AI Fiction Update: The Machines Are Coming For Our Words.

[Theirl]

Well i wanted to read just to see the slop thr ai can make

 

[JayMark]

::raises hand::

untrained and self trained are not the same thing.

 

[DireBadger]

Nonsense. Humans derive from what came before just as much as any AI.
Every one of us is being trained from birth. The only difference is our training materials and a basic ability to use them.

The reason AI will fail on literature is because it's stupid. like TOR authors and the winners of the nebula awards for the last decade.



Sad puppies only whined until they ran out of food, but Rabid puppies will live forever.

 

[Daydreamers]

Some people really underestimate AI in the comments, we are not that special eventually we will be cast aside, it reminds me of chess back in the 90's when they said no AI would ever be better than a human,
now professional chess players are learning from AI to the point where our touch on the game is completly lost, unless you are a beginner ofc but then is it really chess?

 

[CharlesEBrown]

Some people really underestimate AI in the comments, we are not that special eventually we will be cast aside, it reminds me of chess back in the 90's when they said no AI would ever be better than a human,
now professional chess players are learning from AI to the point where our touch on the game is completly lost, unless you are a beginner ofc but then is it really chess?

And ChatGPT was caught cheating against a computer opponent programmed only to play chess - it actually hacked into the code and copied it!

 

[beast_regards]

*sigh*

AI ...

As a fan of sci-fi and fantasy, I sometimes wonder whether the corporations read too much sci-fi, or too little. I suspect it is the latter.

Explaining bureaucrat a technology is famously difficult...

The current so-called "AI" is less a technology, and more Ponzi scheme with the 100% percent investment returns where companies couldn't wait to pour money into something which theoretically could make them money, but in reality it does nothing but hampers both customers, workers, and actual scientist, and would cost them more of their precious money once the bubble bursts. Unfortunately, the bubble didn't burst yet, and until it does, nothing would change. People would lose jobs, and large corporations would alienate both customers and employees.

 

[CharlesEBrown]

*sigh*

AI ...

As a fan of sci-fi and fantasy, I sometimes wonder whether the corporations read too much sci-fi, or too little. I suspect it is the latter.

Explaining bureaucrat a technology is famously difficult...

The current so-called "AI" is less a technology, and more Ponzi scheme with the 100% percent investment returns where companies couldn't wait to pour money into something which theoretically could make them money, but in reality it does nothing but hampers both customers, workers, and actual scientist, and would cost them more of their precious money once the bubble bursts. Unfortunately, the bubble didn't burst yet, and until it does, nothing would change. People would lose jobs, and large corporations would alienate both customers and employees.

Knowing someone who was at the admin assistant/programmer level at Microsoft when they abandoned their AI product (since ChatGPT hit the market and did everything they were working on, most of it better) you are not far off here - but are not absolutely correct either.
Current AI is meant as a tool, and those who understand it can use it to, well, save money (he is now using AI to make movies, something he'd wanted to do with programming in general but landed the job at MS and put his dreams on hold for about four years); not to really MAKE money, but to cut down on the number of people needed to do rote tasks like animating images, cleaning up dialogue, managing schedules, and handling contact with other people working on the project.
It can't be "trusted" to do more than this, but can be used, with a lot of refinement, to do all of it, and far more cheaply and efficiently than conventional staff.

 

[beast_regards]

Knowing someone who was at the admin assistant/programmer level at Microsoft when they abandoned their AI product (since ChatGPT hit the market and did everything they were working on, most of it better) you are not far off here - but are not absolutely correct either.
Current AI is meant as a tool, and those who understand it can use it to, well, save money (he is now using AI to make movies, something he'd wanted to do with programming in general but landed the job at MS and put his dreams on hold for about four years); not to really MAKE money, but to cut down on the number of people needed to do rote tasks like animating images, cleaning up dialogue, managing schedules, and handling contact with other people working on the project.
It can't be "trusted" to do more than this, but can be used, with a lot of refinement, to do all of it, and far more cheaply and efficiently than conventional staff.

The "AI" needs to be fed a constant input of the new content created by the actual content creator, filtered down by the other actual people...

...actually, the worst you could do is to feed it its own content...

...and still it won't work at some form of expression, like writing, which requires the understanding which the "selective randomness generator" isn't able to provide...

...hence the need to call it "AI" to create the false impression it has the capability of understanding, which it doesn't, but it helps to draw in investors which were swayed by the prospect of the free creative workforce, likely after being told some imaginary capabilities from sci-fi, or even just being carried out by imagination of something he got by information osmosis (as I couldn't imagine the CEO being the fan of sci-fi)

To add to the mess, many features called "AI" have very little to do with each other, but are lumped together to ramp up the sales (and more likely, a stock prize)

If you wanted to use the image generator commercially (in other way than fraud promising returns) you need to hire an actual artist doing retouch, actual IT personnel, and a lot of them, and quite a lot of ordinary office workers, and even then, you would struggle to produce anything consistent. That tech may be good for us here, struggling to get the cover for the trash writing done in spare time, but for publishing companies and studios?

It's cheaper (for company) to hire the entire team of artists which would draw you what you want ...

 

[Erysion]

...actually, the worst you could do is to feed it its own content...

Keeping the bloodline pure, aren't we?

 

[beast_regards]

Keeping the bloodline pure, aren't we?

There is one thing the Habsburg dynasty and "AI" have in common: Inbreeding is the bad idea, yet everyone does it.

 

[CharlesEBrown]

...and still it won't work at some form of expression, like writing, which requires the understanding which the "selective randomness generator" isn't able to provide...

Right - it is a tool. Used carefully and well, it is a great tool; used the way about 90% of people are using it ... it is a very, very bad thing.

To add to the mess, many features called "AI" have very little to do with each other, but are lumped together to ramp up the sales (and more likely, a stock prize)

It is replacing (or becoming part of) the "smart-" prefix (smartphone, smarthouse, smartcar, smart-coffeemaker) as a sales tool rather than a real thing, it seems.

If you wanted to use the image generator commercially (in other way than fraud promising returns) you need to hire an actual artist doing retouch, actual IT personnel, and a lot of them, and quite a lot of ordinary office workers, and even then, you would struggle to produce anything consistent. That tech may be good for us here, struggling to get the cover for the trash writing done in spare time, but for publishing companies and studios?

It's cheaper (for company) to hire the entire team of artists which would draw you what you want ...

Not according to this guy who is making a living using it (not as much as he was making at Microsoft but enough to get by) - you do need a small team to clean up the output, but you just need a tiny core of one to four people who really understand how to write prompts to do the original work, and it takes days or weeks instead of the months or years it used to, at least for the kind of animation he's getting into.
Well, you have the team design the characters and sets first, then have the AI insert the characters into the sets and handle the animation, even some dialogue, and polish the writing, then have the team go back in and clean up anything the AI really messed up.

 

[beast_regards]

Right - it is a tool. Used carefully and well, it is a great tool; used the way about 90% of people are using it ... it is a very, very bad thing.

Using it how 90% of the people are using it is not a problem.

Those 90% people are random people on the Internet, which would end up posting the random things with or without so called "AI" ...

Not according to this guy who is making a living using it (not as much as he was making at Microsoft but enough to get by) - you do need a small team to clean up the output, but you just need a tiny core of one to four people who really understand how to write prompts to do the original work, and it takes days or weeks instead of the months or years it used to, at least for the kind of animation he's getting into.
Well, you have the team design the characters and sets first, then have the AI insert the characters into the sets and handle the animation, even some dialogue, and polish the writing, then have the team go back in and clean up anything the AI really messed up.

What kind of animation?

 

[CharlesEBrown]

What kind of animation?

I suspect he's doing advertising given that he's already making money but I did not ask.

 

[DireBadger]

current gen (and next Gen) AI writing stories is like buying an electric razor and expecting it to build your house for you.

 

[dukerino]

One reason AI writing has proliferated--despite the public largely being ambivalent about it at best--is that it always sounds like a linkedin post. Which impresses the c-suite weirdos and ad guys who steer the corporate ships.

 

[Deleted member 266]

OpenAI declares AI race “over” if training on copyrighted works isn’t fair use

National security hinges on unfettered access to AI training data, OpenAI says.

arstechnica.com

 

[Deleted member 84247]

OpenAI declares AI race “over” if training on copyrighted works isn’t fair use

National security hinges on unfettered access to AI training data, OpenAI says.

arstechnica.com

Classic. Reminds me of the fear mongering used to justify all of the government spying with the patriot act.

 

[EverenVale]

I played with ChatGPT a lot because it was one of my biggest concerns that it would get creative and write well-deserved stories.
But it showed me time and time that it is not creative as a human mind and certainly, it doesn't understand the different levels of human emotions.
Most of my stories are based on their inner thoughts and deep psychological emotions, which AI has no context about. It has all the definition but doesn't know how they apply in different situations.
Most of the stories that AI writes are surfaced and user-driven.
So I don't see it coming for at least 10+ years.

 

[melchi]

There is no need to worry. @Ai-chan will only come for your words if they have traps in them. Because Ai-chan is the devourer of traps.

 

[Ai-chan]

There is no need to worry. @Ai-chan will only come for your words if they have traps in them. Because Ai-chan is the devourer of traps.

Yeah! So put in lots of word traps so that Ai-chan will devour you.

For example, say "Uuuh iihh, aah aaah, bing bang, walla walla bing bang" instead of "They fucked."