> These detectors have bad rap among techies, but the objections are often based on outdated assumptions or outright misconceptions.
I did some tests with a the less conservative originality.ai, which flaged around ~2/3 of top HN posts as AI generated, but only 1 out of the 28 blogs in my RSS reader. That one? Machine translated from Italian. It definitely has some false-positives, but it seems like the fraction of LLM generated or LLM assisted content is very high.
Just preface every comment with "I have absolutely no experience as a judge or a lawyer, but I am still very smart and so have a strong untethered-from-reality opinion about this that others will surely want to know. Here it is:"
If the comment still makes sense with that preface, discard it, it's useless.
If it *doesn't* make sense with that preface , those are the valuable comments.
I genuinely don’t understand how some commenters go out of their way to say mean shit. Sadly I’m hooked to the site but never comment or contribute in any way
It makes sense when you consider dead internet theory.
There is inherent benefit arising out of demoralizing a population, both in terms of inducing delusion (so the victim jumps at shadows) as well as subtly manipulating and shaping opinion. Bots make this cost effective at scale.
Some people are skeptical about these things, and to those that are I'd suggest reading case studies from PoWs that were documented in the 1950s. Joost Meerloo and Robert Lifton's Psychology of Totalism cover the methods observationally in depth. Its only gotten better since then.
This is common knowledge in circles that regularly deal with thought reform, irregular warfare, political warfare, or behavior/social engineering (in an adtech context).
Distorting a viewers reflected appraisal is one of the fundamental levers on how these things work when it is coupled with several structural factors designed to impose unrelenting stress beyond a person's ability to cope.
The benefit comes from the behavioral changes imposed by stress as a result of breakdown; such as reduction of rational thought, hollowing out, involuntary hypnosis, and disassociation/delusion/psychosis.
There are also people that seek to cause valuable resources that would normally not be scarce to become scarce; for personal benefit; by drowning out everything in a sea of noise with the generalized intent to cause isolation, and isolation is one of those structural factors.
> These detectors have bad rap among techies, but the objections are often based on outdated assumptions or outright misconceptions.
I did some tests with a the less conservative originality.ai, which flaged around ~2/3 of top HN posts as AI generated, but only 1 out of the 28 blogs in my RSS reader. That one? Machine translated from Italian. It definitely has some false-positives, but it seems like the fraction of LLM generated or LLM assisted content is very high.
Hacker news is easy.
Take a story about a court case, for example.
Just preface every comment with "I have absolutely no experience as a judge or a lawyer, but I am still very smart and so have a strong untethered-from-reality opinion about this that others will surely want to know. Here it is:"
If the comment still makes sense with that preface, discard it, it's useless.
If it *doesn't* make sense with that preface , those are the valuable comments.
I genuinely don’t understand how some commenters go out of their way to say mean shit. Sadly I’m hooked to the site but never comment or contribute in any way
It makes sense when you consider dead internet theory.
There is inherent benefit arising out of demoralizing a population, both in terms of inducing delusion (so the victim jumps at shadows) as well as subtly manipulating and shaping opinion. Bots make this cost effective at scale.
Some people are skeptical about these things, and to those that are I'd suggest reading case studies from PoWs that were documented in the 1950s. Joost Meerloo and Robert Lifton's Psychology of Totalism cover the methods observationally in depth. Its only gotten better since then.
This is common knowledge in circles that regularly deal with thought reform, irregular warfare, political warfare, or behavior/social engineering (in an adtech context).
Distorting a viewers reflected appraisal is one of the fundamental levers on how these things work when it is coupled with several structural factors designed to impose unrelenting stress beyond a person's ability to cope.
The benefit comes from the behavioral changes imposed by stress as a result of breakdown; such as reduction of rational thought, hollowing out, involuntary hypnosis, and disassociation/delusion/psychosis.
There are also people that seek to cause valuable resources that would normally not be scarce to become scarce; for personal benefit; by drowning out everything in a sea of noise with the generalized intent to cause isolation, and isolation is one of those structural factors.
> it has a fair number of toxic commenters
Absolutely, I got used to get insults when I post something there unfortunately, regardless the content of the article.
As for the comment side: https://www.marginalia.nu/weird-ai-crap/hn/