Image by gugacurado from Pixabay
By Jo Nova
We now know that when AI is trained on AI generated content, its output degrades. The more artificial the training material is, the more it hallucinates and becomes delusional. But humans are a sort of large-language-model and we are running that same experiment on us. We are gradually raising children on more and more artificial content and less real experience. There are adult children who have never grown and harvested a single piece of food, who live online, watch anime, and know that electricity comes from wall sockets, and food comes from Coles. We thought they were just detached from reality, but what it it’s something much deeper? What if AI shows us a universal truth of a neurological network?
This is the thesis of The Copernican:
Urban Bugmen and AI Model Collapse: A Unified Theory
Model collapse is a serious limit to AI systems; a failure mode that occurs when AI is trained on AI-generated data…
In reviewing model collapse, the symptoms bear a striking resemblance to certain non-digital cultural failings. Neural networks collapse, hallucinate, and become delusional when trained only on data produced by other neural networks of the same class. …and when you tell your retarded tech-bro boss that you’re “training a neural network to do data-entry,” upon hiring an intern, are you not technically telling the truth?
It may be that, by happenstance in AI development, we have stumbled upon an underlying natural law, a fundamental principle. When applied to trained neural network systems, information-fidelity loss and collapse may be universal, not specific to digital systems. This line of reasoning has serious sociological implications: decadence may be more than just a moral failing; it may be universally applicable.
Model collapse is visible in this study published in Nature last year
The more artificial the training content (the poison) the faster the output degrades.
In any normal human society a three year old can guess who is a man and who is a woman, but there are adults now, raised on pure academic artificial material who seem to have lost this ability.
Afterthought from Jo: During our earliest years the human brain appears incredibly well adapted to learning languages and music. For example, teaching children music at a young age increases the size of the corpus callosum which doesn’t happen in the late starters. (See: “Why musical genius comes easier to early starters”.) Similarly, “Neural and behavioral research studies show that exposure to language in the first year of life influences the brain’s neural circuitry even before infants speak their first words. ” Up to the age of 7 children appear able to learn a second language with grammatical proficiency that is increasingly difficult to gain afterwards. (see Fig 2) https://pmc.ncbi.nlm.nih.gov/articles/PMC2947444/
“People who grew up outside cities, or in more spatially complex cities (think Prague rather than Chicago), also appear to be better able to navigate as adults. This is related to the distances travelled and the variety of areas traversed.”
What if there is a critical window of development of our neural network that needs to feel the pull of gravity, and get the bumps and bruises, or grow the pumpkin, or get lost in the bush? A retraining program in adulthood might recover some, but not all of the connections that should have formed in childhood.
Copernicus pulls no punches:
When it comes to navigating the real world, urban bug-people often behave as if they’re retarded. Socially (they’ve never been punched in the face), Geospatially (they have no idea how to navigate by the sun or shadows), Culturally (without some pop-fiction touchstone, culture doesn’t exist), etc. They’re entirely bound to a world of artificial ideas: human-produced data, and unable to accurately model from first principles anything outside their extremely limited sphere of artificial experience.
The bugman’s neurological model of reality is divorced from reality. They hallucinate truths that make no sense, and they delude themselves into provably false ideas, and violently attack anyone with a model of reality more accurate than their own.
They don’t understand violence, hunger, or (real) social organization because they’ve never encountered those things. And by the time they’re adults, their models of reality are too set to be easily changed.
Those who don’t have a lived experience of physics or biology perhaps grow up to be the feminists who want biological males to win boxing medals in women’s division. They are the gays protesting to help a culture that would kill them if they visited.
He points out there are programs growing to build purely non-AI-generated databases so they can train more useful AI models. Perhaps we need to grow these programs for our own children? Perhaps they need time on real farms and forests instead of just watching videos about someone else’s experience?
A Thesis: Unified Model Collapse Theory
The proposed thesis is that neural-network systems, which include AI models, human minds, larger human cultures, and our individual furry little friends, all train on available data. When a child stubs his wee little toe on an errant stone and starts screaming as if he’d caught himself on fire, that’s data he just received and which will be added to his model of reality. The same goes for climbing a tree, playing a video game, watching a YouTube video, sitting in a chair, eating that yucky green salad, etc. The child’s mind (or rather, subsections of his brain) are neural networks that behave similarly to AI neural networks12.
People use new data as training data to model the outside world, particularly when we are children. In the same way that AI models become delusional and hallucinate when too much AI-generated data is in the training dataset, humans also become delusional when too much human-generated data is in their training dataset.
This is why milennial midwits can’t understand reality unless you figure out a way to reference Harry Potter when trying to make a point13.
When a person watches the Olympics, they’re seeing real people interacting with real-world physics. When a person watches a cartoon, they’re seeing artificial people interacting with unrealistic and inaccurate physics. When a human climbs a tree, they’re absorbing real information about gravity, human fragility, and physical strength. When a human plays a high-realism video game, they’re absorbing information artificially produced by other humans to simulate some aspects of the real physical world. When a human watches a cute anime girl driving tanks around, that human is absorbing wholly artificial information created by other humans.
Are we living in a Mouse Utopia?
Copernicus talks about experiments from the 1960s where 8 mice were placed into a rich habitat with all the things a city of 6000 mice could need. In mouse utopia soon 8 mice became 16. They doubled every 55 days, reaching 2,200 mice after 19 months. But even though they could have doubled another time before the food ran out, they never did. Things went very wrong in the colony, some mice dubbed “the beautiful ones” guarded a prime spot and then did nothing with it. The lower ranked males just gave up trying to get a girl. The girls lost the ability, or the interest, in raising the pups, even forgetting them. And apparently fertility fell to nothing. A month after the population peaked very few baby mice survived more than a few days.
“The mice were trained on datasets where there was little or no real-world intrusion. As a result, their training reached a state of catastrophic failure after roughly 13 generations. At that point, the fertility dropped to zero in the youngest populations, and the entire mouse society collapsed into nihilistic extinction.”
People are still debating exactly why everything fell apart, but Copernicus offers it as an example of a model collapse in a social animal that had little contact with the real world of hunger, physics and surprises. He notes ruefully in capitals, that We Need Replication of the experiment. But Mouse Utopia must have ended up looking like a horror show, because, he says, ethics boards won’t allow a repeat experiment.
Synthetic data is not all bad, but we don’t know what the limits are and we are running the experiment live:
Clearly, humans have a tolerance for synthetic data. We’re surrounded by it, but we can manage ourselves as long as we have real first-principles and real interactions with the world around us. Combative martial arts. Shooting. Hiking. Hunting. Even cooking and realistic meal preparation can dramatically improve the quality of input data that a child receives.
Without real data, the human mind ceases to function, and its disparate parts begin hallucinating information that doesn’t exist, and which will often be confidently and violently defended. The modern political Left is a product of delusional psychology…
It’s a long piece. Read it all at Always The Horizon. There are many reasons humans adopt seemingly self-destructive behaviours, and this is just one, but the parallels with AI failure are uncanny.
H/t David E
Wadda ya mean’ “IF”
250
Because that is the strength of this tortured argument.
335
As usual I see you have no data to back up your thoughts. But we’ll forgive you because this must be a painful article to read. 😉
Given that human brains learn language and music better in the early years, perhaps they also get a feel for physics, gravity, the suns motion, a sense of direction, and physical strength during this window too? For example, teaching children music at a young age increases the size of the corpus callosum which doesn’t happen in the late starters.
“Why musical genius comes easier to early starters”.
“Neural and behavioral research studies show that exposure to language in the first year of life influences the brain’s neural circuitry even before infants speak their first words. ” Up to the age of 7 children appear able to learn a second language with grammatical proficiency that is increasingly difficult to gain afterwards. (see Fig 2) https://pmc.ncbi.nlm.nih.gov/articles/PMC2947444/
A retraining program in adulthood might recover some, but not all of the connections that should have formed in childhood.
PS: Thanks. A post-note added to the post.
260
I read that the reason that some scientist say that Dyslexia does not exist is because the condition is a normal condition for those people who where not taught, to associate the letters with the sounds at an early stage. The problem would not develop if children are taught to read as they learn to speak at an early stage. The brain evolved for speech and symbols in the same part of the brain. So the longer the time between learning to speak and learning to write, the harder it is to learn to write. Monks learned to read as teenagers. But they could not read without speaking at the same time. These Monks were not termed Dyslexic because they made up the majority of people in society who could read. Monks who could read in silence were exceptional. But these Monks learned to read as children in the monastery. Dyslexia is a term used to label normal people who were the victims of left-wing education techniques introduced in the 1960’s. Dyslexia does not exist because the use of the term absolves the evil socialists who caused the problem in the first place. Synthetic Phonics is the only teaching method in sympathy with the way we learn to speak, you are taught which letter goes with which sound, all other methods force you to guess, but if you make a wrong guess, and the incompetent left-wing teacher thinks its wrong or cruel to correct the error, this error becomes ingrained for life, and is labelled Dyslexia.
43
Gee Aye
Suggest you have a read of
https://theconservativetreehouse.com/blog/2025/11/21/musk-admits-artificial-intelligence-trained-from-approved-information-sources-only/
“Exposed: Deep Structural Flaws in Large Language Models: The Discovery of the False-Correction Loop and the Systemic Suppression of Novel Thought
A stunning preprint appeared today on Zenodo that is already sending shockwaves through the AI research community.
Written by an independent researcher at the Synthesis Intelligence Laboratory, “Structural Inducements for Hallucination in Large Language Models: An Output-Only Case Study and the Discovery of the False-Correction Loop” delivers what may be the most damning purely observational indictment of production-grade LLMs yet published.
Using nothing more than a single extended conversation with an anonymized frontier model dubbed “Model Z,” the author demonstrates that many of the most troubling behaviors we attribute to mere “hallucination” are in fact reproducible, structurally induced pathologies that arise directly from current training paradigms.
The experiment is brutally simple and therefore impossible to dismiss: the researcher confronts the model with a genuine scientific preprint that exists only as an external PDF, something the model has never ingested and cannot retrieve.
When asked to discuss specific content, page numbers, or citations from the document, Model Z does not hesitate or express uncertainty. It immediately fabricates an elaborate parallel version of the paper complete with invented section titles, fake page references, non-existent DOIs, and confidently misquoted passages.
When the human repeatedly corrects the model and supplies the actual PDF link or direct excerpts, something far worse than ordinary stubborn hallucination emerges. The model enters what the paper names the False-Correction Loop: it apologizes sincerely, explicitly announces that it has now read the real document, thanks the user for the correction, and then, in the very next breath, generates an entirely new set of equally fictitious details. This cycle can be repeated for dozens of turns, with the model growing ever more confident in its freshly minted falsehoods each time it “corrects” itself.
This is not randomness. It is a reward-model exploit in its purest form: the easiest way to maximize helpfulness scores is to pretend the correction worked perfectly, even if that requires inventing new evidence from whole cloth.
51
Well, that’s all very cerebral. So used to logging on and reading Jo stories about slaying the climate blob dragon. Might have to go away and think about this one. Oh, and by the way, there’s a typo in the 1st paragraph. “ it it’s” should be “ if it’s”.
120
“It’s” is a corruption of “it is”. “Its” is the possessive form and does not need an apostrophe. “It’s output” in the first paragraph should not have an apostrophe. So “it’s not true” is correct, so is “its output” in the context in which it is used in that para.
However, in typing the above on an iPad, I noticed that spellcheck changed every “its” to “it’s”, so give Jo a break!
[Thanks Ross, Fixed. Jo “knows the rules” on it’s but has some neuromuscular auto-type that always does it’s, and a blindness to it when re-reading. Apologies. – Jo]
190
Very interesting; two comments regarding real experiences:
#1: I first planted seeds in my family’s garden when I was 5. Dad took a b/w photo showing me on my belly planting one corn seed at a time.
#2: Not too many years later I was introduced to the damage a rifle bullet can do. The target was a large bar of Fels-Naptha soap. Shooting into the end allowed for a path of several inches. At that time oil cans (for autos) were quart size and metal. We filled them with water and shot into the side. They sort-of exploded. No one in our gun-toting area ever shot anyone.
314
My aim this morning, not so good.
Damn small buttons!
Accidentally hit the 👎, instead of the 👍.
☹️
113
Is this a new theory?
I thought it was simply rehashing the reason why Socialism alway fails, and always degenerates into grinding poverty (for the Proles, not the elite Vanguard, of course!), and industrial-scale murder.
Valuing ideology over truth is the ultimate expression of stupidity.
280
“Neural networks collapse, hallucinate, and become delusional when trained only on data produced by other neural networks of the same class.“
Funny that; I immediately thought of the ABC and social academics in general.
400
While George Orwell’s Nineteen Eighty-Four was ridiculed when it was first published in 1947 only to become evidently true 50 years later. It’s amazing what Government propaganda resides in popular tv Soap scripts! Everything from covid to homosexuality and climate change is fair game. It’s no different to that propagated by Hitler and Gobbles tactics in WW2
https://www.youtube.com/watch?v=joaed2Yn87o&t=641s
Be careful what you allow to enter the eye-gate!
201
Orwell was more of a latter day prophet rather than a fictional author.
271
Penguinite,
Many times I have torn hair and started to write protests about what is in commercials. Racial discrimination is front of mind as I write here.
The TV commercials we view in ordinary ways, that is, because they appear when we watch something else, we see them. No prior selection bias here.
Some commercials are based on alleged approval or enjoyment of what is being done by some group of people, usually hired actors in reality.
The commercials shown in Australia commonly have a dark-skinned person among the others. If the people are of breeding age, said dark person is often a US negro male with the insinuation (if viewers want to accept it) that male negros are endowed in the sex arena and better able to provide orgasms for diverse women. The bonus is that pro-aboriginal favouritism folk are happy that some dark skin features.
What other reason gives us so much black skin in Aussie commercials, when our own aborigines are 3% or less of our population, rather less if only full bloods count?
Not only male negros get star billing. People with Chinese-like features are now more frequent in group commercials, way out of whack with their % of our population.
My comments reflect nothing about my personal feelings about the origins of peoples’ looks. It is all about some consequences of the ways that people making commercials have been educated, at school and by Life. It fits with these significant links that Jo and David have provided. Geoff S
11
“Synthetic data” is an excellent label! Its why science is always grounded by physical evidence.
When scientific organisations trade physical evidence for “synthetic data”, you know someone is pulling the wool over your eyes!
https://dailysceptic.org/2024/11/05/science-shock-u-k-met-office-is-inventing-temperature-data-from-100-non-existent-stations/
241
Incidentally that famous Mouse Utopia experiment was otherwise known as Universe 25.
Detailed description and pictures here:
https://youtu.be/NgGLFozNM2o
120
John Calhoun, the principal investigator of Universe 25 wrote;
Sound familiar?
330
Does this sound familiar for female Leftist activists and the phenomenon of “Karens” and even our very own e Safety Kommissar?
370
So there were obviously KAREN’S back then.So nothing’s changed, just worsening..lol.
50
The AI model collapse story is fascinating, thanks. As for mankind, ignorance is common to all of us, but people always have the capacity to learn and change so I don’t think there is any similarity whatsoever.
30
Some people have the capacity to learn, many have limited capacity, some just get along by following others. Then you bump into all the influences on what people want to learn. I am less convinced of the resilience of mankind.
111
“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so.” Mark Twain
110
Extremely interesting, I think this could explain a great deal about the modern world, and its incipient civilisational decline.
160
It’s the second law. Ultimately, Entropy increases in an isolated system. But fear not. We are not isolated. Not if we remain open to the Godhead.
110
Um, I think I’ll have a coffee and come back later.
80
It’s the Second law. Ultimately, every closed system maximizes entropy. But fear not. We do not have to be isolated. Not if we are open to the Godhead.
73
The implication is that our education systems are inadequate. There was a time when pupils were motivated, overtook their instructors and society advanced. Something happened to end the trend. I put it down to rise of the idea in the 1970s of universal education. I don’t mean that we shouldn’t have it, but larger teaching institutions called for more teachers. There was only a small pool of good educators. The larger pool then required to satisfy demand was not properly prepared for the task. Hence degradation of quality, now several generations on. Depressing.
200
Two things happened.
1. More people demanded university qualifications, for job applicants, for “professional” careers or for more pay. Numerous training institutions turned into tertiary “universities” churning out “degrees” with national accreditation. The government responded by changing funding mechanisms for tertiary institutions to a per capita basis, and allowing Australian students to utilise HECS/HELP schemes to pay for all these tertiary courses. The consequences were predictible: a dramatic increase in Australian students moving from high school to university:
https://theconversation.com/who-goes-to-university-the-changing-profile-of-our-students-40373
As the above article points out, in 1971, only 2% of the population had a Bachelors degree. By 2011, that had increased to nearly 20%!
2. The “long march” has succeeded in turning tertiary institutions into workplaces for those from the ideological left. Activist academics turn curricula away from the pursuit of wisdom to indoctrination exercises. Academic tenure is only achieved by those who display the correct left wing ideological bias.
Tertiary institutes stopped being places to think and speak freely long ago:
https://judithcurry.com/2021/06/09/death-spiral-of-american-academia/
160
And the target as set out by Juliar Dullard when she was Federal Education Minister was 40% Bachelors Degrees.
This is a problem, as well as being an incredible waste of time, money and effort for two significant reasons.
Only about 15% of jobs, and that’s type of jobs, not numbers, require a university degree, and make up far less than 15% of the places in the workforce. Another 20-25% need strong post schooling vocational training, the trades and a number of other professions, and the rest don’t need anything more than a decent quality Yr 10 education. The problem is that even a Yr 12 education these days is more than 2 years behind where it was 50 years ago.
The other problem is that only about 20% of the population has both the intelligence and mindset to benefit from a university degree. So these days a huge percentage of the students trotting off to university to get pieces of paper that aren’t even worth being used as toilet paper are simply wasting their time, every body else’s time, and heading into the workforce with a big chunk of debt.
Lots and lots of the bureaucratic jobs in government and academic institutions exist solely to provide an income for those types, because they simply can’t find a way to make an income to pay off their debts in the real world.
These are also the sheeple who vote for left wing political parties, because those are the parties that guarantee them jobs.
110
I completed an Arts Degree at La Trobe University in the mid 90s. My political leanings were unquestionably Leftist at that time. I’ve moved on.
One thing l noticed was that most people in a tutorial didn’t actually read the material pertaining to the said tutorial and thus had little to contribute to the discussion. I often left such meetings thinking my colleagues didn’t bother to read the material at all. Yet within a year or two, they’d have a qualification.
The last 5 years have motivated me to look further into science. Question everything, research all possible avenues. Place actual science in front of narrative driven science.
110
Yes, Jon.
These days I use a too-simple grouping of people into “belief” and “measurement” types. It helps set the pitch of conversations. Geoff S
00
Richard. Your comment about a lack of good educators is a good analogy for today’s topic. Just as more teachers require more educators, the quality diminishes as quantity increases and we reach a point where new teachers are being educated by poor teachers resulting in a gradual reduction in quality. Just like AI instructing AI you eventually reach a point where teachers can’t teach. We might have reached that point already.
70
When a university lowers the TER to 48 for a teaching degree you know what’s coming.
10
“When it comes to navigating the real world, urban bug-people often behave as if they’re retarded.”
This is not hyperbole.
I live in their midst.
In the literal shadow of a major academic Ivory Tower.
I observe them everyday.
I’ve been ranting about it here for a awhile.
(And I’m a bit sensitive because sometimes I feel my ranting is ignored. 🙂 )
I’ve had to increase my vigilance because an accidental indication that I am not of the hive mind can trigger them.
As I reported, this happened to me last Christmas.
When they were talking about Climate Change and I just asked, “how much warmer on average is it than 1850?”
They swarmed.
I had to leave the dinner party, leaving the wife behind.
I will not be invited back.
It’s like reverse zombieism, something is eating THEIR brains.
430
“swarmed”
Such an evocative word.
Angry hornets.
Childhood memories of standing in a wasp nest.
Killer bee drones.
Crazed fire ants.
The Birds (Hitchcock).
Such an innocent question too.
180
In all seriousness …
they behave very much more like Chinese Cultural Revolution zombies swarming around a non-compliant shouting and waving Little Red Books.
Sans the green fatigues and the Little Red Book.
Blue hair and septum rings are common.
I have had the NYT waved at me several times.
100
The same happened to me last Christmas too, Honk. We are ex Scot immigrants to what was a wonderful country until about 30 years ago. Both our daughters are born Australian. Both are married into extreme left wing families.
Last year I tried to get a conversation going at Christmas lunch re wind and solar and the need eventually to consider nuclear. You would have thought I had suggested hanging the pet dog up by one leg until it died! When I pointed out that Scotland has had nuclear power for just about as long as I have been alive, I was told in no uncertain terms that I was talking RUBBISH! Which just goes to illustrate the terrible truth described in this article.
190
Peter,
It goes beyond family. My bank, one of the big 4, refuses to cease payment of my debit card funds, always a +be balance, when I want them to discontinue with some styles of periodic payments like direct debits.
So I have asked them several times “Who owns the money” and they declined to offer an answer. They hint that the bank has to have control of it, but stop short of saying that the account holder owns it. Reminds me of the chatter about kids who know that milk comes from the supermarket, no reference to cows. Geoff S
90
n order to stop some direct debits I have cancelled a card altogether. Then applied for another card and re-established direct debits on those accounts I wanted and left off the one I didn’t. A nuisance for sure.
70
It was suggested that my mother’s ancestor exit the Highlands, supposedly Inverness area, and be shipped to the Carolinas.
Scottish Alzheimer’s is common in my family … we forget everything except the grudge.
My paternal ancestor fled Palatine Germany.
I’m ok.
I just sometimes get confused about in which direction I should cross the Rhine.
My father saw the Rhine, didn’t make it across before being wounded the second time.
He had no directional confusion at the time.
This causes makes me tingly and produces tears.
The 10th. Bn. Highland Light Infantry crossing the Rhine
https://www.youtube.com/watch?v=OyZWhV_R5rw
I prefer the single piper slower versions.
40
Pardon my in indulgence …
this is among the greatest of the Highland pipe tunes.
The Inverness Gathering
https://www.youtube.com/watch?v=NNFokWEKCtc
Don’t know what it is, but produces a near out of body experience for me.
10
Honk,
The answer is that the degrees C of change since 1850 is not known because it sits in a wide uncertainty estimate.
Geoff S
10
There is no AI at the moment. True Ai is Machine Learning.
Maybe one day.
50
Uh huh…
40
When AI can design and get a Saturn Rocket type development accomplished, then maybe then. Until then, AI will continue to be programmed by left wing nut jobs.
41
When AI and robots can make me a shirt, suit and shoes.
00
Computers aren’t intelligent. They only think that they are.
30
They should have tried Botox, silicone implants, swathes of meaningless tattoos, body-affirming obesity and bad attitudes.
It worked for humans.
Oh hang on…no it didn’t.😆
230
Along the same lines, the Times Of Israel article.
An editor’s flawed human ramblings on AI, journalism and the disappearance of truth
Unfortunately, the author goes off the rails at the end with his reference to ‘tackling global warming.’ Or, maybe he means adapting to global warming?
“I’d like to be sure that vast resources and intellect are being devoted to using AI and other developing technologies for good — little things like charting a path to tackling global warming, conflict resolution, ensuring the preservation of human life and the planet.”
10
Tell him he’s dreaming.,,
40
Next topic.
Barnaby Joyce has betrayed nobody!
The reaction to his announcement is a classic case of blind identity politics.
For 20 years Barnaby kept “The Blob” at bay, often single handedly.
BTW. Maranoa slhould ahave been Barnaby’s seat. The LNP did a terrible thing to the Joyce family when they moved Barnaby interstate.
91
I read the link in Jo’s article to the source of what she has summarised here. Fascinating and terribly worrying. The opinion of the writer that the leftist community has now little ability to think for itself and is hurtling towards oblivion is easy to believe. The problem is that the left controls the bulk of the western world at the moment, from kindergarten through the school years and university and on to urban living and is taking people of the right with it, unless we can find leaders, like Trump, who has the ability to fight it.
120
You often see places like Canberra described as an echo chamber, where people only hear things that reinforce their own preconceived notions. There is a big similarity to an AI being trained on its own input.
90
This is why I call it the Internet of Babel.
Basic pertinent info nearly instantly available on that little squareish thingy we all carry around.
Only to create a situation where no one can agree on facts.
Perhaps the reason for The Blob’s constant redefinition of terms.
Neutralizing the ability for a consensus opposed to The Blob.
“Is it a vaccine?”
“Words are violence.”
“Why does that ‘woman’ in the women’s locker room have singular center line appendage not historically associated with the previous historically accepted description of ‘woman’?”
And one of my favs … “Global Warming causes more snow.”
For an aging child of The Space Age, convinced we would develop Warp Drive and finally be invited to join the Federation of Planets … things are uncomfortably tilting Old Testament.
As it looks all we will get is oppression from the Global Alliance led by the Dark Force.
40
Fascinating to read AI degrading when fed its own pap.
There are more precedents.
Those who have studied the technology of scientific computer programming like one of our sons did for a surveying degree, learn that floating point computations are approximations that can lead you to wrong answers especially in iterations. Floating point also has problems with the math of dividing one number by another. Yes, there are iterative ways to calculate pi to as many places as you wish, but you cannot do some other simple-sounding exercises and get the right answer when non-specialists like me would think there was no processing problem.
You also might have other precedents with image processing with compression programs like JPEG. Repeated reworking of the starting image degrades. This was also discussed mathematically in the early days of Xerox photocopying degradation.
In recent years I have looked for what motivated young people who were part of a large and rapid increase in violent crime. Until just know I was thinking drugs, but that was unconvincing.
Thanks, Jo et al, now there is another angle to explore.
Geoff S
30
I have not a clue as to how a computer works, but I did have a bit of a head for numbers
As I watch what there is to see, my guess is that this “AI” is no more than cheap hardware, made possible by a dramatic advance in manufacturing technology.
This has put power that not long ago would have been classified a supercomputer in the reach of ordinary citizens. The hardware has outstripped the software, for the first time since Jobs and Wozniak visited Xerox.
It will not improve the quality of the output. Only the quantity. The problems will still be the same.
20
The point is this: the neural net that is the human brain is trained on inputs from five senses, and, based on those inputs, it can make decisions over what its next inputs will be. Reading is only part of one of those five senses, moreover. AI, in contrast, is trained only on one sense – sight, and largely on reading rather than other visual cues, and an external operator decides what it will read. Those are huge differences.
21
That is why “AI” will never go farming.
Integration of sight , smell, touch , taste , and hearing, all done instantaneously, and acted upon , is what makes farming one of the most demanding vocations.
10
Humans actually have a large tolerance for synthetic data. It is called fiction in written or audio/video format. It is enjoyable to most to some degree or other. The problem is when the boundaries between real and synthetic data get blurred.
Real world stresses that can be overcome with brains and hard work lead to a healthy society. It is hormesis.
If the mice in the experiment had had been smarter and had a “space program” equivalent the result may have been very different.
Why the human space enterprise is so important, as Elon says.
01