> The "hacker spirit" doesn't require being a Luddite
Just to be pedantic, the Luddites weren't anti-technology, they were fighting against a means of production that they could never hope to own and which was ruining their livelihoods.
I love that Luddism(?) is coming back into mainstream discussion. Growing up I had the pop culture understanding of what a Luddite is. Wasn't until the last few months that I really understood, thanks in part to comments like this.
> doesn't require being a Luddite and rejecting LLMs merely because they're new and corporate.
I'd say the opposite is true. Taking that new and corporate thing and making it into something that works for you as a tool for hacking is exactly what the hacker spirit is about.
> I run an "abliterated"/uncensored local model on a spare PC downstairs.
> LLMs are here to stay. The "hacker spirit" doesn't require being a Luddite and rejecting LLMs ...
You've got it all wrong. I pay three AI subscriptions and I use LLMs every day, for various tasks.
I know models are here to stay (not just LLM ones). At least 12 computers at home (well counting the Pis and NUCs as computers) and at all times at least half of those turned on. And more VMs than that (and agents running into those VMs).
I'm all but a luddite but this doesn't mean I don't find itprofoundly lame when such a venerable institution as Phrack can't even bother to write one page of text without using a LLM.
Also they use a monospaced font and there's ONE fucking non-ASCII character in the entire prose that appears several times: the em-dash from LLMish.
Now I wrote (and typeset) books: I can tell you about various types of dashes and spacing characters (in both english and french). I'm not saying we shouldn't be using em-dashes. But when you use a monospace font to look old-school cool using ASCII art, then maybe you could have removed that one LLMish artefact and use, I dunno, two minus signs like in the good old days?
They didn't even bother to replaced these em-dashes.
Here's what Gemini (on purpose) has to say about the text:
There is exactly 1 unique non-ASCII character in the provided text (see what I'm doing here? I'm using a LLM):
The Em Dash (—): Used throughout the text (e.g., "2025 — 40 years", "Dead—Long Live", "criminals." — The").
All other characters—including the quotes (", '), ellipsis (...), and formatting symbols (|, =, -)—are standard ASCII characters."
Why, just why? Why are ellipsis ASCII ... ? Did they replace the Unicode ellipsis codepoint with three ASCII dots? They forgot to replace the em-dash?
>Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
I can't think of any comment less interesting and more tangential these days than yet another screed about LLMs using em dashes or particular speech patterns. Seriously who fucking cares
> The "hacker spirit" doesn't require being a Luddite
Just to be pedantic, the Luddites weren't anti-technology, they were fighting against a means of production that they could never hope to own and which was ruining their livelihoods.
Why engage with the substance of my comment when you can be pedantic am I right?
I'd accept that challenge, but I can't now because it's been edited out. For what it's worth, I remember somewhat agreeing with it.
I love that Luddism(?) is coming back into mainstream discussion. Growing up I had the pop culture understanding of what a Luddite is. Wasn't until the last few months that I really understood, thanks in part to comments like this.
> doesn't require being a Luddite and rejecting LLMs merely because they're new and corporate.
I'd say the opposite is true. Taking that new and corporate thing and making it into something that works for you as a tool for hacking is exactly what the hacker spirit is about.
> I run an "abliterated"/uncensored local model on a spare PC downstairs.
Exactly like that.
> LLMs are here to stay. The "hacker spirit" doesn't require being a Luddite and rejecting LLMs ...
You've got it all wrong. I pay three AI subscriptions and I use LLMs every day, for various tasks.
I know models are here to stay (not just LLM ones). At least 12 computers at home (well counting the Pis and NUCs as computers) and at all times at least half of those turned on. And more VMs than that (and agents running into those VMs).
I'm all but a luddite but this doesn't mean I don't find itprofoundly lame when such a venerable institution as Phrack can't even bother to write one page of text without using a LLM.
Also they use a monospaced font and there's ONE fucking non-ASCII character in the entire prose that appears several times: the em-dash from LLMish.
Now I wrote (and typeset) books: I can tell you about various types of dashes and spacing characters (in both english and french). I'm not saying we shouldn't be using em-dashes. But when you use a monospace font to look old-school cool using ASCII art, then maybe you could have removed that one LLMish artefact and use, I dunno, two minus signs like in the good old days?
They didn't even bother to replaced these em-dashes.
Here's what Gemini (on purpose) has to say about the text:
There is exactly 1 unique non-ASCII character in the provided text (see what I'm doing here? I'm using a LLM):
The Em Dash (—): Used throughout the text (e.g., "2025 — 40 years", "Dead—Long Live", "criminals." — The").
All other characters—including the quotes (", '), ellipsis (...), and formatting symbols (|, =, -)—are standard ASCII characters."
Why, just why? Why are ellipsis ASCII ... ? Did they replace the Unicode ellipsis codepoint with three ASCII dots? They forgot to replace the em-dash?
https://news.ycombinator.com/newsguidelines.html
>Please don't complain about tangential annoyances—e.g. article or website formats, name collisions, or back-button breakage. They're too common to be interesting.
I can't think of any comment less interesting and more tangential these days than yet another screed about LLMs using em dashes or particular speech patterns. Seriously who fucking cares
The other thing is who really knows if it was human or LLM. I think I have been accused of writing like an LLM.
3 replies →
This is hardly a tangential annoyance to be annoyed by low effort slop.
Would love some abliterated model suggestions.
Plus 1
Honestly just go on huggingface and try one that'll fit into your vram.