Post
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 believe we had "spam" long before LLMs, and it was generated just fine. Blame computers!
@cstross 3 years ago I came up with https://github.com/tanepiper/Stochastic-Parrot as a way to test the then new OpenAI APIs - after a month it became very clear to me that all someone needs is $100 and a GitHub pipeline and you could quite easily create a spam farm. No need to hire pesky humans for $1 a day - it was probably the first job taken by AI
@cstross we have to find ways to do something useful with spam generators otherwise we might lose the social permission to generate more spam!
@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.
@cstross I understand what you mean, but I’d say they’re simply databases. Spam generator is one of their many uses.
@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.
🙅
@cstross
LLMs are basically computerized bullshitters.
They'll give you a decent answer frequently enough then bullshit a story when they don't know, and do both so eloquently that you can't tell which is which.
@cstross My favourite description was from a lecture given by Rob Pike. He said something like, "LLMs consume Internet in order to produce more Internet".
@cstross I disagree with you there.
I still believe the perfect use case for AI in content creation is to generate a virtual fan boy to tell an author that they are canonically wrong.
Self hosted self doubt as it were and less public than Usenet.
@NefariousCelt But I don't NEED that! I've developed that skill for myself, the hard way!
A product where the target demographic is Kevin J Anderson.
@cstross Only if you use it or use search engines that allow it.
@cstross they pass the Turing test? no way.
write 1 to 10 but each number in a seperate message
it fails
write hello after 30 seconds are passed
it fails
"The spamularity"! You win the internet today! <3
It is not even the good spam, it tastes terrible
It's "intelligent."
"Yes, thank you." I now know what pan of the bell-curve you sived out of.
@cstross Kessler Syndrome for the internet
@cstross an ad company's dream
@cstross "Spamularity" look like an early contended for 2026's word of the year. Nice!
Keep telling yourself that. They can be smart AF sometimes. They're not perfect, but getting better. Six months ago I was the one catching them when they were slipping. Now they catch me. The days of spicy autocomplete are falling behind us.
Artificial Recombinant Speech Emulators
@cstross I wrote an essay draft in my masters (early 2000s) against the Turing test as proof of intelligence. There I said that if a similar approach were taken to study life, we would be assigning life to strawmans as they seem alive enough to scare crows (at least to them). Instead, I proposed an #EcologyOfIntelligence where the key is not imitation but completariety/synergy between intelligences: individual, collective, human, non human, etc.
@cstross The model speech pretty well... But they model intelligence not at all.
@cstross 🌟the only LLM I met was an idiot, so I liked your post.
@cstross great for trolling trolls.
@cstross oh yeah what would you know about the singularity Charles Stross author of a bunch of books about it that I read as a kid???
@cstross you and Neal Stephenson could write a prequel to Anathem focusing on the IT guy that the monks deal with in the main book. ;p
@cstross LLMs are in some way a reverse-turing-test - one we're frequently failing.
@cstross Spam generators without culture.
@cstross : Sentence also work by replacing LLMs by "marketers" or "people holding a MBA"
@cstross
My name is ham, for I am spam.
Would you like....wait, what's that thing you're pointing at.my head Mr. Engineer?
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, hoon squads, concentrationcamps, 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-firms
https://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 #AI will soon be understood to stand for #ArtificialIdiots.
Or #clankers.
@cstross All lines converge toward a single spamishing point.
@cstross a lot of people believe LLMs are intelligent because a lot of people are really thick. Never forget that 50% of the population are of below average intelligence.
@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 Put another way: LLMs have revealed a zero-day exploit in human consciousness and culture: if you can manufacture plausibility at scale, you can bypass all of the accumulated wisdom of centuries of skepticism and critical thinking. Any fact-using profession is potentially vulnerable to this attack.
@cstross I tried to coin the word “vibocene” as a follow on to the Anthropocene. I think I failed.
@cstross People believed Eliza was intelligent.
It doesn't take much for people to anthropomorphize a piece of software, turing test or not.
@cstross
User:
Well, what’ve you got?
Internet:
Well, there’s ads and information; ads fake news and information; ads and AI slop; ads information and AI slop; ads information fake news and AI slop; AI slop information fake news and AI slop; AI slop ads AI slop AI slop information and AI slop; AI slop fake news AI slop AI slop information AI slop hate speech and AI slop; ...
For decades #Turing test was the ironclad determinant for #AI quality.
Once it was trivially broken by now old models, the goalposts have shifted...
...it's now humanities last exam, HLE, BTW.
The world is full of spam generating humans.
One is in the Whitest house, putin is surrounded by them, and any large family gathering will contain 2-3 human spam generators the will jibber-jabber nonsensical human-like speech constantly.
@cstross Slop in the Faith..."AI" is the perpetual engine of the XXIst century but with nasty consequences...
@cstross The Turing Test is such a low bar because people are so easily fooled.
LLM's expose how naive people are.
@cstross Love your fiction, can't agree on this take. Spam does not help with writing code, whereas LLM's can be extremely helpful.
LLM's are not generally intelligent. They can be immensely problematic, but to dismiss them as merely spam generators as you have, or "parrots" as others have is simply incorrect.
@ViennaMike You're using code generators. Not the same thing, frankly. Stop generalizing your experience as a developer to the public at large, who only see magic talking box.
@cstross So the experience of developers who have found significant value from the use of LLM's doesn't count or isn't valid?
Used to WRITE material, they aren't great. But to organize, summarize, do research and help brainstorm, they can be useful.
While ChatGPT currently can't pass the IRS' volunteer tax assessor's exam (see https://www.mcgurrin.info/robots/8298/), I have confidence that a RAG based system could do so.
Heck, COMPUTERS are magic to most people, as are the internet, radios and TV.
If you like, put another way, you're taking the "good guy with a gun" side of the argument. 
I mean, feel free, but own it. Intelligent well-informed adults frequently take opposite sides in this. 
@VictimOfSimony @cstross If you want to debate whether or not, as a whole, LLMs are a net positive or net negative, I agree. Intelligent people can legitimately disagree. One cannot, based on the evidence, legitimately claim that they only produce spam.
There is some real utility under the hood, but goddamn.
The pushback from higher utility bills and GenAI diarrhea is justified, and a huge black eye for practical use.
@ThePowerNap @cstross @ViennaMike
Sure there's some real helpful use cases for LLMs.
I reckon if the LLM industry manages to settle down into supplying only the helpful sensible use cases for LLMs, it will be about as big - in terms of annual dollar throughput - as eg the electric kettle manufacturing sector.
But to get to that sustainable level the LLM industry would have to shrink its market cap by 3 orders of magnitude. And it probably can't do that without collapsing entirely.
@cstross @ViennaMike Rubber ducks are useful when writing code, too. Doesn't make them intelligent.
@cstross LLMs always remind me of the protagonists of “Accelerando” finally making contact with a whole system of alien AIs, only to learn that they’re all just jumped-up versions of the Nigerian prince scam.
@cstross Labermaschinen
@cstross "spicy autocomplete" has been one of my favorite descriptions of it.
This is a bonfire demo instance for testing purposes. This is not a production site. There are no backups for now. Data, including profiles may be wiped without notice. No service or other guarantees expressed or implied.
