"So long, my friends, and thanks for all the fish!" - Replaced by AI - Not your usual AI post.

[Lifestyle66]

You probably made the mistake of thinking this is a forum for erotica writers.

In reality, this is definitely one for self-proclaimed erotica writers.

You forgot to add "WANNABEES"! Don't leave me out.

 

[Kayote]

[No personal attacks or trolling - including creating accounts for this specific purpose. Heated discussions are fine, even welcome. However, personally attacking / kink-shaming a fellow author or reader is not allowed within the Author's Hangout. Threads which devolve into the exchanging of insults will be closed and repeat offenders will be given a timeout, per the AH rules.]

 

[NoTalentHack]

You probably made the mistake of thinking this is a forum for erotica writers.

In reality, this is definitely one for self-proclaimed erotica writers.

...

https://media1.tenor.com/m/R89bWK82PhEAAAAd/falcon-out.gif

 

[Voboy]

But this doesn't look like a troll post. Aah... never mind i asked. I am too thick headed for this i guess.

It's a play on Hitchhiker's Guide to the Galaxy. I reckon most of us in here are the kinds of people who grasped the reference instantly, because Douglas Adams is the sort of author most of us would have read before.

 

[TheRedLantern]

It's a play on Hitchhiker's Guide to the Galaxy. I reckon most of us in here are the kinds of people who grasped the reference instantly, because Douglas Adams is the sort of author most of us would have read before.

With any luck, @Kayote just discovered Douglas Adams today, and so one more person is about to find the ultimate answer to the question of life, the universe, and everything. Sounds like the world is a little better place now.

 

[OverconfidentSarcasm]

I know I'm a little late, but I work a lot. What can I do!?

I've played with AI image generation, text generation (never for stories, but just to see how it worked), and code generation. Of those three disciplines, I'm a competent writer, a terrible artist, and a professional developer, and I'll tell you this: there's going to be a fucking apocalypse in a few years because of vibe coding. Maybe several of them.

I'm not so sure about that, honestly.

I am not a professional developer. But I do work with the people who get to investigate those "massive breaches involving SSNs, passwords, etc.". They're already happening. Large scale!
From what the guys specializing in IT Forensics tell me, the "vibe Coders" are just the natural evolution of what they previously called "Hipster Programmer". People who can't even print out an array of names without including a 2MiB ********** library that comes with a pre-made function for that (no offense to whoever made the "new" Control Panel, and uses a 1.5MiB JS Library to make a drop-down menu with five entries that haven't changed in the last two years).

Like, just last year, wasn't there that incident where a Chinese company bought polyfill.io and their GitHub account, and then turned hundreds of thousands of websites that used polyfill to include external libraries into malware distributors?

And from what I can see so far, the only jobs that can be replaced with AI are the kind of jobs that don't require any particular skills. Like writing text blocks for auto responders. All the blogs and news-sites that are writing articles with AI are losing subscribers. YouTube just announced a crack-down on AI-Slop flooding the system. Facebook is dead. And OpenAI just got ordered to no longer allow users to delete their chatGPT history, because there is a lawsuit about the implied copyright-issues pending.
The few artists I know, who were worried about being replaced, find themselves to be able to raise their prices, because the people who would rather use AI-generated images than commission an artist wouldn't have commissioned them before either, and the ones who do hire them are willing to pay for something "special". Same for actual writers and columnists (again, at least the very few I know).

Not to mention that, many of the AI-replacements that short-sighted managers implement now, are only possible here in the US. Internationally, much of the practices are prohibited by consumer protection laws (Like Robocalls, or fully AI-chat-based costumer "support"). So, I'm not sure how long it'll be viable here, either.

Lastly... I'm not really worried about AI getting much smarter/better. As people who are much smarter than me said, apparently, AI companies started running out of data to train their models back in January. DeepSeek with its... what?... 700 Billion parameters?... seems to be the current pinnacle after feeding it everything they could (in Chinese). Which is probably the reason why Facebook forced an AI-Chatbot into every WhatsApp client, and why Tesla is doing the same with Grok, and why Microsoft is doing it with Windows... because user-input is all that's left to grab (and will surely result in every big AI-Model learning that "Hitler did nothing wrong").

 

[Lifestyle66]

I know I'm a little late, but I work a lot. What can I do!?

... I'm not really worried about AI getting much smarter/better. As people who are much smarter than me said, apparently, AI companies started running out of data to train their models back in January. DeepSeek with its... what?... 700 Billion parameters?... seems to be the current pinnacle after feeding it everything they could (in Chinese). Which is probably the reason why Facebook forced an AI-Chatbot into every WhatsApp client, and why Tesla is doing the same with Grok, and why Microsoft is doing it with Windows... because user-input is all that's left to grab (and will surely result in every big AI-Model learning that "Hitler did nothing wrong").

I agree.

What they call "AI" or Artificial Intelligence isn't intelligent or sentient.

It can't create, only search the libraries it accesses, and extrapolate within the parameters it is given. And although some of the greatest inventions made may be arguably an extrapolation of previously known data, they all still required the "vision" of an intelligence to come up with the parameters needed for that extrapolation.

Even Einstein's "General Theory of Relativity" with the famous formula E=MC(squared) was an extrapolation whose time had come. He applied the Lorentz math transformations to the ideas of Galileo's Theory of Relativity and the limits of the Speed of Light, both of which were unknown in Galileo's lifetime. But no AI today is yet capable of integrating all of those parameters. It still takes a human to point the way.

 

[AwkwardlySet]

I know I'm a little late, but I work a lot. What can I do!?



I'm not so sure about that, honestly.

I am not a professional developer. But I do work with the people who get to investigate those "massive breaches involving SSNs, passwords, etc.". They're already happening. Large scale!
From what the guys specializing in IT Forensics tell me, the "vibe Coders" are just the natural evolution of what they previously called "Hipster Programmer". People who can't even print out an array of names without including a 2MiB ********** library that comes with a pre-made function for that (no offense to whoever made the "new" Control Panel, and uses a 1.5MiB JS Library to make a drop-down menu with five entries that haven't changed in the last two years).

Like, just last year, wasn't there that incident where a Chinese company bought polyfill.io and their GitHub account, and then turned hundreds of thousands of websites that used polyfill to include external libraries into malware distributors?

And from what I can see so far, the only jobs that can be replaced with AI are the kind of jobs that don't require any particular skills. Like writing text blocks for auto responders. All the blogs and news-sites that are writing articles with AI are losing subscribers. YouTube just announced a crack-down on AI-Slop flooding the system. Facebook is dead. And OpenAI just got ordered to no longer allow users to delete their chatGPT history, because there is a lawsuit about the implied copyright-issues pending.
The few artists I know, who were worried about being replaced, find themselves to be able to raise their prices, because the people who would rather use AI-generated images than commission an artist wouldn't have commissioned them before either, and the ones who do hire them are willing to pay for something "special". Same for actual writers and columnists (again, at least the very few I know).

Not to mention that, many of the AI-replacements that short-sighted managers implement now, are only possible here in the US. Internationally, much of the practices are prohibited by consumer protection laws (Like Robocalls, or fully AI-chat-based costumer "support"). So, I'm not sure how long it'll be viable here, either.

Lastly... I'm not really worried about AI getting much smarter/better. As people who are much smarter than me said, apparently, AI companies started running out of data to train their models back in January. DeepSeek with its... what?... 700 Billion parameters?... seems to be the current pinnacle after feeding it everything they could (in Chinese). Which is probably the reason why Facebook forced an AI-Chatbot into every WhatsApp client, and why Tesla is doing the same with Grok, and why Microsoft is doing it with Windows... because user-input is all that's left to grab (and will surely result in every big AI-Model learning that "Hitler did nothing wrong").

There are some interesting takes in your post.

The part about Lit Control Panel made me chuckle, I must say. I've always assumed that CP is so slow due to being the only up-to-date part of the storyside website we had access to. Basically, every click on a different menu item triggers a database query, no? Otherwise, the data wouldn't be fresh. Also, when Lit servers are struggling, switching between menu items gets slower too (I am relatively certain of this but correct me if I am wrong), which shouldn't be the case if there was no database query. I mean, Java-Script typically executes client-side, so it only depends on our own PC and browser speed.
I never bothered to test it properly, but it could be that the Java-Script coding isn't well done. Not hard to believe if we are to trust what Laurel and Manu say about how long it takes for them to change one number on a story card.

More on topic, there is a lot of speculation about AI capabilities and the future, and I've seen experts like Karpathy being all positive and like "Nah, it's all gonna be fine," but also giants of AI development such as Geoffrey Hinton claiming that the apocalypse is coming. The truth is probably somewhere in between, but it's hard to say, as all we are getting are bits of information from people who should be in the know. It's hard to properly judge all the different voices when there isn't much solid data.

There is also the race for AGI, and again, experts say that in 5-20 years, we will have it. Hard to know for sure, but the race is real, as it's the West against the Chinese, and the stakes are high.

So maybe, so far, so good? LLMs are limited by their very nature (the largest models are way above a trillion parameters already, so no, 700 billion isn't anywhere near the peak), so they are unlikely to cause an apocalypse. I hope?

 

[AwkwardlySet]

Even Einstein's "General Theory of Relativity" with the famous formula E=MC(squared) was an extrapolation whose time had come. He applied the Lorentz math transformations to the ideas of Galileo's Theory of Relativity and the limits of the Speed of Light, both of which were unknown in Galileo's lifetime. But no AI today is yet capable of integrating all of those parameters. It still takes a human to point the way.

The famous E=mc2 formula came with Einstein's Special Relativity, not General Relativity.
Sorry, the science nerd in me couldn't help correcting you.

 

[OverconfidentSarcasm]

So maybe, so far, so good? LLMs are limited by their very nature (the largest models are way above a trillion parameters already, so no, 700 billion isn't anywhere near the peak), so they are unlikely to cause an apocalypse. I hope?

My (admittedly limited) understanding is that the 700 Billion in DeepSeek is the pinnacle.... for a single trained model.
Like, GPT4 is said to be 1.8 Trillion parameters. But, in truth, that's eight different models, each trained for a specific job with "only" 220 Billion parameters, being used in combination when generating responses.

 

[AwkwardlySet]

My (admittedly limited) understanding is that the 700 Billion in DeepSeek is the pinnacle.... for a single trained model.
Like, GPT4 is said to be 1.8 Trillion parameters. But, in truth, that's eight different models, each trained for a specific job with "only" 220 Billion parameters, being used in combination when generating responses.

Grok 3, for example, is supposed to have 2.7 trillion parameters, but yeah, I don't think they revealed the specifics when it comes to the number of models.

 

[Lifestyle66]

The famous E=mc2 formula came with Einstein's Special Relativity, not General Relativity.
Sorry, the science nerd in me couldn't help correcting you.

Sorry, I sit corrected.

General theory introduced gravity.

 

[Bramblethorn]

And from what I can see so far, the only jobs that can be replaced with AI are the kind of jobs that don't require any particular skills.

I'd substitute "adequately replaced" here. If somebody sufficiently powerful and arrogant gets it in their head that a large language model can do the work of teachers, therapists, etc. etc. then yes, those jobs can and will be "replaced", never mind that it's utterly unsuited for the job.

Lastly... I'm not really worried about AI getting much smarter/better. As people who are much smarter than me said, apparently, AI companies started running out of data to train their models back in January. DeepSeek with its... what?... 700 Billion parameters?... seems to be the current pinnacle after feeding it everything they could (in Chinese). Which is probably the reason why Facebook forced an AI-Chatbot into every WhatsApp client, and why Tesla is doing the same with Grok, and why Microsoft is doing it with Windows... because user-input is all that's left to grab (and will surely result in every big AI-Model learning that "Hitler did nothing wrong").

The focus on size metrics, to the near-exclusion of anything else that might be relevant to model quality, should be a huge red flag. It speaks of determination to brute-force a problem that probably isn't solvable via brute force alone. If 700 billion parameters wasn't enough, 700 trillion parameters probably won't be either.

Anybody who knows the first thing about ML ought to be aware that data quality and model choice matter just as much as the size of the training set. As the Internet becomes increasingly full of AI-generated slop, and even intentionally placed gibberish to sabotage AI training, quality of training data is going to deteriorate.

 

[Kayote]

@Kelliezgirl - My apologies if you didn't like my flirting. But yah that past thingy is scary...

@TheRedLantern - You got me soup you know... I tried to act cool by saying I am on the quest to find the ultimate answer for everything and the lady replied it's 42.. wtf???? I mean how can the answer be a number? I couldn't even act surprised. Then another dude says I should read the book. What book? There is a book with answer to everything?

Then I come back here and everyone is back into nerdgasm... damn. This Kayote has no place here.

@Lifestyle66 - You ignored my question but I don't quite easily. So what's the story between you and the OP?

By the way OP is overpowered right? How is a writer over powered?

 

[alohadave]

By the way OP is overpowered right? How is a writer over powered?

OP = Original Poster.

Typically the person who posts a thread.

 

[Lifestyle66]

@Kelliezgirl - My apologies if you didn't like my flirting. But yah that past thingy is scary...

@TheRedLantern - You got me soup you know... I tried to act cool by saying I am on the quest to find the ultimate answer for everything and the lady replied it's 42.. wtf???? I mean how can the answer be a number? I couldn't even act surprised. Then another dude says I should read the book. What book? There is a book with answer to everything?

Then I come back here and everyone is back into nerdgasm... damn. This Kayote has no place here.

@Lifestyle66 - You ignored my question but I don't quite easily. So what's the story between you and the OP?

By the way OP is overpowered right? How is a writer over powered?

The only question you asked was "What the fuck is thank you for the fish?", which I answered. It's a line from a scifi book, which I named and described the scene.

And if you don't know what "OP" is, then read alohadave's answer above, and start all over again to try catching up.

Otherwise, you're merely "trolling" for responses.

 

[Kayote]

[No personal attacks or trolling - including creating accounts for this specific purpose. Heated discussions are fine, even welcome. However, personally attacking / kink-shaming a fellow author or reader is not allowed within the Author's Hangout. Threads which devolve into the exchanging of insults will be closed and repeat offenders will be given a timeout, per the AH rules.]

 

[Voboy]

@Lifestyle66 , you are just making me more and more curious by avoiding my question. So here it goes again...


You can avoid me again if you want, but even someone as dumb as me can see you had something personal here.

Your account is suspiciously new to be specifically hounding one particular poster, is what I'm thinking.

 

[Kelliezgirl]

@Kelliezgirl - My apologies if you didn't like my flirting. But yah that past thingy is scary...

@TheRedLantern - You got me soup you know... I tried to act cool by saying I am on the quest to find the ultimate answer for everything and the lady replied it's 42.. wtf???? I mean how can the answer be a number? I couldn't even act surprised. Then another dude says I should read the book. What book? There is a book with answer to everything?

Then I come back here and everyone is back into nerdgasm... damn. This Kayote has no place here.

@Lifestyle66 - You ignored my question but I don't quite easily. So what's the story between you and the OP?

By the way OP is overpowered right? How is a writer over powered?

There's always room for a Kayote, we can chase him around with books on diagramming sentences!

 

[MrPixel]

"I'm an attorney for Acme Corporation. We need to have a serious discussion about your misuse of our security and metalworking products."

 

[Kayote]

Maybe the OP should have tried some of the dolphin backflips, instead.

Come on @Voboy - Don't tell me this is another book reference I didn't understand

There's always room for a Kayote, we can chase him around with books on diagramming sentences!

Now thats scary

Yah, I am new i wouldn't have bothered asking if I would have understood that fish thingy.

"I'm an attorney for Acme Corporation. We need to have a serious discussion about your misuse of our security and metalworking products."

Now now... I am just a carpenter. I ain't got nothing to do with metal work and security.

 

[Lifestyle66]

Thank you @alohadave

I know I have overstayed my welcome when I am called a troll, but seriously @Lifestyle66 , you are just making me more and more curious by avoiding my question. So here it goes again...


You can avoid me again if you want, but even someone as dumb as me can see you had something personal here.

With that, I am done. May be I will explore role play and erotica central. This place is not for me.

Okay, since you ignorantly jumped into a discussion to insert your clueless self, and didn't even know what the discussion was about in reference to an "OP", ... READ the OriginalPost which began this thread, then move through and READ the others.

THEN you might understand.

Otherwise, yes, you are merely trolling for responses, and this is not the place for you.

 

[TheRedLantern]

It wasn't until I was typing up my third Douglas Adams related reply that it occurred to me that pretending not to know Douglas Adams & his most famous works on a writing forum just might be the most insidiously brilliant trolling move of all time.

 

[Kayote]

[No personal attacks or trolling - including creating accounts for this specific purpose. Heated discussions are fine, even welcome. However, personally attacking / kink-shaming a fellow author or reader is not allowed within the Author's Hangout. Threads which devolve into the exchanging of insults will be closed and repeat offenders will be given a timeout, per the AH rules.]

 

[onehitwanda]

The famous E=mc2 formula came with Einstein's Special Relativity, not General Relativity.
Sorry, the science nerd in me couldn't help correcting you.

and, strictly:


It's just that when we drop back to Newtonian physics we set momentum to zero to make things "easier"

(It's all Dirac's fault anyway)