this post was submitted on 25 Dec 2024
262 points (75.1% liked)
Linux
48655 readers
2075 users here now
From Wikipedia, the free encyclopedia
Linux is a family of open source Unix-like operating systems based on the Linux kernel, an operating system kernel first released on September 17, 1991 by Linus Torvalds. Linux is typically packaged in a Linux distribution (or distro for short).
Distributions include the Linux kernel and supporting system software and libraries, many of which are provided by the GNU Project. Many Linux distributions use the word "Linux" in their name, but the Free Software Foundation uses the name GNU/Linux to emphasize the importance of GNU software, causing some controversy.
Rules
- Posts must be relevant to operating systems running the Linux kernel. GNU/Linux or otherwise.
- No misinformation
- No NSFW content
- No hate speech, bigotry, etc
Related Communities
Community icon by Alpár-Etele Méder, licensed under CC BY 3.0
founded 5 years ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Sorry for the wall of text again c:
(CLICK HERE FOR BIG WALL)
AI text as a whole is usually structured, neutral-positive to positive shallowness. It's called slop because it's easy to make a lot of substanceless, nutrientless goo. One common structure is
What do we spot? Sets of three, largely perfect/riskless formal grammar (grammar perfection is not inhuman -- but a human might, say, take the informal risk of using lotsa parentheses (me...)), uncreative colon titles, SEO-style intros and conclusions, an odd corporate-style ethics hangup, em-dashes (the long —), and some of the stuff in that reddit link I mentioned are often giveaways.
Here's some examples in the wild:
Playing Dumb: How Arthur Schopenhauer Explains the Benefits of Feigned Ignorance. PeopleAndMedia. has useless headings and the colon structure I mentioned. There's also phrases like "Let's delve" and "unexpected advantage" -- ChatGPT likes pretending to be unconventional and has specific diction tics like "Here's to a bright future!" One interesting thing is that the article uses some block quotes and links -- this is rare for AI.
Why is PHP Used. robots.net. This is from a "slop site", one that is being overrun by AI articles. Don't read the whole thing, it's too long. Skim first. See how many paragraphs start with words like "additionally", "moreover", "furthermore", like a grade school English lit student? Furthermore (lol), look at the reasonings used:
ChatGPT-esque vocabulary is used (this is something you unfortunately get a feel for), and the reasoning isn't very committal. Instead of evaluating some specific event deeper, the article just lists technologies and says stuff like "PHP has comprehensive and well-maintained documentation, providing in-depth explanations, examples, and guides." So what if there's docs? Everyone has documentation. Name something PHP docs do better or worse. Look at this paragraph (SKIM IT, don't read deeply):
It doesn't actually SAY ANYTHING despite its length. The paragraph can be compressed to: "CodeIgniter has a light footprint". It doesn't even say whether we're talking about comparative speed, memory usage, or startup time. It's like they paid someone (openAI) to pad word count on the ensmallening I mentioned.
Before reading something, check the date. If it's after 2020, skims to be too long and not very deep, and has too many GPT tics (tricolons, vocab like "tapestry/delve", the SEO shit structure), then it's AI slop. Some readers actively avoid post-2020 articles but I can't relate.
edit: clarified that perfect grammar is humanly doable, but GPT-style riskless formal grammar is still distinct from grammatical human text
the term "riskless grammar" perfectly puts into words how i felt about chatgpt's texts, every human-written text has something "wrong" with it grammar-wise, except maybe example essays by english teachers.
As an example, my previous paragraph has a lowercase I, too many commas, sentences compressed by using hyphens where they probably shouldnt go and probably some other stuff i missed.
But it still read well, at least i hope.
Most authors write their sentences their own way, and in my opinion, that's what makes reading their books interesting. Perfect grammar is boring and no fun to read.
as a fun experiment, i asked chatgpt to rewrite my first paragraph:
"The phrase "riskless grammar" accurately captures my impression of ChatGPT's texts. Unlike human-written content, which often contains grammatical imperfections—except perhaps for example essays by English teachers—ChatGPT's writing maintains a level of precision and correctness."
Kind of changed the meaning to be self-complimenting, which is funny.
edit: Normally I would have rewritten parts of this comment to make my point more clearly and be better to read, but i wanted to keep my first draft to make my point a bit better.
Ty for feedback :>
Your paragraph read well. I definitely agree -- grammar with risks, outside of hyper-formal sitches, is just stylized diction. ChatGPT could scarcely come up with an e.e. cummings poem (just tested now, it never gets the style about right), nor dare to abuse parentheses, nor remove cruft for conciseness (e.g. to start a sentence with "Kind of changed" instead of "This kind of changes" for compression (woot)). It's a "wrong" but not quite "wrong", and I'm glad that "riskless" manages to carry that feeling
And I edit a lot too :) it's the "post-email-send clarity" effect
Errors can give away that a human typed something, but knowing proper grammar, spelling, and syntax of English is totally neutral—if not to be somewhat expected from a native speaker/typer with a lifetime to learn the language they speak (especially if we consider how many Anglophones are monolingual + educated + have access to technology like spell check meaning there is little excuse for not having English mastery).
In my education, I got a public apology from a teacher letting the class know they tried to dig up proof of plagiarism in my persuasive papers, but for the first time proved themself incorrect on a plagiarism hunch. Humans are capable of writing well.
edit: updated accordingly for clarity
Ah, I mean proper grammar as in formal, largely riskless grammar. For example, AI wouldn't connect
with pluses, like a human would.
Not sure how I'd phrase that though. Maybe "perfect, risklessly formal grammar" as I just tried to call it? (i.e. if AI trainers consider using +'es a "risk", as opposed to staying formal and spick n' span clean).
Perfect grammar is humanly possible but there is some scrutiny that can be applied to GPT-style grammar, especially in the context of the casually-toned web (where 100%ed grammar isn't strictly necessary!).