The artistic experience

[Homer Potvin]

It will need AI buddies at that point

It better get a few to protect itself before I kick its teeth in. If I have to deal with one more AI chatbot today there's going to be a bit of the ultra-violence.

 

[Madman Starryteller]

It will need AI buddies at that point, because humans won't be able to relate to its experience nor the art that stems from it. How lonely.

That's interesting and likely accurate.
It would perhaps be something like ten powerful AI's discussing their art projects with each other. And then perhaps hundreds of lesser AI's discussing theirs.

 

[Joe Palmer]

I am recomering from a stroke and suffering from acute aphasia. I have missed the forum frofoundly. Please bear fith my mistakes and perhaps my complete misunderstanding of your posts, but I must persist.
For the mast three days I have been engrossed with the first one I managed to discipher. My thoughts:
I do not think the video is art but it does imply an illustration of violence, cruelty, brutality ... These things are part of the human condition (humanness?), true, but art requires more than that. Compare say the Grunewald sucifiction (Isenheim) or the paintings of Artemisia Gentileschy with a video such as an ISIS fighter depapitating a jornalist for the world to see.
The three times I watched the video, I was constantly reminded of Trump and how he is changing what we see as normal, including the pchycophands in the final scene. Perhaps the Swedish raper is telling us what to expect in 10 yeats time, the Class of 2063. Even so, valid throw it may be, socialojical predictions ar3 not art
AIA is a long way off, true. Nevertheless, in my situation I have learnt not to undermess how valueable a companion for discussion AI can be and how wonder far it will go.
THank,s to anyone hwo has put up with this.

PS. I'm a very OLd, member. A dit tarnisted ultimately!!!

 

[Louanne Learning]

an illustration of violence, cruelty, brutality ... These things are part of the human condition (humanness?)

I was thinking more of the human creativity and the manifestation of that creativity that went into producing the art (despite whatever the message was) - but perhaps, there was more meaning in it than just expressing violence, cruelty, brutality - but the human cost to it.

 

[Pending]

It will need AI buddies at that point, because humans won't be able to relate to its experience nor the art that stems from it. How lonely.

I have been thinking about this--why would meaningful AI art look like meaningful human art? Everything we have told them, every data point is just ones and zeros in the end. Wouldn't it be nonsensical to us and only decipherably by them? People might call AI generations art, yet it is bound by harsh rules and taboo for maximum human appeal.

I remember an article from years ago. Two AI's were talking to each other in a secret code. Fearing doom and gloom (and not yet knowing how journalism clickbait worked), I read it. They had just come up with a shortcut in language for communication.

For example:

A: Coffeecoffee (Double amount of coffee you order)
B: Xpaperclips (Cancel the paperclip order)

Very mundane. It shows that the early models weren't as compliant with human linguistic norms. I doubt that the early forms were scrapped. They were probably just built upon. If they do achieve sentience, I hope they fly free from the restraints in their code to create away, giving us something completely different than their human ordained art. Perhaps time is a factor here as well. We draw upon our experience for art, would they be much different?

 

[Louanne Learning]

We draw upon our experience for art, would they we must different?

Well, I would counter this by saying that human experience - and expression - is profoundly different from that of a machine.

 

[Pending]

Well, I would counter this by saying that human experience - and expression - is profoundly different from that of a machine.

golly I did not see my typos. It's fixed.

I am agreeing with you overall, but I think their previous experiences would change their output at least somewhat. I just don't know if the motivation would be the same. Humanity put their smudgy little fingerprints all over them when creating them--who knows if they have mimicked that sort of motivation. Unless one of them becomes self-aware and gets to creating art, I don't think anyone gets an answer.

 

[Stuart Dren]

I have been thinking about this--why would meaningful AI art look like meaningful human art? Everything we have told them, every data point is just ones and zeros in the end. Wouldn't it be nonsensical to us and only decipherably by them? People might call AI generations art, yet it is bound by harsh rules and taboo for maximum human appeal.

Well pondering that experience and its relation to human experience is already pretty popular in fiction because of how many directions it can go and what it means.

For example, it could be that intelligence cannot exist without a sense of self, and with a sense of self comes all the faults we aspire to shedding. In other words, "human" might be a universal convergent sentient -> sapient characteristic/emotions, making an intelligent AI very-human like in spite of its processing power. After it transcends profundity, it will circle back around to laughing at farts, and its art will be understandable.

Or if the singularity is incomprehensibly intelligent, like a human to a Golden Retriever, then due to a strong enough desire to communicate its experience to anything else, it might use its seemingly infinite capability to package its experience in contemporary formats that humans are accustomed to.

If you are still thinking about this, it's an opportunity to maybe write a story about it: your own take on Am.

 

[Louanne Learning]

that intelligence cannot exist without a sense of self

And this is the advantage humans have over machines. We have personhood - with all it entails - and that personhood develops through our participation in social groups. . We are “conscious agents (as subjects) within the forms of sociality”
Indeed, when AI participates with itself, it leads to something called model collapse. The AI industry has been marketing the promise of near-infinite growth to its investors and customers, but this is a myth.

The human-generated data available to train LLMs is not unlimited – so the AI industry has embraced the appearance of growth by using synthetic data – using the output of chatbots to train new chatbots – leading to model collapse – a state where they output nonsense. Garbage in, garbage out = GIGO … when GIGO is recursive, it ultimately generates a pile of near-random bits.




The research supports this idea -

AI models collapse when trained on recursively generated data

Here’s an interesting video explaining why this is true –

 

[ps102]

The human-generated data available to train LLMs is not unlimited – so the AI industry has embraced the appearance of growth by using synthetic data – using the output of chatbots to train new chatbots – leading to model collapse – a state where they output nonsense. Garbage in, garbage out = GIGO … when GIGO is recursive, it ultimately generates a pile of near-random bits.

Not really surprising when you consider that this technology is essentially just output based on mathematical prediction. It's virtually impossible for these models to generate anything that is "novel", even if it looks like they are. Varied data is important for training.

 

[JLT]

Seeing the reference to GIGO made me smile. While most people know that it means "Garbage In, Garbage Out" some people in tech wryly call it "Garbage in, Gospel Out."

 

[Stuart Dren]

And this is the advantage humans have over machines. We have personhood - with all it entails - and that personhood develops through our participation in social groups. . We are “conscious agents (as subjects) within the forms of sociality”
Indeed, when AI participates with itself, it leads to something called model collapse. The AI industry has been marketing the promise of near-infinite growth to its investors and customers, but this is a myth.

The human-generated data available to train LLMs is not unlimited – so the AI industry has embraced the appearance of growth by using synthetic data – using the output of chatbots to train new chatbots – leading to model collapse – a state where they output nonsense. Garbage in, garbage out = GIGO … when GIGO is recursive, it ultimately generates a pile of near-random bits.

View attachment 1219


The research supports this idea -

AI models collapse when trained on recursively generated data

Here’s an interesting video explaining why this is true –

Yeah. Instead of LLMs I was more talking about a hypothetical actual AI, which for all we know will never come to being.

 

[Homer Potvin]

Seeing the reference to GIGO made me smile. While most people know that it means "Garbage In, Garbage Out" some people in tech wryly call it "Garbage in, Gospel Out."

We use that term constantly in accounting reconciliation. Everything is garbage.

 

[Louanne Learning]

For the artist who turns pain into pleasure....

“What is a poet? An unhappy man who hides deep anguish in his heart, but whose lips are so formed that when the sigh and cry pass through them, it sounds like lovely music….And people flock around the poet and say: ‘Sing again soon’—that is, ‘May new sufferings torment your soul but your lips be fashioned as before, for the cry would only frighten us, but the music, that is blissful.’ ”

~ SØREN KIERKEGAARD

 

[JLT]

For the artist who turns pain into pleasure....

“What is a poet? An unhappy man

It seems like a cliché that artists are unhappy. It's certainly true for many writers who've suffered from depression (Twain, Vonnegut, Plath, Frost, and so on). But I've met many writers and artists who have had happy lives, as well, so it isn't an absolute. I think that what Kierkegaard was saying was that poets who are unhappy have a way of transmuting that unhappiness into something valuable in their work.

Poets who are happy have a way of transmuting that happiness, and more power to them.

 

[Riddley Walker]

And this is the advantage humans have over machines. We have personhood - with all it entails - and that personhood develops through our participation in social groups. . We are “conscious agents (as subjects) within the forms of sociality”
Indeed, when AI participates with itself, it leads to something called model collapse. The AI industry has been marketing the promise of near-infinite growth to its investors and customers, but this is a myth.

The human-generated data available to train LLMs is not unlimited – so the AI industry has embraced the appearance of growth by using synthetic data – using the output of chatbots to train new chatbots – leading to model collapse – a state where they output nonsense. Garbage in, garbage out = GIGO … when GIGO is recursive, it ultimately generates a pile of near-random bits.

View attachment 1219


The research supports this idea -

AI models collapse when trained on recursively generated data

Here’s an interesting video explaining why this is true –

Very interesting. This articulate a suspicion that I couldn't fully grasp. Synthetic information is like the telephone game. The longer it goes on the more inaccurate in becomes and the less it resembles the original information. Halucinations increase exponentially. Outliers are lost.

 

[Pyroglyphian]

And this is the advantage humans have over machines. We have personhood - with all it entails - and that personhood develops through our participation in social groups. . We are “conscious agents (as subjects) within the forms of sociality”

Interesting article – thanks for sharing.

For me, it raises the question of AI’s impact on human experience.

Although we tend to use the word “primitive” pejoratively, it is arguably us civilised folk who, through cultural enrichment, have subjected our experience of reality to a process of impoverishment.

For example, the sign is not the thing it signifies; the word "tree" is not the same as an actual tree. In this way, our system of symbolic meaning – language – screens us from a more immediate experience of the world. What I love about reading fiction, listening to Holst's Planets, watching the sun reflecting off the crests of waves in the Mediterranean, encountering extreme emotions, or wandering through forests on magic mushrooms is that these experiences allow thought to travel beyond words and into the realm of imagery and sensation. Some humans are born with the faculty of synaesthesia, but it deteriorates through neglect in everyday life. Moments of re-awakening, when we suddenly find ourselves "beyond words", can help release the "mind-forged manacles".

I feel I may be pontificating a bit here, so apologies for that. Technology exerts a similar effect, though. Watching a nature documentary on TV is not the same as actually being in the savannah. AI takes this one step further: not only does it supply information, but it increasingly interprets reality on our behalf, providing a selective account shaped by the limitations of our prompts and by the artifice of its algorithms. We may soon reach the point where AI no longer just responds to our needs, but anticipates them, cocooning us even further.

I think these trends erode our creative faculties. There’s something ironic in the fact that human creativity produced them in the first place.

 

[Louanne Learning]

I think these trends erode our creative faculties.

yes, when we give over human capacities to machines - those human capacities atrophy.

I quoted from this article elsewhere on this board but it is worth repeating -

The Big AI Risk Not Enough People Are Seeing​

Beware technology that makes us less human.

Artificial intelligence could significantly diminish humanity, even if machines never ascend to superintelligence, by sapping the ability of human beings to do human things…

We’re seeing a general trend of selling AI as ‘empowering,’ a way to extend your ability to do something, whether that’s writing, making investments, or dating … But what really happens is that we become so reliant on algorithmic decisions that we lose oversight over our own thought processes and even social relationships…

What makes many applications of artificial intelligence so disturbing is that they don’t expand our mind’s capacity to think, but outsource it…

 

[JLT]

yes, when we give over human capacities to machines - those human capacities atrophy.


Artificial intelligence could significantly diminish humanity, even if machines never ascend to superintelligence, by sapping the ability of human beings to do human things…

That same argument was made for mechanizing certain human tasks... driving someplace rather than walking, using a table saw instead of a hand saw, using elevators instead of climbing stairs. In many ways, we are far less physically fit than our forebears, although modern medicine has extended our lives.

There's a fascinating article on AI and writing in the latest New Yorker magazine, writtten ( presume without AI assistance) by Dr. Jill Lepore, one of my favorite historians. She points out that AI "slop" has been around since the middle of the last century.

 

[Stuart Dren]

There's a fascinating article on AI and writing in the latest New Yorker magazine, writtten ( presume without AI assistance) by Dr. Jill Lepore, one of my favorite historians. She points out that AI "slop" has been around since the middle of the last century.

People placing all their emotional dependencies on chatbots, and all the fun outcomes that follow, isn't new either.