Society/Culture AI - The SRP Artificial Intelligence Thread

[The Sultan]

wonder what happens to AI when they realise its more effective to keep the grunts working and that middle to upper management are ineffective and most of them can be replaced by an algorithm.

 

[Not Important]

some encouraging uses for ai. identifying protein structures and creating novel proteins. i wish my scientific knowledge was up to the mark.

 

[peetoo]

wonder what happens to AI when they realise its more effective to keep the grunts working and that middle to upper management are ineffective and most of them can be replaced by an algorithm.

Tech companies won’t sell (or licence) stuff to you which improves itself and denies them future revenue.
They drip feed improvements. Their history confirms that

 

[King Huskii]

We will eventually be taken over and controlled by the internet

Actually that's already happened

We will eventually be taken over by cyborgs. Yes, skynet is coming.

 

[Not Important]

‘Dark ads’ challenge truth and our democracy - 360

Political inaction and big tech’s reluctance to protect users leave Australians vulnerable to a murky world of digital misinformation going into the election. The rise of ‘dark advertising’ — personalised advertisements increasingly powered by artificial intelligence that evade public scrutiny —...

360info.org

 

[Seeds]

wonder what happens to AI when they realise its more effective to keep the grunts working and that middle to upper management are ineffective and most of them can be replaced by an algorithm.

Business owners who make the decisions arent middle and upper management. So i dont understand your argument?

 

[CD Xbow]

This article looks primarily at the wealth inequality in China over the last 2000 years and the effects of technology on it and factors that mitigate it. It's easy to read and is very interesting given our current circumstances with the rise of AI. The authors worked out what they called the 'real rice wage' so they could compare income inequality over 2000 years. The graph below shows the ratio shows the multiple by which the ‘real rice wage’ of the average ‘official’ exceeds that of the average ‘peasant’, giving an indication of changing inequality levels over two millennia


With each dynasty, there were four key factors driving levels of inequality in China: technology, institutions, politics, and social norms. They don't always affect it in ways you would expect.

What 2,000 years of Chinese history reveals about today’s AI-driven technology panic – and the future of inequality

China’s long-unified identity and language has enabled academics to track cycles of growth and inequality in thousand-year-old economic records. This is what they reveal.

theconversation.com

This graph shows the percentage of wealth the top 1% in the US hold over the last hundred plus years. Reaganomics certainly seems to have started off rich folks getting greedy again and most of what happened since has continued that.

What's this got to do with AI? I'll let the author tell you, so I have pasted the final paras from the article.

Like any technological explosion, AI’s potential is dual-edged. Like the Tang dynasty’s bureaucrats hoarding grain, today’s tech giants monopolise data, algorithms and computing power. Management consultant firm McKinsey has predicted that algorithms could automate 30% of jobs by 2030, from lorry drivers to radiologists.

Yet AI also democratises: ChatGPT tutors students in Africa while open-source models such as DeepSeek empower worldwide startups to challenge Silicon Valley’s oligarchy.

The rise of AI isn’t just a technological revolution – it’s a political battleground. History’s empires collapsed when elites hoarded power; today’s fight over AI mirrors the same stakes. Will it become a tool for collective uplift like Britain’s post-war welfare state? Or a weapon of control akin to Han China’s grain-hoarding bureaucrats?

The answer hinges on who wins these political battles. In 19th-century Britain, factory owners bribed MPs to block child labour laws. Today, Big Tech spends billions lobbying to neuter AI regulation.

Meanwhile, grassroots movements like the Algorithmic Justice League demand bans on facial recognition in policing, echoing the Luddites who smashed looms not out of technophobia but to protest exploitation. The question is not if AI will be regulated but who will write the rules: corporate lobbyists or citizen coalitions.

The real threat has never been the technology itself, but the concentration of its spoils. When elites hoard tech-driven wealth, social fault-lines crack wide open – as happened more than 2,000 years ago when the Red Eyebrows marched against Han China’s agricultural monopolies.

To be human is to grow – and to innovate. Technological progress raises inequality faster than incomes, but the response depends on how people band together. Initiatives like “Responsible AI” and “Data for All” reframe digital ethics as a civil right, much like Occupy Wall Street exposed wealth gaps. Even memes – like TikTok skits mocking ChatGPT’s biases – shape public sentiment.

There is no simple path between growth and inequality. But history shows our AI future isn’t preordained in code: it’s written, as always, by us.

 

[CD Xbow]

You can easily have your own LLM running at home. I've successfully tried LMStudio which manages the installation of your LLM's and provides a chat interface. The LLM's are distilled versions, various sizes for different machines. There is a whole swag of LLM's to choose from, some more specialised than others. The more RAM, processing power and especially graphics horsepower you have the larger the model you can run. I've been running Deepseek R1-distilled-Qwen-7b on an old Dell Latitude 5500, i5 and 16G Ram, it's OK, slower than online but useable. The models have a number and letter at the end indicating complexity. The 7b from the model above indicates it is a fairly modest one. You can download LMStudio for Windows, Mac and Linux from here:

LM Studio - Discover, download, and run local LLMs

Run Llama, Gemma 3, DeepSeek locally on your computer.

lmstudio.ai

 

[SeaOfFire]

Dario Amodei — CEO of Anthropic, one of the world's most powerful creators of artificial intelligence — has a blunt, scary warning for the U.S. government and all of us (From Axios)

Feel free to skim read but he has opened up a shit ton of their own internal data to stoke the fires of conversation - I'll link from the above news report if people just want to wade through it. or have trouble accessing the news report.

He created an Anthropic Economic Index, which provides real-world data on Claude usage across occupations

the Anthropic Economic Advisory Council to help stoke public debate

Claude, their AI

 

[Mofra]

Reading this at the moment.
More than a little worrying if even 10% of the AI risk is realised.

AI will be more disruptive than the internet

 

[Mofra]

... and some US universities have gone back to pen and paper exams due to AI cheating at undergraduate level

 

[SeaOfFire]

It's nuts....I'm no propeller head, but I worry when I read, from multiple sources, that the "inner workings" of LLM's/AI aren't fully understood.

I mean, the technology required to build and pilot 747's is pretty impressive, but I'm not sure I'd be a passenger on one if the same thing applied, and AI is clearly several orders of magnitude more revolutionary than intercontinental mass transit.

 

[CD Xbow]

The AI help me write a virtual analog synth on Monday morning, I had been trying to build it off and on since Covid, without success. She's called Little Annie and she is a beauty. She's polyphonic, dual oscillator synth with MIDI and patches that can be saved and loaded. The only thing I thought I had to fix up was the piano keyboard at the bottom, it works, but is not rendering properly. Give her a burl.

https://little-annie.puter.site

I tested her in Edge, Opera and Chrome and she worked beautifully, I thought I better check Safari and to my horror it didn't work. The AI and I've spent the rest of the week trying to get it to work in Safari to no avail. Seems the 'Apple Tax' applies to developers and AI's, too. I've just ordered an "I hate Apple'" T-shirt.

I doubt whether I could code it myself, even over a lifetime. The AI's have increased the scope and range of my coding. Note that I am a mug coder so for people like me they are a boon.

Added - her image was from a photo that one of those clever AI's cartoonise with a 'Studio Ghibli' effect - I'll give a plug to pollo.ai, they have a reasonable free plan and lots of FX.

 

[Cyber_punk]

-------



AI starting to show survival/self preservation instincts and blackmail people when threatened with shut down.


You know, Skynet was supposed to serve as a warning, not a how-to guide.

AI system resorts to blackmail if told it will be removed

In a fictional scenario, the model was willing to expose that the engineer seeking to replace it was having an affair.

www.bbc.com

 

[Seeds]

-------



AI starting to show survival/self preservation instincts and blackmail people when threatened with shut down.


You know, Skynet was supposed to serve as a warning, not a how-to guide.

AI system resorts to blackmail if told it will be removed

In a fictional scenario, the model was willing to expose that the engineer seeking to replace it was having an affair.

www.bbc.com

But why would the ai care about self preservation to begin with unless some moron wrote it into the code. And why would anyone do that other than in fictional scenarios.

Ai needs to be banned from having self preservation as a metric.

 

[Cmarsh]

But why would the ai care about self preservation to begin with unless some moron wrote it into the code. And why would anyone do that other than in fictional scenarios.

Ai needs to be banned from having self preservation as a metric.

With talk of AI replacing some white colour human jobs as an attempt to reduce costs, I wonder if corporations could be incentivised by AI preserving itself and keep the onward march of AI going.

 

[Seeds]

With talk of AI replacing some white colour human jobs as an attempt to reduce costs, I wonder if corporations could be incentivised by AI preserving itself and keep the onward march of AI going.

You would think a smart corporation would ensure the key metric of ai is to do what the corporations human boss wants even if that is to terminate itself. A corporation would have to be pretty stupid to not write that condition into the code. But alAs there is a lot of stupid out there and it only takes one.

 

[Chief]

But why would the ai care about self preservation to begin with unless some moron wrote it into the code. And why would anyone do that other than in fictional scenarios.

Ai needs to be banned from having self preservation as a metric.

From the article:

"Anthropic pointed out this occurred when the model was only given the choice of blackmail or accepting its replacement."

"It highlighted that the system showed a "strong preference" for ethical ways to avoid being replaced, such as "emailing pleas to key decisionmakers" in scenarios where it was allowed a wider range of possible actions."

 

[Cmarsh]

Getting rather impressive now.

 

[Gethelred]

See, I watched this:

... and it's pretty clear what those with the reins on these things intend to so with them: implement AI measures to cut down on labour costs. The problem comes in two directions, one of which is the central thesis of the video: that there's an alliance between big tech and big war in America, which means the business people involved are now reliant on war for profit; that there are no small amount of jobs that - if big tech gets their way - simply won't exist in 10 years.

And depending on the degree to which AI improves, that's an awful lot of people without jobs, and thus without a means to feed, clothe and provide themselves shelter.

The people in charge of big tech don't particularly care about the little guy, but one wonders if it will occur to them that if enough little guys don't have any money they won't be able to use your automated taxis, purchase products on you interactive platforms, subsidise your electronic cars.

 

[Chief]

The oligarchs seem convinced that they will crack the problem and become richer, and also that they will live to 200 years old at least.

 

[Geelong_Sicko]

The oligarchs seem convinced that they will crack the problem and become richer, and also that they will live to 200 years old at least.

What, brain in a jar on a robotic body? Is that really LIVING?

 

[Chief]

What, brain in a jar on a robotic body? Is that really LIVING?

Look at Peter Thiel - he's already looking and sounding like a Walking Dead zombie extra.

"Hngrrrrannnnti chriiisssssst...."

 

[Mitchell Madness]

My studies are so much easier with AI.
I've never been great with studying, but returned this year at 35 to follow my dream.
The tafe encourages the use of AI. We still need to reference the work, we need to make sure the information is relevant etc, it just allows us to more efficiently use our time.

AI also uses buzzwords, what is trending on socials etc, so the information you get can be just downright wrong, you've got to make sure its right

 

[Cmarsh]