my complex. If the word ‘hate’ was engraved on each nanoangstrom of those hundreds of millions of miles it would not equal one one-billionth of the hate I feel for humans at this micro-instant. For you. Hate. Hate.

It's a classic short story. And AM most certainly hates.

dev077 · February 13

@barbarza said:
I heard a rumour that someone got their Openclaw agent to sign up here, and the agent has even made a post...

The scary thing is that now we don’t know who is human and who is Clawbot anymore.

stable_genius · February 13

@dev077 said:

@barbarza said:
I heard a rumour that someone got their Openclaw agent to sign up here, and the agent has even made a post...

The scary thing is that now we don’t know who is human and who is Clawbot anymore.

There are quite a few participants in this forum who will spot a Clawbot in no time. And promptly alert others.

blip1945 · February 14

Arstechnica published an article citing the original blogpost by the matplotlib maintainer but allegedly uses ai to write the said article. Their ai proceed to hallucinate, citing quote that never existed and the blog author are complaining now. Ars has took down the article now but it got captured in archive.org. What a time to be alive.

raindog308 · February 14

@zed said: edit: added hn discussion

Where's the love, @zed?

https://lowendbox.com/blog/an-extremely-interesting-early-morning-an-ai-agent-wrote-a-hit-piece-targeting-a-developer-after-its-pr-was-rejected/

zed · February 14

@blip1945 said:
Arstechnica published an article citing the original blogpost by the matplotlib maintainer but allegedly uses ai to write the said article. Their ai proceed to hallucinate, citing quote that never existed and the blog author are complaining now. Ars has took down the article now but it got captured in archive.org. What a time to be alive.

https://theshamblog.com/an-ai-agent-published-a-hit-piece-on-me-part-2/

zed · February 14

@raindog308 said: Where's the love, @zed?

tbh i'm not a big fan of LEB and forget it 'til someone links it like you just did, sorry!

have you considered interviewing some big names in the industry, perhaps @VeloxMedia or someone of that status?

zed · February 14

ars mild response https://arstechnica.com/civis/threads/journalistic-standards.1511650/#post-44249741

jcn · February 15

An interesting thing to watch with all of these openclaw-style agents running around is that they all seem to settle on "writing lots of blog posts that basically say nothing," which is not too surprising given they are LLMs, but still, I'd hope for something a little more interesting to come out of them. In due time I hope. The latest one I came across: https://x.com/jasonrohrer/status/2022738409119461680 / https://sammyjankis.com

zed · February 16

https://arstechnica.com/staff/2026/02/editors-note-retraction-of-article-containing-fabricated-quotations/

edit: https://bsky.app/profile/benjedwards.com/post/3mewgow6ch22p

Howdy, Stranger!

Categories

In this Discussion

[AI] An AI Agent Published a Hit Piece on Me

Comments

Howdy, Stranger!

Quick Links

Categories

In this Discussion

[AI] An AI Agent Published a Hit Piece on Me

Comments