LLMs are spam generators.
Uncategorized
120
Posts
79
Posters
8
Views
-
@cstross LLMs are replacing every aspect of civilization - institutions, process knowledge, organizational competence - with civilization fan-fiction.
@cstross Because so many people boosted my post, I want to explore this topic in more detail.
First, a thought experiment: suppose we had cheaply scalable mind control, even if partially effective (LLMs are not mind control but they touch some of the same wires). Democracy would end, since the controllers could sway elections at will. The so-called free market would instantly become a command economy. Science, and any other activity requiring independent thought, would be a dead letter.
-
@cstross Because so many people boosted my post, I want to explore this topic in more detail.
First, a thought experiment: suppose we had cheaply scalable mind control, even if partially effective (LLMs are not mind control but they touch some of the same wires). Democracy would end, since the controllers could sway elections at will. The so-called free market would instantly become a command economy. Science, and any other activity requiring independent thought, would be a dead letter.
@cstross In the book _Rainbows End_, Vinge proposes a form of mind control called "YGBM" technology, which stands for "You Gotta Believe Me", a way to get affected individuals to accept arbitrary propositions as truth.
What this has in common with LLMs is technologically enhanced plausibility.
-
@cstross In the book _Rainbows End_, Vinge proposes a form of mind control called "YGBM" technology, which stands for "You Gotta Believe Me", a way to get affected individuals to accept arbitrary propositions as truth.
What this has in common with LLMs is technologically enhanced plausibility.
@talin I was going to mention "Rainbows End"!
-
@cstross In the book _Rainbows End_, Vinge proposes a form of mind control called "YGBM" technology, which stands for "You Gotta Believe Me", a way to get affected individuals to accept arbitrary propositions as truth.
What this has in common with LLMs is technologically enhanced plausibility.
@cstross Now, we know from the earliest days of machine learning that algorithms are capable of exploiting any loophole in their fitness function with seemingly godlike competence.
Unfortunately, we can't make a fitness function for intelligence, truthfulness, or integrity. What we can do, at great expense, is make a fitness function for plausibility.
When an LLM goes through its reinforcement learning, it's behavior is rewarded based on whether some human reviewer believed the result.
-
@cstross Now, we know from the earliest days of machine learning that algorithms are capable of exploiting any loophole in their fitness function with seemingly godlike competence.
Unfortunately, we can't make a fitness function for intelligence, truthfulness, or integrity. What we can do, at great expense, is make a fitness function for plausibility.
When an LLM goes through its reinforcement learning, it's behavior is rewarded based on whether some human reviewer believed the result.
@cstross As LLMs improve, they will get better at chasing their fitness function, that is, at convincing people that what they say is true. This is what LLMs have in common with YGBM.
Those who know history know that the invention of cheap printing sparked of centuries of bloody religious conflict. In this regard, LLMs are like "hold my beer".
-
Yet Saudi Arabia alone is laundering $100 billion of its looted national treasury on spam generation?
Perhaps 10% is being used on the spam-making software, the rest is being spent on a fossil fuel funded fascist movement, complete with civil rights erosions, state surveillance platforms, goon squads, concentration camps, and financial fraud.
We continue to underestimate how badly the fossil fuel industry wants permanent rule.
https://www.bloomberg.com/news/articles/2025-09-16/ai-deals-saudi-arabia-eyes-artificial-intelligence-partnership-with-pe-firmshttps://www.semafor.com/article/11/07/2025/uae-says-its-invested-148b-in-ai-since-2024
@Npars01 Yep, 💯
-
Yet Saudi Arabia alone is laundering $100 billion of its looted national treasury on spam generation?
Perhaps 10% is being used on the spam-making software, the rest is being spent on a fossil fuel funded fascist movement, complete with civil rights erosions, state surveillance platforms, goon squads, concentration camps, and financial fraud.
We continue to underestimate how badly the fossil fuel industry wants permanent rule.
https://www.bloomberg.com/news/articles/2025-09-16/ai-deals-saudi-arabia-eyes-artificial-intelligence-partnership-with-pe-firmshttps://www.semafor.com/article/11/07/2025/uae-says-its-invested-148b-in-ai-since-2024
@Npars01 @cstross
Spam making software also assists the nazis too.
"The ideal subject of totalitarian rule is not the convinced Nazi or the convinced Communist, but people for whom the distinction between fact and fiction (i.e., the reality of experience) and the distinction between true and false (i.e., the standards of thought) no longer exist." ~ Hannah Arendt -
@cstross Well, they are statistical models of language. To be precise, the corpus they were trained on, and there are a number of other aspects, like representation, what cleanups were applied.
Yes, one way to use these models, the one that seems to fascinate most people (but IMHO not necessarily the most useful, depending upon what one wants to achieve), is to complete a prefix with a plausible ending, statistically speaking.
Generally, they are a compelling NLP technology.
But yes, if you generate text with an LLM, you generally get text that is basically a "probable" text, based upon the training and the hyperparameters. And let's be honest, people that rely upon the training data of an LLM as a "search engine data corpus", they do have a problem.
🙅 -
LLMs are spam generators. That is all.
They're designed to generate plausibly human-like text well enough to pass a generic Turing Test. That's why people believe they're "intelligent".
But really, all they are is spam generators.
We have hit the spamularity.
@cstross Well, they are statistical models of language. To be precise, the corpus they were trained on, and there are a number of other aspects, like representation, what cleanups were applied.
Yes, one way to use these models, the one that seems to fascinate most people (but IMHO not necessarily the most useful, depending upon what one wants to achieve), is to complete a prefix with a plausible ending, statistically speaking.
Generally, they are a compelling NLP technology.
-
@VictimOfSimony @ViennaMike@mastodon.social If I was suddenly planetary overlord, I'd ban the advertising industry. Seriously, it's cultural poison. Doesn't matter whether it's cheap spam or brand marketing for YSL, it's garbage all round, designed to manipulate us into doing things we wouldn't otherwise do (for someone else's profit).
As a #PublicPolicy wonk I'm already remembering how different our common law legal systems have become in just three centuries. Over here they're still worried about the anti-porn #Censorship case where the #SupremeCourt refused to define what they were censoring. They famously said, "I know it when I see it," grumble grumble prurient use in commerce, harrumph typical opinion of ordinary firmness, followed by thirty years of chipping off corners without admitting they were sort of making shit up.
-
@davidgerard @cstross The article you posted included a very useful study. However, contrary to the article, that study is NOT the only study on the topic. Other, far larger, more comprehensive, studies show measured, meaningful productivity improvements. You asked for measurements, so here they are: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4945566 and https://www.youtube.com/watch?v=tbDDYKRFjhk The Stanford study found that the average productivity boost is significant (~20%) but some teams see productivity decrease.
-
LLMs are spam generators. That is all.
They're designed to generate plausibly human-like text well enough to pass a generic Turing Test. That's why people believe they're "intelligent".
But really, all they are is spam generators.
We have hit the spamularity.
@cstross I understand what you mean, but I’d say they’re simply databases. Spam generator is one of their many uses.
-
@davidgerard @cstross The article you posted included a very useful study. However, contrary to the article, that study is NOT the only study on the topic. Other, far larger, more comprehensive, studies show measured, meaningful productivity improvements. You asked for measurements, so here they are: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4945566 and https://www.youtube.com/watch?v=tbDDYKRFjhk The Stanford study found that the average productivity boost is significant (~20%) but some teams see productivity decrease.
@ViennaMike @cstross your paper's key metric is pull requests, which is addressed in the link you answered but didn't read (one jira leaves, three jiras enter). but thanks anyway.
(over and over i see this from AI guys, rebuttals that ignore the question asked)
-
@cstross Kessler Syndrome for the internet
-
@ThePowerNap @dragonfrog @cstross @ViennaMike I suspect that it'll be good for the ML space -- the conflation of LLM prompting with the harder skills required to actually train a model has been disastrous for the market. The industry isn't sophisticated enough for people to distinguish between the two skillsets, so the latter market has just been absolutely deleted in Australia.
@ludicity @dragonfrog @cstross @ViennaMike
I'm kinda betting on this, but it doesn't mean the next couple of years won't be painful.
-
@ThePowerNap @dragonfrog @cstross @ViennaMike The best purpose of a LLM system is to provide a natural language interface to a complex but not mission critical system such as home automation, media libraries and switchboards.
Machine learning on the other hand should lean into seeing without eyes (is a dramatic jump in intensity cancer?)
Both of these cases should be powered through a single IEC C13 socket and not some gW data centre.
@NefariousCelt @dragonfrog @cstross @ViennaMike
95% agree. I'm currently working an LLM project that is interfacing with something a touch more than home automation. Aside from fine tuning the model and putting constraints on the output (think https://github.com/dottxt-ai/outlines), you can use existing approved software as validation of llm output.
All of this requires in-depth knowlege, patience and exaustive testing.
-
@NefariousCelt @dragonfrog @cstross @ViennaMike
95% agree. I'm currently working an LLM project that is interfacing with something a touch more than home automation. Aside from fine tuning the model and putting constraints on the output (think https://github.com/dottxt-ai/outlines), you can use existing approved software as validation of llm output.
All of this requires in-depth knowlege, patience and exaustive testing.
@ThePowerNap @dragonfrog @cstross @ViennaMike I suspect I am coming from a very risk averse standpoint. The “summarise this” is what scares me with LLM marketing. I have been burned by poor software requirement specifications. Language simplification can do that in multiple sectors. But yeah mapping natural language to a fixed set of goals, that’s fine. Trust AI with my life, no. Sudden random steering inputs WTAF stelantis.
-
@talin @cstross I don't think there was ever a time in human history when critical thinkers were more than a small minority of the population. LLMs are unprecedented in the scale at which they can produce misinformation. Who knows? Exposed to so much bullshit, people may become more critical of what they see. It's already happening with images and videos, right?
-
@dragonfrog @ThePowerNap @cstross @ViennaMike i have occasionally been using an AI tool called Phind. It was basically a “help with coding” tool. It recently rewrote itself completely into a generic AI tool… and then a month later closed.
I cant help but wonder… Is this the start of the great collapse?
@leadegroot @dragonfrog @ThePowerNap
I can certainly see a bubble bursting, but I'm not sure that Phind is a sign. Hard to compete with those with billions to develop core models, and not going to hit any profit home runs reselling access to other's models. -
LLMs are spam generators. That is all.
They're designed to generate plausibly human-like text well enough to pass a generic Turing Test. That's why people believe they're "intelligent".
But really, all they are is spam generators.
We have hit the spamularity.
@cstross
The Turing Test will not fail an LLM for the same reason humans are ready to accept them as persons: humans evolved with a hardcoded behavior to perceive any language-using process as a cognitive entity and create a mental model of it. Rejecting that behavior takes a metacognitive ability that has not been taught to most people, so the problem isn’t rampant stupidity (look at how many clearly intelligent AI researchers fail this test) but rather ignorance, sometimes wilful, granted.
Feed RSS
Gli ultimi otto messaggi ricevuti dalla Federazione
-
@VeroniqueB99 too much windy, eh?!
-
This post did not contain any content.
-
Per chi utilizza Linux ed ha bisogno di un editor pdf esiste Master Pdf Editor. Software molto professionale, a pagamento. Se usate molto i pdf lo consiglio.
Se ne fate un uso sporadico, la vecchia versione 4 (ora siamo alla 5) è ancora "gratis per uso personale". Chiaramente ha qualcosa in meno rispetto alla 5, ma se appunto non fate cose sofisticate....
Dal sito ufficiale code-industry.net non si trova, ma potete trovare qui i deb dell'ultima versione usabile gratis per usi non commarciali
Il deb sul mio portatilino 13" con mint funziona benissimo.
https://www.linuxuprising.com/2019/04/download-master-pdf-editor-4-for-linux.html
-
@evan Yes, but only for federal matters (assuming a government with local/provincial/federal splits) and, as a matter of taste, only if you plan to return or for your family to return.
If you haven't got a stake in that country's future (and will not feel the effects of its policies) I do not think you _should_ vote, but given that's unenforceable, then I think you should _be allowed_ to.
-
@evan Yes, but not on local elections if you've been away for a long time. Another but: That there's some balance between the voting rights of expatriate citizens and people living in my country without citizenship (many of whom currently can't vote, but have lived here longer than some expatriate citizens and have a higher stake in what happens to the country).
-
Larry's not thinking straight. Bear spray all over his bedding? Using an air horn on mice? Welcome to crazy town, population: 1.
-
@evan no taxation without representation
-
@evan Americans are required to pay taxes to the US when we work abroad, though we're certainly not consistent on "no taxation without representation" :)
