/lit/ - Literature

[Anonymous No.17587188]

What will literature look like in 10 years when AI starts writing it better than humans?

[Anonymous No.17588018]

>>17587188

[The following reply was written entirely by GPT-3 a natural language model developed by OpenAI...be afraid]

>What will literature look like in 10 years when AI starts writing it better than humans?
It will likely be a gradual transition, with AI-assisted writing followed by AI-directed writing, and eventually AI-authored writing by the end of the decade. As AI can generate thousands of pages in seconds, the volume of a text will matter less than it’s content.
Literature with easily replicated structures, such as romance, crime, self-help, and mystery, will likely see an uptake in AI authorship, since the genre’s structures can likely be easily replicated within the constraints of what makes a good story.
While it is impossible to predict which genres will be most impacted by AI authorship, one thing is clear: When AI starts writing, the number of books produced will increase significantly.
This is because AI doesn’t have all the same time considerations as human authors—AI can write continuously, 24 hours a day, 7 days a week. This will place pressure on editors, who will also be using AI to screen the new submissions, effectively forming an adversarial network.
Short-form writing and self-publishing may become more prevalent. A focus on humanistic writing, like poetry, haiku and creative non-fiction, will likely remain the same, since these genres involve a more human authorial touch.
The farther away we get from AI-authored writing, the more discussion there will be about the “creative” nature of literature. It is possible that a new type of appreciation will arise for both good and bad writing – similar to how we can appreciate the beauty in sculptures, even if they were created by artificial intelligence.

[Anonymous No.17588153]

>>17588018
>look mom I used the glorified Markov chain generator again

[Anonymous No.17588195]

>>17588153
>He typed into his 1+1 machine

[Anonymous No.17588237]

>>17588018
>>17588153
>>17588195
explain, please

[Anonymous No.17588244]

>>17588018
wtf how come AI is a pseud

[Anonymous No.17588362]

>>17588244
‘pseud’ perfectly describes AI. It is optimized to be extremely convincing despite being completely hollow inside.

The fun part is to just hit ‘refresh’ and get an entirely contradictory answer.

[Changed the question a bit...]
>What will literature look like in 10 years when we realize AI can never write better than humans?
Twenty years ago, Google was a one-trick pony. Today we know it as the world's most powerful search engine. We also know it as a leading source of answers for our questions on the Internet. But we didn't always know this in advance. Search engines of that era were considered a novelty, unreliable, and were often used by technically savvy people who were willing to tolerate their quirks just to get information quicker than by using traditional search methods such as books or Web sites.
It follows that many suspected Artificial Intelligence (AI) would have a similar destiny. Early AI researchers were more excited about their research than anyone else; they were convinced that AI would change the world (though they disagreed whether the change would be for for better or worse). There are significant limitations to AI in general. For example, there are clear limits to the knowledge that AI can acquire (and there is evidence that many researchers have not fully appreciated this limitation). In addition, it may be impossible to write code for creativity in general.
While AI may be a valuable tool in the creation of content, as spelling and grammar checkers, for example, it is clear that AI cannot write better than a human. If the intent of literature is to entertain, the AI may be capable of entertaining, but if the intent is to move readers the AI will never be able to achieve this. In fact, the AI may not even be able to write stories with more than a handful of "identical" endings.
So, what will literature look like in 10 years when we realize that AI cannot write better than humans? I believe that it may not change from what it looks like today. There are thousands of writers who have a gift for enlightening us with their words, and this gift cannot be replicated by an AI.

[Anonymous No.17588365]

>>17587188
It won't

[Anonymous No.17588387]

>>17587188
Exactly like it looks now. 90% will schlock like “How Unf*ck Your Life in 10 Days or Less” and endless novellas about colonialism or something, which no one will read. This medium has been reduced to rubble by so many actors already, it’s almost irreparable no matter what.

[Anonymous No.17588395]

won't differ much from bug-pilled refutation of god and beauty and range of human sensation that we have now. I don't see how a robot could believe in god, and thus, I don't see how a robot could believe in higher art. Would probably just pump out the same kind of YA dreck that dominates the literary scene now

[Anonymous No.17588639]

>>17588362
Link?

[Anonymous No.17588668]

What will have a greater impact:
AI-assisted writing or AI-assisted content discovery?

Look at this board, everyone’s looking for recommendations and they’re all just pitching the same ~200 books to each other. We really need a better way to discover good work, rather than simp over the same shit.

[Anonymous No.17588677]

>>17588639
https://openai.com/blog/openai-api/

[Anonymous No.17588918]

>>17588362
>Twenty years ago, Google was a one-trick pony. Today we know it as the world's most powerful search engine. We also know it as a leading source of answers for our questions on the Internet. But we didn't always know this in advance. Search engines of that era were considered a novelty, unreliable, and were often used by technically savvy people who were willing to tolerate their quirks just to get information quicker than by using traditional search methods such as books or Web sites.
>It follows that many suspected Artificial Intelligence (AI) would have a similar destiny. Early AI researchers were more excited about their research than anyone else; they were convinced that AI would change the world (though they disagreed whether the change would be for for better or worse). There are significant limitations to AI in general. For example, there are clear limits to the knowledge that AI can acquire (and there is evidence that many researchers have not fully appreciated this limitation). In addition, it may be impossible to write code for creativity in general.
>While AI may be a valuable tool in the creation of content, as spelling and grammar checkers, for example, it is clear that AI cannot write better than a human. If the intent of literature is to entertain, the AI may be capable of entertaining, but if the intent is to move readers the AI will never be able to achieve this. In fact, the AI may not even be able to write stories with more than a handful of "identical" endings.
>So, what will literature look like in 10 years when we realize that AI cannot write better than humans? I believe that it may not change from what it looks like today. There are thousands of writers who have a gift for enlightening us with their words, and this gift cannot be replicated by an AI.
Is this written by AI? Forgive me anon but I haven't caught up. If so, is this open to public?

[Anonymous No.17589313]

>>17588918
Yes (OpenAI’s GPT-3)
If you want to access a bit of what it can do, you can try out Shortlyai. I’ve been using it to assist in short story writing in the ffa thread here:
>>17573960
Some of the stuff I wrote:
>>17523786

[Anonymous No.17589349]

>>17587188
You'd have to be an imbecile to trust an AI to tell you what it will be like when AI replaces human writing.

Pic related is what it will be like

[Anonymous No.17590861]

>>17589349
genuinely frightening

[Anonymous No.17590918]

>>17589349
Yes, trust some random 4chan pseud instead.

[Anonymous No.17591188]

>>17589349
>linking an article from 2009 while posting in 2017
Kek yes one isolated incident surely means that the unwashed masses are ready to revolt.
Poors and middle classers dont have the language or the capacity to imagine a real, meaningful revolution. They are easily swayed by symbolic victories are have been conditioned to loathe minor discomforts and anticonsumerism
You will never, ever see this kind of violence occur in our lifetimes. Though it would be beautiful if a bunch of Wall Street goons got drawn and quartered while the other vultures get tossed out of ten story windows

[Anonymous No.17592067]

>>17591188
>implying that that's what it's about, and not algorithms entrenching themselves into every facet of the market

[Anonymous No.17592615]

>>17591188
The last part is saying that in the 2020s people will start murdering content creators not because they hate richfags, but because the content creators are pushing a type of algorithmic brainwashing that disagrees with the lynchers own algorithmic brainwashing.

[Anonymous No.17592756]

>>17588018
>https://www.youtube.com/watch?v=LSHZ_b05W7o
Yeah, nah.

[Anonymous No.17592773]

>>17587188
00001010100001010101101010101010110

[Anonymous No.17592989]

>>17588395
a robot will have the exact same ability to believe in god as a human as theres is no fundamental difference between the two

[Anonymous No.17593010]

>>17588677
im retarded how does it work

[Anonymous No.17593016]

>>17588395
>I don't see how a robot could believe in god, and thus, I don't see how a robot could believe in higher art.
This is your brain on humanism

[Anonymous No.17593029]

>>17592989
Atheists actually believe this sincerely?

[Anonymous No.17593038]

>>17593029
yes

[Anonymous No.17593056]

>>17588018
why are people brushing this off, this already much better than the average person could do

[Anonymous No.17593246]

>>17593029
It's epistemological. Over a text medium such as this board, you have no idea who/what else is posting. The only way you know we probably aren't bots is your pattern recognition. Once bots become able to give similar enough responses, you will have no way of knowing whether you're the only human on the board. What if you are discussing religion and the robot strongly espouses its belief in God? I could be convinced the robot believes more strongly than a lot of the pretenders at my church. There are entire families of people all pretending to believe just to fit in and not disappoint each other.

[Anonymous No.17593258]

>>17593056
It beats a large part of academic leftists for sure. We can automate the writing of leftist word salads now.

[Anonymous No.17593264]

>>17593246
Bots are already able to do this kind of thing. For 95% of the post here, they could have been written by bots for all I know.

[Anonymous No.17593296]

>>17588362
It's trained on hundreds of thousands of webpages and articles. It's ideas will only venture in directions written about by other humans. It's literally impossible for it to be novel.

[Anonymous No.17593335]

>>17588018
>didn't read the first part
>thought 'yeah this makes sense'
>tfw I check the reply and then read the first part
>tfw AI-generated answers are indistinguishable from pseudposts
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAa

[Anonymous No.17593495]

>>17593296
STRONG disagree here! The AI isn’t human, and we are...that means I the ideas it comes up with can be way outside standard thinking. One of the things I find it does best is assist in coming up with ideas.

40% are derivative
30% make no sense
20% are right but impractical
10% are pure fucking gold

For example, asked it to come up with some Squirrel themed Valentine’s Day cards. It generated about 100, but I kept the top 10:
>1. You mean the sq-world to me.
>2. I want someone to hold me like a squirrel in your arms.
>3. You make my heart feel like a small animal being eaten by a really huge bird.
>4. You are the shell to my nut.
>5. I wish I could give you all my fallen acorns; that’d be a good Valentines Day gift.
>6. You make my bushy tail flick out of control.
>7. You’re my oak leaf.
>8. When I looked into your eyes I feel like a squirrel looking at an oncoming car
>9. It’s impossible for a to tell you how much I love you; let’s just say, I love you a bushel and a peck.
>10. The depths of my affection are downright creepy.

Not saying it’s better or worse than I could come up with...but it certainly is different! I’m sure writers will be able to use AI to interrogate their ideas, or propose 50 endings that may inspire them to right the 51st...

[Anonymous No.17593612]

>>17593495
That is surprisingly funny. Reminds me of Tay AI making the zodiac killer joke. That shit was genuinely funny and surprising.
I think I agree with you that 10% is gold but if I were to continue playing devil's advocate I'd say something about how it's possible that some human had made most if not all of those jokes somewhere in the data that the AI was fed. But the dataset is so huge it would be impractical to search for it.

[Anonymous No.17593617]

>>17593495
The sq-world pun is genuinely impressive

[Anonymous No.17593650]

>>17593495
>>You mean the sq-world to me.
Kek

[Anonymous No.17594117]

>>17593612

Okay, let’s find something no one has possibly done before...two anons give me names of some two figures or fictional characters.

>[character1] and [character2] had three children:

We’ll see what it comes up with (I tried Count Chocula and the Babadook before and the results were fantastic)

[Anonymous No.17594139]

>>17587188
Dreadfully boring and meaningless.

[Anonymous No.17594214]

>>17594139
Sure, why not...


>Dreadfully, Boring and Meaningless had three children: A girl named Change, a girl named Theater, and a boy named Addiction. Change got married and had a child. It was brought up in a cult, and its name was Question. She hated her mother, and performed terrible acts. One day, she decided to commit suicide by jumping off a cliff. As she fell, she prayed to god.

>God answered, but upon hearing her name, blushed in embarrassment. She landed on her head and died. This is a painless and quick form of death. I would recommend it to anyone.

>Theater (unlike her name) hated being seen on stage. She was a librarian in a small town in Massachusetts. She had a horrible sense of direction, as a result of having no eyes.

>Addiction was addicted to cocaine. All the time. He was incredibly short, and also wore a toupee. It was an exceptionally large afro, it seemed. He was married to a woman who had been dead for years, and had a son that was half dead too. The boy had a large tumor on his back. It seemed to be growing.

That looks pretty original to me...

[Anonymous No.17594217]

art is evolutionary warfare. to do "good" art you basically need to act at a point in time, at a certain place and you need to do it with a kind of temperament that feels daring, eksiting and promising for that particular moment
for ai to do good art it doesn't have to replicate literature, it has to replicate human evolution and the human mind in total, providing us with the kind of cultural material that elevates us in this specific historical contekts

[Anonymous No.17594247]

>>17594214
This is AI generated?

[Anonymous No.17594264]

>>17594217
Yup, “good” is whatever ppl at a particular time think it is. “Art” is a medium to move people or make them see the world differently. I think AI nails “art”, jury’s out on “good”...

Also your affectation around using the letter ‘x’ is cringey as fuck. Stop.

[Anonymous No.17594273]

>>17594247
Yup. I can’t come up with shit that fast (on my phone at the moment)...

[Anonymous No.17594293]

>>17594273
See that's why I said it'd be dreadfully boring and meaningless. Perhaps this hyper-autism could be interesting in the post-modern interpretative sense where you nitpick texts for the meaning the author tried to infuse them with, but AI can not infuse meaning into anything. The very concept makes no sense to it. The best AI can do is form without substance.

[Anonymous No.17594313]

>>17594293
>post-modern interpretative sense where you nitpick texts for the meaning the author tried to infuse them with,
you double wat

[Anonymous No.17594331]

>>17594313
what

[Anonymous No.17594404]

>>17594264
people are seriously misunderstanding the point which ai is at. it's not impressive, it's legible but there's not much going on, mentally speaking

the "point" of art is to simulate a temperament which makes the world come together in a pleasing way, freshness is a part of that equation. zooming in on the madeleine cake as proust did, is fresh. ai can remiks that story with another pastry but it's not the same thing, the train has already left the station. constantly building and rebuilding your models is a part of art and it's done with a certain purpose in mind. changing the world in some way. ai would need a comprehensive idea of how it wants to change the world before it could create cutting edge art

sorry about the x, my keyboard is broken

[Anonymous No.17594405]

>>17588018
If someone were to use this AI regularly to post in random threads, do you guys think it would be noticed? What if we get to a point where this becomes a regular thing? We would have to collectively play “hunt the AI” games.

[Anonymous No.17594513]

>>17594404
It doesn’t matter what Proust had in his head when he wrote Remembrance...what made it powerful was it’s effect on you.

The thing that makes AI interesting to me isn’t the fact that it speaks English (pretty passably), but it’s that it ‘thinks’ in a way completely alien to us.
When AI learned to play chess, it made everyone better at chess, since they could play against a strong opponent anytime they liked. Imagine writing a paper and being able to bounce ideas off of an endlessly patient and curious AI as you write. Even if the AI didn’t draft a word, it will change the course of that discussion.

>>17594405
That world exists, check out Twitter.
Good thing 4ch doesn’t have anything worth wasting a bot on, y’all can troll each other just fine without help.

[Anonymous No.17594548]

>>17594405
The funny thing is that they'd be ten times better than the usual posts so yeah it'd be immediately noticeable, most of the time especially the /pol/ baiters genuinely read like if they were AI. Like maybe if you made a shitpost AI then we wouldn't notice them as easily

[Anonymous No.17594574]

>>17594548
Like this thread for example >>17594345 might as well had been made by a fucking bot. Now that I think about it discussion with a super knowledgeable AI would be comfy, the board's quality would actually improve a lot

[Anonymous No.17594604]

>>17594548
Idk about /pol/ but I’m sure there are boards or threads in which this quality of reply could fall under the radar quite easily. Maybe a little probing could figure it out, but then again, maybe not

[Anonymous No.17594610]

>>17594574
But I get what you mean, it would be very suspicious and would require a special kind of AI to camouflage itself there

[Anonymous No.17594637]

>>17594513
>what made it powerful was it’s effect on you.
I get that and I don't think there's any magic prohibiting ai from producing great work but it's simply not there yet. I wont worry about art when it is *there* because then we'll be entering a new era for humanity

>Imagine writing a paper and being able to bounce ideas off of an endlessly patient and curious AI as you write.
it's a tool, much like randomizers for music producers. I'm not sure that stuff revolutionized music but I guess it's nice

[Anonymous No.17594655]

>>17588362
Christ, this one is really scary

[Anonymous No.17594658]

AI can barely handle talking to humans without breaking apart; I'm not convinced it will be able to create literature. People who do believe it will create "literature" typically have an absurd understanding of what literature is - usually something like "something that just happens to read a lot at a particular time:.

A hundred AIs can write a 1,000,000,000 novels, but not one of them would be Bleak House. Even if they were to write an iteration that resembled Bleak House, you would still require Charles Dickens to recognize it as such.

[Anonymous No.17594665]

>>17594548
Maybe the alphaglobros already have shitpost AIs botting the net

[Anonymous No.17594672]

>>17594513
>>17594637
Bouncing back ideas sounds nice and all, but it’s gonna end up in a world where in any given type of text half is written by AI or at least has its ideas generated by them

[Anonymous No.17594677]

>>17594637
I’m getting major ‘of what use is a newborn baby’ vibe...
AI is the most interesting before it’s ‘there’...the accidents and misunderstandings and leaps from year to year make it fascinating to see.

[Anonymous No.17594789]

>>17594117
Lieutenant Thomas Glahn from Hamsun's Pan.

[Anonymous No.17594865]

>>17594789
Okay, need one more...

[Anonymous No.17594940]

>>17594117
And Frederick Barbarossa if no one else wants to answer.

[Anonymous No.17594943]

>>17594865
Remidios the Beauty from 100 years of solitude

[Anonymous No.17595028]

>>17594677
I'm not going to piss on anyone's parade and I don't profess to know a whole lot about the writing process but I have one thing that stands out in my mind. the good stuff I've written has come out in complete chunks, with the sensation of what's being produced being almost like an object that I could touch. there's something powerful about it, not 'interesting' or 'amusing' or a quaint oddity. this is not me preventing a script writer from using new technological tools in a brainstorming session, this is just me doubting the level you're going to reach with ai assistance. not because I doubt technology but because I doubt the temperament and mental rhythm of what's being produced in that scenario

[Anonymous No.17595038]

>>17594789
>>17594943


>Thomas Glahn and Remidios the Beauty had three children: a boy, Johannes, and two twin girls, Charlotte and Lucia.

>Johannes took after his father, equal parts honorable and obstinate; his elder twin girls were beautiful, like their mother, and were well cared for.
>Johannes served, happily enough, in the Swedish Navy, but he could never get used to war. He finished his service after only 3 years, and moved to Denmark to start a new life. No one has heard from him since, and it is believed that he has died.

>It is said that after Johannes left, Remidios spent her entire energies on her two girls. Over time, she became obsessed with their beauty: she dressed them different but alike, got them educated in the same manner and made them practice the same activities. If it was painting, Charlotte and Lucia painted together; if it was music, they played on the piano together; if it was dancing, they danced together. It was like this for everything. In their bed, each had to have a copy of the same book (both in Swedish and English), and their clothing had to be exactly the same.

>Charlotte was to marry Jofred, a fishmonger in Copenhagen; Lucia was to marry Gaston de Costos, a French nobleman and future Marquis of St. Ouen; and Johannes was engaged to marry Ester, a rich widow of Danish birth, before his disappearance. Thus were the children of Thomas Glahn to be settled.

[Anonymous No.17595119]

>>17594789
>>17594940

> Frederick Barbarossa and Lieutenant Thomas Glahn had three children: two sons, Frederico, and Conrad Glahn; and one daughter, Sofia Glahn. The family lived in the Grand City of Antioch.

>Frederick, convinced that the loss of the Holy Land was his fault, was an absent father. Without his affection, Frederico and Conrad Glahn grew up to hate the Holy Roman Empire, which had begun to fade over the centuries. Sofia Glahn, however, born in the Holy Land, was an extremely loving and kind girl, so much so that she was the only one Frederick could call his favorite. He often included Sofia on his conferences, and would read Arabic books with her.

>"That's how I learned to speak Arabic," Sofia told her brothers.

>As a man, Frederico strives to be all that his father was not, and his grief and anger, which had been festering inside him from time to time, were now in full swing. Cries of "Long live the Glahn house!" could now be heard from even the most pacifistic peasants.

>Conrad followed scholarly pursuits. His goal was to find a way to re-instile the Holy Roman Empire. He didn't seem interested in offering his services at the Battle of Antioch, but he was sitting on that hill, watching. The two brothers didn't seem to get along, but that was to be expected.

>The three children were growing older. Sofia Glahn was now in her early twenties, and looked very much like her father. She had his smile and his bright eyes. She was very intelligent, and had a love for reading and history, adored nature, and enjoyed poetry. The Glahn's were a remarkable family, and despite everything, Conrad Glahn, in particular, still loved his father, even if he didn't love him anymore. He loved his mother, too, although she was not even related to him.

This one goes off the rails a bit...lmao at the last line!
As a reminder, I’m not posting these to show that AI is perfect...only to show it can inspire creativity/originality.

[Anonymous No.17595552]

>>17595038
>>17595119
This isn't very high-quality writing and it is quite derivative but it's deeply disconcerting how far these kinds of programs have advanced in even just the past year.

[Anonymous No.17596408]

>>17594293
Now THIS response sounds AI generated.