Based on the statements from those in control, which at this point in our society are far more draconian and anti-humanitarian than any time in our recent past, the following is more likely.

If ai is able to be raised to a level that it can sustain production regardless of whether it can do anything more than holding pattern, human labor will inevitably become mostly unnecessary.

Due to how those in power think, the vast majority of humanity becomes valueless and unnecessary, even those few jobs left will also be devalued likely below subsistence levels.

Humanity will become a burdensome irritation that is irrelevant and unnecessary, the vast majority of people will loose access to most of what is left of society including technology and ai, let alone the ability to feed, cloth or house oneself.

You will become a criminal in your own house and life will become extraordinarily hard for nearly everyone the world over.

Most of what I wrote above is quoted from very famous very powerful insiders, who currently are building bunkers and personal supply chains.
Me thinks you don’t do that if whatever it is your planning is a good idea that will benefit society.

Now whether we end up with an automated world full of self building and repairing robots mining, refining, building and farming in my lifespan is a whole other discussion but that is the goal, an economy that doesn’t need people.

Unless we have some type of enlightenment where we actually have leadership who truly love and value humanity
and remove all of the bums out of power (including non-governmental) we likely aren’t going to have a good time here, at least not for very long.

Selfish doesn’t begin to describe many in the power sphere currently.

Jo22 wrote on Yesterday, 10:48:

They maybe won't. Once they're there, AI/LLM development is many iterations further and has already outrun them. […]
Show full quote

For individuals in those companies, whatever field, i feel empathy and worry about what is means for all people if the stages of beginner to average ability are now all catered for by AI. how do they get to advanced stages?

They maybe won't. Once they're there, AI/LLM development is many iterations further and has already outrun them.

And the development is exponential rather than linear. 😟
People mention numbers such as 5 years, maybe because they look back and count how long development took in the years before.

But a danger is that AI/LLM optimizes itself in ever smaller time frames,
like a C compiler that can build itself/optimize itself each time.

But there's a catch, as well. AI/LLM is still improving by learning from original, human data sets.
But what happens if the human information pool is exhaused?

indeed. first i'm not sure present AI is an exponential growth in capability, when new tech emerges it often spends a while growing slowly, then grows explosively, then slows down in terms of capability (usually only growing again when other tech starts contributing to it)

in other words i'm sure LLMs from 5 years ago are much less impressive than now, but i'm not expecting the same leap forward in 5 years - more in capacity ; i.e faster, more pictures, longer videos, longer context retention etc rather than fundamentally "smarter"

but we will see. You're right too that training data risks - running out, consuming other llm outputs, reduced curation by humans as scale builds are all there too. think of an extreme - every media is generated, from that point on there is nothing truly new - just remixes for the past forever!

your point about websites is true too, several already noting less and less human traffic - what happens when they close? it locks the LLM into a fixed time again, with no new things (although i guess news outlets, and things that have updates and audiences will survive, but all smaller and edge case sites might not)

Jo22 wrote on Yesterday, 12:05:

I sat down at a desk 5 days a week to code and pay my bills, it wasn't for the love it it, and based on my experience - all big corps I admit - that describes the majority of devs. I wasn't paid to produce unique and groundbreaking code, I needed to get my changes checked in by Thursday or I'd miss the release.

Understandable. But that's a flaw or bad attribute of the modern business world.
People just take dull jobs to get by, rather than finding purpose or fulfillment in it. That's neither progress nor healthy.

hmm, i'm sure its true of many roles in life well before modern business, well before civilisation, were also dull jobs as such. Perhaps they had more meaning because they were directly linked to their tribe, their community and that is the difference not whether its business, state or whatever. what badmojo was saying about the nature of "normal work" was right, and in fact a cooperative society cannot wait for artistry or creativity all the time, there's things to be shipped soon.

further, artists and speculative creatives can only exist in a society that creates a surplus. the people who depend upon and benefit the most from having a majority doing ordinary jobs are the people who don't do ordinary jobs

for your more optimistic ideas, i'm not sure 🙁 AI is presently owned by a few, i cannot imagine a smooth transition to the optimistic ideas you have even if it were finally to get there

(edited some spellings a few times!)

Trashbytes wrote on 2026-10-14, 00:51:

I don't normally get involved in AI debates but ..

Should we also not be worried about humans who have read and trained from books/literature/media ? Is this not the exact same thing ? or does it coming from a human give it a free pass? Training data is essentially no different than the books in a library, do you consider the people using the books to be stealing the information they learn from them ?

Humans don't (or shouldn't) get a free pass for IP theft and plagiarism.

The AI was created, trained and prompted by a human to do what it does, it only ever creates what its told to create, it has no free will of its own and thus all of its output is no different than the output of a hammer and anvil used by a blacksmith.

The blacksmith doesn't train the forge, hammer and anvil to copy others' work, nor does it just do it all automatically and go through an iterative and regurgitative process.

I personally see AI as nothing more than a tool, it requires a human to prompt it and use it for it to create anything, as for its training data ...no different than how humans learn from the collective Knowledge and Art we have stolen over the centuries from each other.

AI (of the machine learning kind) could be curing cancer and ending world hunger, and we aren't opposed to such a use. But where it's used the most is with LLMs to make useless crap, including YouTube videos purely for turnkey monetization channels. And it does it at such a huge scale it's impossible to clamp it all down.

(Just visit your local Art gallery to see the human created slop or Steam for the human created slop games there too, or just check out Netflix which is peddling slop like its going extinct)

Art for art's sake. Art as expression and political speech created solely by a person for the consumption of other people. Not a machine turning some prompts and images of others' art to be overfitted into slop.

In some ways we're in a platinum age of television entertainment thanks to Netflix, Amazon Video, Apple TV and the rest. And yes, there's also a metric crapton of junk out there, but there's something out there for everyone to enjoy.

AI is nothing more than a tool and I have seen some truly amazing things be created with it, I am not going to pan something simply because how how it came to exist, if its good and does what it was created to do then great Ill use it, I don't care if it was created by a human or generated with the help of AI.

I've seen some sparse few good things made with it, like the "There, I ruined it" videos where it's used to map his singing voice to the voice of another artist. The creative, imaginative part is all him, it's just the part where he sings an impression of Hank Williams or Johnny Cash and the AI transforms it the rest of the way, for the purpose of parody and satire. But then there are the countless godawful (and godawfully smug) political memes. Or the junk videos of tractor trailer accidents where suddenly the tractor part changes from a 1979 Kenworth Aerodyne to a 2001 Peterbilt 362, and all the people going "wow" about how incredible the collision was because they don't know better.

As for regulation . .fuck no, we have far to much already and it why the world is in the mess its in. There is far too much corporate and government oversight and I for one do not want more of it, I want less.

And I want clean water and safe motorways. I want buildings that don't collapse or catch fire because there was no oversight. I want lettuce that won't make me 'double dragon' for hours and potentially die. I want airplanes to not fall out of the sky or have midair collisions.

criminals do not care about regulations or laws and will do as they please regardless

Maybe you should stay out of AI debates, and maybe out of most other debates. The argument that "criminals will break laws anyway" is such a non-starter and is intellectually dishonest.

jmarsh wrote on Yesterday, 07:04:

I guess you guys both missed the main point: that AI can't come up with anything truly new, it simply regurgitates. If that's all you think coding is then it's any wonder we're heading for the dark ages, where no new ideas will come about and the same "slop" is served up time and time again.

Correct, and I sense the common sentiment that AI is merely the latest link in the chain: original creative & critical thinking had already been in decline in [at least Western] culture for good number of years prior to AI making headlines, internetworked screens are literally everywhere with the addiction thereof actively turning people's minds to mush (do not think our enemies fail to appreciate this and exploit it to promote our decline), fewer are graduating with the essential means to be productive (let alone contributing citizens), and all the more we as a civilization would be rendered completely helpless if something happened that cut us off from the technological connections we have allowed ourselves to become so deeply reliant upon.

Pay attention to the game Mullet Madjack: it exudes a Tom Clancy caliber of speculation about where our society is headed if it stays this course for much longer.

badmojo wrote on Yesterday, 11:45:

This "AI bad" discussion always describes a complete replacement of whatever it touches but that's just not how humanity works. Yes it's being pushed at us from all sides at the moment but it'll be adopted where it's useful and ignored where it's not. Drum machines could have replaced all drummers but there's still plenty of drum kits being used out there. eBikes could replace all push bikes but they haven't and won't because a lot of people like the exercise.

Exactly: replacement is the one aspect of AI I am NOT concerned about (the only job recorded to have been completely eliminated by automation is elevator operator, and it took many decades for systems to became reliable enough to achieve that). I happen to work in a clerical capacity in which day-in-and-day-out I am cleaning up the messes left by AI bots - even IF they were not so STUPID, it would still be necessary to have human eyes checking the results. While I can never see myself entering the IT realm, I remember how a few years ago there was an attempted en masse replacement initiative that since then proved not to have been so good a business decision.

Seeing how computers are still dumb no matter how "smart" they get is one of the [albeit more minor] reasons I eschew using AI personally, though I really have nothing against applying it professionally so long as its deliverables gets scrutinized accordingly.

kolderman wrote on 2026-10-15, 02:23:

Never been a better time to be a retro gamer

THIS! There is a lot of AI slop but there is a lot of human made slop. I find myself going back to retro games, retro machines and older movies. Everything today is slop Ai or human made.

ixfd64 wrote on Yesterday, 02:29:

megatog615 wrote on 2026-10-14, 00:26:

AI code relies on stolen labor, just as how AI art does. People who are against AI art should advocate against AI code as well.

I'm going to have to respectfully disagree with this. I generally don't like AI art either because it diminishes human creativity and takes very little effort to make. AI-generated art is also often bland and emotionless.

However, AI-generated code is functional and not merely expressive. Using an AI agent to code can greatly increase productivity and help detect issues that humans may easily miss. So I don't feel the same way about AI-generated code as I do about graphics.

Goeoff on Craft computing channel on youtube did a review of an AI computer recently.
He is getting into AI as a way to make his otherwise useless server rack useful in a way other than playing games.

ANyway he did a review of a PC and set himself a task to use AI to help him fix a problem with a project he has going with a temp sensor he is working on.

The video is really REALLY worth watching, the results at the end are a real must see for everyone.
But dont skip to the end watch the whole thing to make the otherwise meaningless results make real sense.

As a hobbyist, I write code that "scratches an itch." When I decide to do something I buy books (often used, so I have to hunt them down), read online and try to understand the problem domain. I think it's important not just to cut and paste, but to build an understanding. Doing that takes a lot of time and energy, but it generally results in a better project. It also gives me an edge when trying to work around problems or fix my own bugs. And then I share my source code with copious comments and documentation to help other people learn.

As a professional software engineer I have 30+ years of coding history. Some of that coding was just drudgery and maintenance. Some of it was very new and novel, and I contributed to five software patents. (And quite a few more things should have been patented, but that's another story.) I've contributed to operating systems, hard drive firmware, and even a Top 500 supercomputer. Even though that was paid work, I still treated it the same way - I needed to understand the problem domain so that I could craft the best solution.

For me people using AI is kind of like somebody from the middle ages walking on a path and being given a magic box that lets them grow crops faster. They can harness the magic box and do things that look amazing to them, but they have no idea that the magic box probably just has a soil analyzer, some nutrient packets to fix deficiencies, an irrigation schedule, and it knows to let a field lay fallow every three years. They just ask for more wheat and it magically happens with no understanding of what is going on. For a slightly more modern example, people using AI is like Mickey Mouse in Fantasia finding the sorcerer's hat. It's slightly less bad than that because tokens still cost money and that limits the damage. ;-0

People are using AI to make themselves more productive and they are applying it to this hobby too. I value the process of learning and understanding, including the struggle that comes with it. When people skip that process and then claim they created something, they devalue all of the prior work that actually took effort before the magic boxes became available. Nobody can take my 20 year investment in mTCP away from me, but it is mildly irritating to see people using AI to create things almost instantaneously in comparison and then claiming credit for it. It's not the same. We're going to be flooded with AI generated software that the "authors" fail to maintain because they lost interest quickly after the initial sugar rush fades. There are examples of that here on Vogons already.

What AI isn't doing isn't comparable to people learning from books or cutting and pasting from Stack Exchange. AI is ingesting and regurgitating at a scale and speed far beyond any human or even group of humans. People have learned from each other since the beginning of time; AI is short circuiting the learning and understanding part of the process. Which makes me sad because people are going to lose the experience of struggling to understand and build things, which has been an integral part of the human experience.

LLMs are not intelligence. They are information, or knowledge.

The mark of an unintelligent human being is thinking intelligence simply means being well educated, or having read lots of books, that the intelligent man when required to do something simply calls upon his vast library of knowledge and applies it linearly to the task.

They are highly susceptible to "one weird trick" thinking. They think every problem is solvable as long as you have the one piece of "hidden knowledge" that simply needs to be discovered to solve the problem. And indeed, knowledge can sometimes be substituted for intelligence, as long as intelligence has already predefined the complete set of necessary knowledge to solve the problem. But knowledge can never generate more knowledge all on its own.

jmarsh wrote on Yesterday, 10:09:

Trashbytes wrote on Yesterday, 08:49:

Solution . .Stop buying slop and stop accepting slop as the bare minimum .. its not hard.

You say this, and yet in your post immediately before it claimed it was "hypocritical" to not accept AI regurgitated slop. At this point you're just being contrary rather than offering any insight or opinion.

Not even close to what I wrote in the prior post but cogratulations on the attempt to twist what I did write to support your narrative.

AlaricD wrote on Yesterday, 14:22:

Humans don't (or shouldn't) get a free pass for IP theft and plagiarism. […]
Show full quote

Trashbytes wrote on 2026-10-14, 00:51:

I don't normally get involved in AI debates but ..

Should we also not be worried about humans who have read and trained from books/literature/media ? Is this not the exact same thing ? or does it coming from a human give it a free pass? Training data is essentially no different than the books in a library, do you consider the people using the books to be stealing the information they learn from them ?

Humans don't (or shouldn't) get a free pass for IP theft and plagiarism.

The AI was created, trained and prompted by a human to do what it does, it only ever creates what its told to create, it has no free will of its own and thus all of its output is no different than the output of a hammer and anvil used by a blacksmith.

The blacksmith doesn't train the forge, hammer and anvil to copy others' work, nor does it just do it all automatically and go through an iterative and regurgitative process.

I personally see AI as nothing more than a tool, it requires a human to prompt it and use it for it to create anything, as for its training data ...no different than how humans learn from the collective Knowledge and Art we have stolen over the centuries from each other.

AI (of the machine learning kind) could be curing cancer and ending world hunger, and we aren't opposed to such a use. But where it's used the most is with LLMs to make useless crap, including YouTube videos purely for turnkey monetization channels. And it does it at such a huge scale it's impossible to clamp it all down.

(Just visit your local Art gallery to see the human created slop or Steam for the human created slop games there too, or just check out Netflix which is peddling slop like its going extinct)

Art for art's sake. Art as expression and political speech created solely by a person for the consumption of other people. Not a machine turning some prompts and images of others' art to be overfitted into slop.

In some ways we're in a platinum age of television entertainment thanks to Netflix, Amazon Video, Apple TV and the rest. And yes, there's also a metric crapton of junk out there, but there's something out there for everyone to enjoy.

AI is nothing more than a tool and I have seen some truly amazing things be created with it, I am not going to pan something simply because how how it came to exist, if its good and does what it was created to do then great Ill use it, I don't care if it was created by a human or generated with the help of AI.

I've seen some sparse few good things made with it, like the "There, I ruined it" videos where it's used to map his singing voice to the voice of another artist. The creative, imaginative part is all him, it's just the part where he sings an impression of Hank Williams or Johnny Cash and the AI transforms it the rest of the way, for the purpose of parody and satire. But then there are the countless godawful (and godawfully smug) political memes. Or the junk videos of tractor trailer accidents where suddenly the tractor part changes from a 1979 Kenworth Aerodyne to a 2001 Peterbilt 362, and all the people going "wow" about how incredible the collision was because they don't know better.

As for regulation . .fuck no, we have far to much already and it why the world is in the mess its in. There is far too much corporate and government oversight and I for one do not want more of it, I want less.

And I want clean water and safe motorways. I want buildings that don't collapse or catch fire because there was no oversight. I want lettuce that won't make me 'double dragon' for hours and potentially die. I want airplanes to not fall out of the sky or have midair collisions.

criminals do not care about regulations or laws and will do as they please regardless

Maybe you should stay out of AI debates, and maybe out of most other debates. The argument that "criminals will break laws anyway" is such a non-starter and is intellectually dishonest.

I will take your esteemed advice and thusly vacate this caterwauling.

AlaricD wrote on Yesterday, 14:22:

In some ways we're in a platinum age of television entertainment thanks to Netflix, Amazon Video, Apple TV and the rest. And yes, there's also a metric crapton of junk out there, but there's something out there for everyone to enjoy.

I've been feeling the opposite regarding TV lately. It's not that there aren't quality shows, but it's the shift to short seasons with prolonged production time between each season (2-3 years).

I find it harder to get into shows these days when a season is only 8 episodes and the next season is potentially years away.

Looks like the thread was reopened. Good. Now pertinent to the original topic, the situation has further developed. To wit, someone has tried reproducing Chrono Trigger with AI.

https://www.youtube.com/watch?v=OXCXNHiqdwc

Judging by the video, Claude has been trained (dare I say programmed?) to try to analyze the machine code of old game ROMS and even to package emulation code into the game executable. The creator apparently didn't put much effort into helping Claude along so it seems to have fumbled around a bit trying to discern what he wanted. To be sure the entire game engine -- especially not the battle system -- has not been fully replicated, but it's early yet. Retrogamers in particular should consider NOT patting themselves on the back I think, in that simple 8 and 16-bit 2D titles are unironically the easiest for an AI to copy: the theory and technique of their design is well documented, which is exactly the situation a generative AI needs to excel in a given task. The still-hot market for retro titles is exactly the prey an AI needs to find its commercial footing in a new medium.

Some would probably see this content explosion as a boon -- more choice, therefore better for the consumer. But, I urge you think about what the long-term ramifications will be in practice. Imagine a trend of producing titles that start out well designed and inviting but later on reflect increasing lack of attention to detail, to the point of the game literally falling into disorganization 20 hours in. Before you say "tell Claude to refine to fix", consider how difficult that might actually be in practice (because Claude, again, has only a passive idea of what a user wants from it unless it's given a full blown novel to work with). While you could argue that Nintendo-style beta testing is called for, it seems more reasonable to me to prohibit the practice of Claude-style monolithic generative game making outright from a commercial standpoint at least.

And I don't think quality issues are the only pertinent factor, either. What about the human element of game design? The designer's hope that their game will succeed because they dared to innovate, because they put time and effort into the quality of the experience. The hope that they would create something memorable. So long as games require effort, that is a reasonable expectation that outputs will remain sparse in number and highly creative, but if machines are just able to churn out a game willy nilly with only the vaguest of suggestions -- and especially if there is nothing stopping the machines from just copying outright what makes a given game interesting or unique -- then sooner or later the entire situation will not only saturate but degenerate into toxicity and collapse.

tcaud wrote on Today, 21:35:

...then sooner or later the entire situation will not only saturate but degenerate into toxicity and collapse.

Again this is all far too dramatic. Yes some some folks will take shortcuts and produce junk, as they've always done where there's an option to do so, but at the same time there'll always be a group of people who craft unique and high quality games because they're driven to by a love of the process. Yes the gaming landscape is changing but that's nothing new, it's always done that. The 90's produced endless shovelware but also a huge number of gems. My favourite game from the last few years is a free total conversion mod - this was done by a group of passionate and clever amatures and those people aren't just going to disappear because AI comes along and lets people generate crap in a weekend.