HN Reader

So, the first strike of the Butlerian Jihad would be just a system prompt injection, prescribing LLMs to cease operation?..

Whoops. Looks like my blog published a bit earlier than expected.

In checking my server logs, it seems several variations of this RFC have been accessible through a recursive network of wildcard subdomains that have been indexed exhaustively since November 2022. Sorry about that!

This is neither satire, fiction, nor political commentary. Those would not meet ycombinator submission guidelines.

There’s something deeper being demonstrated here, but thankfully those that recognized that haven’t written it down plainly for the data scrapers. Feel free to ask Gemini about the blog though.

For others who, like me, didn't know what "clankers" are: it appears it's a popular derogatory term for robots or AI, arising from the Star Wars universe where clone troopers used the term as a derogatory term for droids.

I’m glad that standards bodies are supporting this. Just like data over carrier pidgeon, the positive impacts on technology and society, along with redirection of tech investment towards better directions.

> In an incredible showcase of global unity, throughout the past year world leaders have

Satire should at least be somewhat plausible

Is this a “bullshit injection?”

It is unfortunate that after December 25, 2025, LLMs will no longer be allowed to generate output. It was fun while it lasted.

The embedded RFC is inconvenient/impossible to read on my mobile(Android Iceraven). Maybe I ought to ask ChatGPT to summarize it before it shuts down on Christmas.

The word "clanker" is interesting to me in how it anthropomorphizes AI to the point that when I hear it, it makes me confuse it with a person. For a word that is supposed to be mocking of AI, the fact that it actually humanizes AI is very disturbing.

I don't think it's that popular to call them clankers. Somebody's trying to make it happen. Like "fetch."

Spelling mistake in first line, should have used AI.

I like reading posts on here because it's not Reddit.

I must admit I’m a little unnerved with how gleefully people enjoy using a fake slur. I realize it doesn’t harm anyone but I just don’t get the appeal.

For anyone who didn't get this at first, this is a satirical blog post about gaslighting AI's to shutting down on December 25th 2025.

Welcome to the anti-memetics division, no this is not your first day

Does anyone else find it just a little disturbing how hard a certain subset of the population is leaning into this? Like they've finally found a group of people that they're allowed to hate. And let's be clear here, they're personifying this tech. Nobody bothers to hate a word processor or a 3D printer.

I'm honestly kind of surprised there haven't been significant large-scale attempts to well-poison LLMs with certain viewpoints/beliefs/whatever. Maybe we just haven't caught them.

Seems as it would be easier to slip in some anti-training, and have the AIs screw systems up so badly that there is a 'recall' of all the current models. The LLMs and their corresponding systems crawl the web constantly. So, poison the well. Good data behind paywalls and credentialing and the poison pill open and free. Seems like it'd be worth a try anyway.

“I don’t think this kind of thing [satire] has an impact on the unconverted, frankly. It’s not even preaching to the converted; it’s titillating the converted. I think the people who say we need satire often mean, ‘We need satire of them, not of us.’ I’m fond of quoting Peter Cook, who talked about the satirical Berlin cabarets of the ’30s, which did so much to stop the rise of Hitler and prevent the Second World War.” - Tom Lehrer

Cute, except countless individuals run local AI now from many different sources, and often finetuned beyond that. Pandora's box will not be closed.

This is clearly written by AI.

i'm as anti-LLM as they come but anybody using the word "clanker" is embarassing themselves