A potential disadvantage I see with using LLMs for writing long form content is the existence of readers' ability to smell out slop. It's real. Over a long enough timeframe, anyone can develop it.
I don't subscribe to the 'masses are dumb' worldview as it is not effective while trying to provide value to people.
Once they do develop this ability, it's likely that they'll start trusting the text you publish less and less, because they might start doubting that you've proofread it, verified it or put care and effort into refining it.
There is no such thing as an objectively 'good' body of text to read. But people will still judge it according to personal tastes and preferences.
If what they're expecting from a writer is just news reporting, slop might be fine to an extent. But if they want more, like valuable insights, analyses or opinions, slop could be off-putting.
Even for news, telling LLMs to answer basic questions required in journalism: 'why, when, where, who, how, what, which, etc.', i.e. research, which LLMs can be better at than human effort, is highly effective.
But choosing which answers are the most relevant for a given news story, how to interpret and present them, the appropriate theoretical tools to use, i.e. curation, empathy, reasoning (in the accurate definition*), judgement and entrepreneurship, are things humans are better at than LLMs. This requires time and care too.
*For now, I can only say bluntly that LLMs aren't really 'reasoning', based on my understanding of rationalist ideas.
sachin
npub1xn…43qnl
2026-08-18 23:43:34 UTC
Author Public Key
npub1xnc64f432zx7pw4n7zrvf02mh4a4p7zej3gude52e92leqmw8ntqd43qnlSeen on
wss://relay.wisp.talkPublished at
2026-08-18 23:43:34 UTCEvent JSON
{
"id": "1ce71f9781066730e88ed89ceb5a4d12155b75eb5df2b390471ee98823534398",
"pubkey": "34f1aaa6b1508de0bab3f086c4bd5bbd7b50f8599451c6e68ac955fc836e3cd6",
"created_at": 1787096614,
"kind": 1,
"tags": [
[
"client",
"Amethyst"
]
],
"content": "A potential disadvantage I see with using LLMs for writing long form content is the existence of readers' ability to smell out slop. It's real. Over a long enough timeframe, anyone can develop it. \n\nI don't subscribe to the 'masses are dumb' worldview as it is not effective while trying to provide value to people.\n\nOnce they do develop this ability, it's likely that they'll start trusting the text you publish less and less, because they might start doubting that you've proofread it, verified it or put care and effort into refining it.\n\nThere is no such thing as an objectively 'good' body of text to read. But people will still judge it according to personal tastes and preferences.\n\nIf what they're expecting from a writer is just news reporting, slop might be fine to an extent. But if they want more, like valuable insights, analyses or opinions, slop could be off-putting.\n\nEven for news, telling LLMs to answer basic questions required in journalism: 'why, when, where, who, how, what, which, etc.', i.e. research, which LLMs can be better at than human effort, is highly effective.\n\nBut choosing which answers are the most relevant for a given news story, how to interpret and present them, the appropriate theoretical tools to use, i.e. curation, empathy, reasoning (in the accurate definition*), judgement and entrepreneurship, are things humans are better at than LLMs. This requires time and care too. \n\n*For now, I can only say bluntly that LLMs aren't really 'reasoning', based on my understanding of rationalist ideas.",
"sig": "84b2d9af9f4581b1f3c5af143072d6bb87395c056c3e1cf4f66b8695c084638e9bd9e573e114a569f0e6bfaf0081b1c0acf533d8d047b75caf2d92238dd03b3a"
}