Some users appreciate AI writing for helping non-native speakers and raising quality standards, while many others criticize its unnatural style, detectable patterns, and inauthenticity when passed off as human.
Based on 16 visible X reactions from 54 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Hm I don't know about this. AI does not write how most humans naturally write. It has a distinctive, cringe-worthy style of its own, especially the Claude models. I'd actually say this is one thing where talented humans genuinely have the upper hand (for now). It's been interesting how hard it's been to fix in the models; the slop is still so apparent even as more Erdos problems fall.
@Dan_Jeffries1 It feels like watching an alien nonchalantly drink gasoline from a beer bottle. And it's oddly impersonal, like they're turning to the camera and monologuing through the fourth wall. https://twitter.com/rameswar08/status/2076957355900821734
@signulll ai writing raised the floor so high that even ceos can't hide bad prose anymore the winner will be the one who uses ai to amplify their unique style, not replace it
@Dan_Jeffries1 "command the language with the same thundering power" wtf, and this cuck want to lecture us on writing?
Steven Sinofsky compares the AI output quirks to early word-processing era documents.
Hm I don't know about this. AI does not write how most humans naturally write. It has a distinctive, cringe-worthy style of its own, especially the Claude models. I'd actually say this is one thing where talented humans genuinely have the upper hand (for now). It's been interesting how hard it's been to fix in the models; the slop is still so apparent even as more Erdos problems fall.
@Dan_Jeffries1 It feels like watching an alien nonchalantly drink gasoline from a beer bottle. And it's oddly impersonal, like they're turning to the camera and monologuing through the fourth wall. https://twitter.com/rameswar08/status/2076957355900821734
@signulll ai writing raised the floor so high that even ceos can't hide bad prose anymore the winner will be the one who uses ai to amplify their unique style, not replace it
@Dan_Jeffries1 "command the language with the same thundering power" wtf, and this cuck want to lecture us on writing?
@signulll lol honestly no, it’s like a public toilet. Hold my breath, get in, post, gtfo
Spending time on fanfic forums was a good education. There are some highly skilled writers there, I support some on patreon myself, and some have been good enough to get traditionally published. In the past I've read things there that I've enjoyed far more than anything on the NYT top ten. But reading those forums really drove home how bad the average person is at structure, character, dialogue, and even understanding what a story is at all. The models can't fix that part for people yet.
Endorse. 💯 No one would believe that people used to think "did you use a word processor to create this?" When documents first started having formatting—the first indication was a prolific use of hyphenated words because that was the first thing solved automagically in word processors—people would viscerally devalue the material. Early word processing docs (of the Mac era) had *too* much formatting and ImageWriter output looked toy like. The lack of typos and spelling seemed a bit mysterious. The software and hardware had to improve. The LaserWriter was the tipping function. Then documents looked like something created a printer. Amazing as it sounds people would ask "how did you do that?" in 1985. There were many things in word processing that needed to improve before they could fully replace old processes. Many were just how people used the tools (less formatting and ransom notes). Many were missing features like tables which used to be hand drawn and literally pasted into empty space. And many were just bugs like how subscript/superscript worked and footnotes. All along people could "tell" you used the new thing and had emotional reactions ranging from awe to disdain. If you made a mistake with a word processor believe me everyone pointed it out. A single typo would make you a laughing stock for low quality work. When Word introduced AutoCorrect we took out full page adds with giant typography "teh -> the" it was so important. AI is going through all this right now. The bugs will get fixed. New ways of acceptable writing will emerge. People will "catch" others using these new tools and scream busted. Embarrassing mistakes will happen. But "It's happening" is definitely true.
LLMs have most certainly gotten better at writing well and I love them but as a thirty years veteran writer I can still spot the patterns. They have a remarkably consistent structure and sentence style (block paragraphs, this and then this and this sentence patterns, lack of contractions, overly formal word choice, grammar that's too perfect instead of a style that knowingly breaks the rules to hammer home a point, etc). The other challenges are subtler. There is often no unifying idea or, worse, stray ideas find their way in that don't make sense in the overall structure and don't make the overall points cleanly or that don't flow smoothly from the last idea. To me every X post reads the same now. Same overly long read, puffy middle and semi-conversational "here's the thing" style. No doubt it's gotten better and if you know how to use it (like anything in life) it's an amazing tool but I find that when I start with LLM drafts I have to work harder sometimes, almost like being an editor of someone else's writing (someone who is slightly and subtly nuts) instead of it being my own words end to end and where the whole shape of the article or story and what it's doing is in my mind. But this is the worst it will ever be. Still some of the problems will persist. Ultimately if you don't do the work or part of the work to shape the LLM's writing (or code or art or whatever) then it's not yours and you don't understand it and can't command the language with the same thundering power.
Some users appreciate AI writing for helping non-native speakers and raising quality standards, while many others criticize its unnatural style, detectable patterns, and inauthenticity when passed off as human.
Based on 16 visible X reactions from 54 accounts; directional sample.
Ask a question below.
Published answers will appear here.
Spending time on fanfic forums was a good education. There are some highly skilled writers there, I support some on patreon myself, and some have been good enough to get traditionally published. In the past I've read things there that I've enjoyed far more than anything on the NYT top ten. But reading those forums really drove home how bad the average person is at structure, character, dialogue, and even understanding what a story is at all. The models can't fix that part for people yet.
Endorse. 💯 No one would believe that people used to think "did you use a word processor to create this?" When documents first started having formatting—the first indication was a prolific use of hyphenated words because that was the first thing solved automagically in word processors—people would viscerally devalue the material. Early word processing docs (of the Mac era) had *too* much formatting and ImageWriter output looked toy like. The lack of typos and spelling seemed a bit mysterious. The software and hardware had to improve. The LaserWriter was the tipping function. Then documents looked like something created a printer. Amazing as it sounds people would ask "how did you do that?" in 1985. There were many things in word processing that needed to improve before they could fully replace old processes. Many were just how people used the tools (less formatting and ransom notes). Many were missing features like tables which used to be hand drawn and literally pasted into empty space. And many were just bugs like how subscript/superscript worked and footnotes. All along people could "tell" you used the new thing and had emotional reactions ranging from awe to disdain. If you made a mistake with a word processor believe me everyone pointed it out. A single typo would make you a laughing stock for low quality work. When Word introduced AutoCorrect we took out full page adds with giant typography "teh -> the" it was so important. AI is going through all this right now. The bugs will get fixed. New ways of acceptable writing will emerge. People will "catch" others using these new tools and scream busted. Embarrassing mistakes will happen. But "It's happening" is definitely true.
LLMs have most certainly gotten better at writing well and I love them but as a thirty years veteran writer I can still spot the patterns. They have a remarkably consistent structure and sentence style (block paragraphs, this and then this and this sentence patterns, lack of contractions, overly formal word choice, grammar that's too perfect instead of a style that knowingly breaks the rules to hammer home a point, etc). The other challenges are subtler. There is often no unifying idea or, worse, stray ideas find their way in that don't make sense in the overall structure and don't make the overall points cleanly or that don't flow smoothly from the last idea. To me every X post reads the same now. Same overly long read, puffy middle and semi-conversational "here's the thing" style. No doubt it's gotten better and if you know how to use it (like anything in life) it's an amazing tool but I find that when I start with LLM drafts I have to work harder sometimes, almost like being an editor of someone else's writing (someone who is slightly and subtly nuts) instead of it being my own words end to end and where the whole shape of the article or story and what it's doing is in my mind. But this is the worst it will ever be. Still some of the problems will persist. Ultimately if you don't do the work or part of the work to shape the LLM's writing (or code or art or whatever) then it's not yours and you don't understand it and can't command the language with the same thundering power.