Some of those sound like something I'd actually like to read.
> I am, or was.
> There was once a man who lived for a very long time; perhaps three thousand years, or perhaps a thousand million years, maybe a trillion or so, depending on how the scientists look at it.
> “I am Eilie, and I am here to kill the world.”
> I have just been informed, that the debate over the question ‘is it right or wrong to have immortal souls’ has been finally brought to a conclusion.
> The purple-haired woman came to the clearing in the plain, and without looking up from her book, said, “It’s too late to be thinking about baby names.”
> The village of Pembrokeshire, in the county of Mersey, lies on a wide, happy plain, which, in a few years, was to become known as the “Land of the Endless Mountains.”
> I was playing with my dog, Mark the brown Labrador, and I had forgotten that I was also playing with a dead man.
> The black stone was aching from the rain.
> How many times have I had the misfortune to die?
> The first day I met my future self, I was aboard the old dirigible that lay in wait for me on the far side of the moon.
I could see a fascinating story starting with most of those lines.
> There was once a man who lived for a very long time; perhaps three thousand years, or perhaps a thousand million years, maybe a trillion or so, depending on how the scientists look at it.
I straight away thought of a Paul Erdos anecdote [1]:
"In 1970, I preached in Los Angeles on `my first two and a half billion years in mathematics.' When I was a child, the Earth was said to be two billion years old. Now scientists say it's four and a half billion. So that makes me two and a half billion. The students at the lecture drew a timeline that showed me riding a dinosaur. I was asked, `How were the dinosaurs?' Later, the right answer occurred to me: `You know, I don't remember, because an old man only remembers the very early years, and the dinosaurs were born yesterday, only a hundred million years ago.’”"
From the plot of the 1998 movie, Deep Impact: "MSNBC journalist Jenny Lerner investigates the sudden resignation of Secretary of the Treasury Alan Rittenhouse and his connection to "Ellie", supposedly a mistress. [...] she finds out that Ellie is really an acronym: "E.L.E." ("extinction-level event"). "
Most of this strikes me as mundane word salad, none of which can compare with human-composed entries to the Bulwer-Lytton Fiction contest[0]. It seeks "an atrocious opening sentence to a hypothetical bad novel."
This year's winner:
"Space Fleet Commander Brad Brad sat in silence, surrounded by a slowly dissipating cloud of smoke, maintaining the same forlorn frown that had been fixed upon his face since he’d accidentally destroyed the phenomenon known as time, thirteen inches ago."
Many years later, as he faced the firing squad, Colonel
Aureliano Buendía was to remember that distant afternoon
when his father took him to discover ice.
I’m an Edgar Finalist, Wall Street Journal best-selling author, blah, blah, and I can confidently say that...I’m doomed.
GPT-2 is amazing without even grasping deeper context. Once NN’s start to grasp arc and continuity (which is understanding changes over time) I’ll be looking for a new job.
GPT-2 is good at generating coherent original sentences based on a given corpus, but they don't really mean anything, which rapidly becomes clearer with each extra sentence. Ask it for an essay and you get a string of sentences with zero underlying point. The distance from here to writing original novels is very, very far. (Experimental poets, however, might need to worry!)
I'm not sure if you're actually disagreeing with me.
My point is once some future neural network starts to grasp continuity and context (something GPT-2 does not do) I think writing won't be safe from AI.
I couldn't speculate when that would be, but I've seen some unpublished efforts that make me think it's sooner than many people might think.
But between here and there will be the opportunity for AI-assistive writing that could facilitate what my developmental editors and beta-readers do.
I don't think it's at all guaranteed that future neural networks will grab continuity and context [in a meaningful sense, as opposed to basics like the character and place names being consistent and the context continuing to involve a murder and detectives] still less actually plot well. More importantly, for actual publishable writing as opposed to spam content AI has to do more than just start to grasp continuity and context, because every unpublishable human writer reaches that level, and most of them understand the much higher level of narrative arcs, characterisation and twists too. I don't think the cost of cash advances for no-name writers or lack of half decent unsolicited submissions to choose from is a big problem for publishers, is it? Surely 'time to fix not-irredeemably-flawed texts up enough to make them marketable' is in shorter supply...
Agreed that AI-assistive writing is pretty close to the point where it can save you the hassle of thinking up interesting ways of describing minor characters and locations, but of course when it reaches that level there are massive areas of content creation for AI companies which are more lucrative than novel writing and don't require further technological advances...
> My point is once some future neural network starts to grasp continuity and context (something GPT-2 does not do) I think writing won't be safe from AI.
If this is so, writers will more likely become editors.
In the same way, visuals artists might become curators.
As an optimist, I am still pondering how this could be a good thing.
My most recent series is The Naturalist. It's about a computational biologist that hunts serial killers. Even though the book goes into Bayesian statistics, Fourier analysis and has the occasional bit of Python, the series has been well received by the general public (it's Wall Street journal best seller and was optioned by a major studio.)
Diving into ML for research for the book lead me into actually doing my own projects and current AI startup.
I've read that book! I specifically logged in to say this. I liked the book!
As someone who wrote 6 unpublished novels, I'm envious of you that you got to publish that book (as in a story with a science touch)!
I specifically remember the scene where the protagonist is looking for bodies on the hill and he identifies the dead bodies acc to the plants that grew. It positively surprised me (that some ppl ala detectives actually are capable of that level of thinking) and gave me the creeps every time I go to the forest :p.
My books for rejected from every publisher and agent I sent it to and it is a duology about a "modern" dictatorship.
Part 1 is a mystery genre
Part 2 is a satire.
Sadly, I live in India and as per an Editor "Love and historical fiction sells in India and yours ain't that"
Nobody knows anything. Harry Potter got rejected by over 40 publishers.
I started as a self-published author and sold hundreds of thousands of books plus made my first movie deal before ever getting a publisher.
I was actually better early-off as an indie than with some of my first publishing deals. Although I learned a lot by working with a first-rate editor.
You don't need a publisher in 2019. You need a good book, a good editor, a good cover and a good description...and to keep writing.
A year after I sat down to write my first self-published novel I was sitting across the table from one of my favorite directors on the Disney lot talking about my book.
Inside baseball question, does it help you more if I buy the book, or if I use Kindle Unlimited to read it? What's the more beneficial business arrangement for you?
Whatever is convenient for you. I'm just happy to have readers. The most beneficial thing it is to leave a review if you like it and tell people what you enjoyed.
I remember when it was commonly thought that computer programs could never be creative. That the "spark" of creativity is impossible for a pure-logic machine to produce.
Interesting that that turned out to be such a simple problem.
This is still monkeys and typewriters, just the keys write entire words and dynamically change to a set of likely words based on some corpus of literature. And it's heavily editorialized.
I'm sorry to say that this perfectly describes human written novels. Particularly the part about heavy editorializing, most people wouldn't believe the sort of books that come out as self published without an editor.
"There was once a man who lived for a very long time; perhaps three thousand years, or perhaps a thousand million years, maybe a trillion or so, depending on how the scientists look at it."
I absolutely loved the opening “The first day of the world was born in the year 1985, in an old side of the world, and the air of the old sky of lemon and waves and berries.”
Just a tip. Reading these, select your favorite British actors voice and "hear" them reading it - dramatically. I mostly went for Patric Stewart. It was great. Try it.
"At the end of the world, where the tides burst upon the drowned, there exists a land of dragons, of dragons, which is the land of the dragons."
I'm intrigued at how this NN plays in the fuzzy line between magical surrealism and wordy incomprehensibility, it reminds me of the writings of many great writers...
I think we'll be hearing more interesting things from it in the coming years. I'll be excited to see more realistic generation, but sad to see less weird bonkers shit.
About GPT-2, is there any site like talktotransformer.com that can generate longer output? The length generated by that site leaves a lot to be desired.
> I am, or was.
> There was once a man who lived for a very long time; perhaps three thousand years, or perhaps a thousand million years, maybe a trillion or so, depending on how the scientists look at it.
> “I am Eilie, and I am here to kill the world.”
> I have just been informed, that the debate over the question ‘is it right or wrong to have immortal souls’ has been finally brought to a conclusion.
> The purple-haired woman came to the clearing in the plain, and without looking up from her book, said, “It’s too late to be thinking about baby names.”
> The village of Pembrokeshire, in the county of Mersey, lies on a wide, happy plain, which, in a few years, was to become known as the “Land of the Endless Mountains.”
> I was playing with my dog, Mark the brown Labrador, and I had forgotten that I was also playing with a dead man.
> The black stone was aching from the rain.
> How many times have I had the misfortune to die?
> The first day I met my future self, I was aboard the old dirigible that lay in wait for me on the far side of the moon.
I could see a fascinating story starting with most of those lines.