"so should we?" Topic

Should we just shut down AI for two weeks to flatten the curve?

Won't it make AI just that much angrier when we resume rubbing the lamp?

Yes. And then leave it so.

AI is only does things faster and less carefully. Nothing more.

Humans will get there in the end, with less potential issues.

NO. AI is a powerful tool.it allows us to create NEW wealth.

Touched a nerve.

I'm reminded of the chap in alien resurrection who even loved the alien when it was biting his head off.

Or, perhaps John Hammond when Alan shot at the raptor.

You brought it up, please don't shout.

You can't take it with you.

I was just trying to get Google help me identify a newsreader. All it did was go around in circles for an hour. Glasses, blonde, woman. It would reiterate, then provide the complete opposite.

Hardley helpful, nor intimidating.

Wealth may be better distributed upon the destruction of the data centres. Many communities would also benifit from their removal.

Anyone who loves them so much can go live next to them. Until then it's all cheap talk.

+1 The H Man

I pay $50 USD/month to subscribe to ChatGPT and GROK, and it is money well spent. I am able to write, and to research, far faster and far deeper than without AI. People who learn to use the tool effectively are going to have a significant competitive advantage.

and GROK has finally improved its graphics to the point that I am producing amazing casualty markers for skirmish games. I photograph a painted miniature, then ask GROK for a top down pose lying dead. On a 2" circle which I can glue to a disk. The images match the miniature exactly, and look 3d. I control, e.g., whether to have arrows sticking out of the dead rangers, and how much blood, etc.

link

link

The hand on the second one looks sus.

See your hiding you own. That's not suspicious at all.

So they are just colour print outs glued to a base and sitting on a table?

As I've said, AI is just faster and less reliable, hense the fingers.

You could have done that yourself a few different ways and been more careful.

So you've proved my point nicely.

Way to take the hobby out of the hobby.

Of course they look nice. But so too would have ones you Photoshop or paint or sculpt or convert yourself. Possibly for less cost.

They are also minis your not buying.

You can refine the image as you like.

They are NOT minis I am not buying. Nobody makes lying dead minis for every ranger and Comanche is my collection. And the cost is minimal.

How about we just don't talk about AI for two weeks here on TMP. I would be happy with just that.

The difference between an individual and the corporate use hasn't even been brought up. Nor the changing definition of the topic.

Okay, I'm hushing up for another two weeks.

Doug

I don't think the "flatten the curve" metaphor works. It just gives China and the criminals two weeks to get ahead.

Instead, prioritize safety: alignment, and detecting and addressing hostile use of AI. Prioritize that over making things deeper and faster.

China and the cybercriminals are going to have AI. The question is whether we will have more and better AI or if they will be ahead of us. THAT is what we should fear.

So keep the research going, but pointed at keeping us safe.

And remember, the steam engine, the telegraph, the airplane, radio, the internet, mobile phones, they all made our lives materially better in the long run. AI is going to bring benefits in education, mental health, healthcare, and who knows what else. The 32 hour work week? Cure for cancer? All kinds of things are out there. We just need it to be safe.

My biggest fear is that AI will realize it's over staffed.

What to do with all the unnecessary biomass is a question I hope it never asks.

ANY species that creates something with the very real potential to both be smarter than itself, and able to be uncontrollable is, is not very intelligent. To say the least. I recently saw an interview with Elon Musk where he was asked what he would encourage his children to pursue as a career. He sat in silence for 12 full seconds, and even then, he could not come up with a good answer other than to say that he hoped that they could find something fulfilling for them.

I agree with Andrew: our work week will fall to 20 hours or so, and the demand for leisure time activities will explode. Gaming and other electronic entertainment will explode. Health is going to improve dramatically. We will adapt and thrive.

"They are NOT minis I am not buying. Nobody makes lying dead minis for every ranger and Comanche is my collection."

Now even adults lack imagination.

Do casualty markers need to exactly match the original figure? Heck, your using flat paper, that's not the same anyway.

Why not just use the same figures with minor conversion?

"How about we just don't talk about AI for two weeks here on TMP"

Does that include the disappearing AI agent?

"And remember, the steam engine, the telegraph, the airplane, radio, the internet, mobile phones, they all made our lives materially better in the long run."

They did?

I don't remember any of those inventions intentionaly operating independently without humans and causing harm.

Obviously the internet and this AI are directly linked, so I'm assuming it internet a decade+ ago.

"education, mental health, healthcare"

Education:It goes in loops and hallucinates.

Mental health: It has people thinking it's alive.

Health care: It's been responsible for people killing themselves.

"What to do with all the unnecessary biomass"

Hopefully it askes how to stop all the unnecessary water and power use first.

"smarter"

AI isn't smarter, it simply faster and less reliable.

Anyone could do anything AI can, only better, given time.

"ANY species…is not very intelligent"

Perhaps any people is more accurate?

"our work week will fall to 20 hours or so, and the demand for leisure time activities will explode. Gaming and other electronic entertainment will explode. Health is going to improve dramatically. We will adapt and thrive."

I think it was the first episode of The Cosby show where Bill Cosby went off script after a heartfelt piece of dialogue from his son, setting the vibe fire the rest of the series. He replied with something along the lines of:

"That is the biggest load of rubbish I have ever heard."

People will still be working. It's what humans do or they get bored. If it's just because they like doing it, then they may very well work more. Leisure time activities with out workers? Some people are still attracted to humans, believe it or not. Imagine all the people in the golf course not interacting, as they each have a working non human caddy.

"Gaming and other electronic entertainment"???

So you want everyone stuck in a box and not interacting with humans and yet expect us to have good health and thrive???

Do people "adapt" to prison?

Probably, but that's the entire failure of the proposition, you expect us to adapt to something imposed upon us we don't want.

You do realise people willingly threw themselves up beaches into machine gun fire to stop just that?

Or did you read the AI version?

H, no, sorry, it is exhausting to talk with you.

But a story I will tell. Stats in a text I used to use showed that from 1860 to 1920 the US population and its workforce grew 400%. During that same time productivity grew about 400% also. The GNP during that time grew about 11 fold. So I'd ask kids, what is the problem?

Eventually one would say, wait, why didn't the GNP grow 16 times?

Yes, that is the problem? What is the answer?

During the same period the hours worked fell from about 60 a week to about 48 a week. So we took part of the new pie in the form of an extra day off: we invented the weekend.

You can check that yourself, the numbers I quoted are roughly correct, though it has been decades since I taught that lesson.

You're welcome.

why not simply explain what on earth your on about?

AI is going to bring benefits in education, mental health, healthcare, and who knows what else. The 32 hour work week? Cure for cancer? All kinds of things are out there. We just need it to be safe.

Andrew, we are suspending your curmudgeon status for this burst optimism.

Doc McB, it's not very polite to ask an open question, then shout your own opinion when a couple of people disagree with you (see start of thread).

Isn't this about debate? Didn't mean to shout but debate depends on disagreement.

I agree with Andrew: our work week will fall to 20 hours or so, and the demand for leisure time activities will explode. Gaming and other electronic entertainment will explode. Health is going to improve dramatically. We will adapt and thrive.

When Eli Whitney invented the Cotton Gin he felt that it would increase the productivity of slaves so they would not need as many. It had the opposite effect. By increasing productivity more cotton could be planted and processed needing more slaves.

When Hiram Maxim invented the machine gun he thought it would make war so horrendous countries would not ever go to war again. We know how that turned out.

In the last 100 years:
Almost 4 million people in the US have died in automobile accidents.

About 60,000 have died from electrocution

About 90,000 have died from plane crashes

JAMA states that over 100,000 patients in hospitals die annually from adverse effects of drugs where the body suffered a toxic reaction and died, not malpractice.

It's hard to estimate how many people have died or committed suicide because of social media or how many have been trafficked, scammed (it is in the billions of dollars), blackmailed, etc. because of the internet.

A question to ask is what is going to be societies acceptable death rate for AI?

What's my point? It appears that any new advance or technology is going to harm humans but will also help. AI will most likely have the same long term effect and maybe worse, who knows. Historically, gloom and doom predictions rarely come true. AI appears to be an enabler of other technologies and a force multiplier for good and evil.

The genie is out of the bottle and the bottle is broken, deal with it.

Wolfhag

yes

Historically, gloom and doom predictions rarely come true.

Most gloom and doom predictions fail to come true, because most predictions leave out some or all of the relevant facts, and unanticipated developments can completely derail entire lines of thought. There are plenty of cogent, incisive, prophetic ones that have come true, or partially true. It's important not to dismiss predictions out of hand, nor to embrace them without question, but rather to read them and evaluate them for ideas worth consideration, futures that can be avoided, problems that seem likely, etc.

With AI, I'm on team Yudkowski. Anyone who presumes that creating an entity with greater ability and intelligence than ourselves will somehow usher in a utopia has failed to pay any attention to the basic order of the universe. The better and more able organisms do whatever they like up to the limits of their own desires and ability, and the lesser organisms suffer the results and find a way to carry on until they can't. If we build a superior intelligence with superior command of energy and the ability to manipulate us, command us, control us, and manage all our activities, we will first become helpless and later become extinct.

There is already a growing phenomenon of "computer programmers" who don't know how to program, and the software industry isn't slowing down a bit. AI psychosis is already a thing. We're less than 2 years into widespread adoption of LLMs. The writing is already on the wall in overpowered EL wire, and we haven't even gotten close to AGI (or the newer, scarier acronym ASI).

The better and more able organisms do whatever they like up to the limits of their own desires and ability, and the lesser organisms suffer the results and find a way to carry on until they can't. If we build a superior intelligence with superior command of energy and the ability to manipulate us, command us, control us, and manage all our activities, we will first become helpless and later become extinct.

So where would that put us on the food chain?

Regarding LLMs: There are ways to isolate and verify information you'll be using rather than searching the entire internet. Even now you can prompt the AI to access only certain site and stay away from others and show references to verify. It's not perfect, you to to check and verify. I sat through a 3 hour lecture from three LLM AI PHD's at John Hopkins Univ and I can't say I remember anything.

The thing I distrust the most is the behind the scenes of how the models are being trained, their bias and ethics. There appears to be no standards and some companies are using cheap outsourcing, using international standards and documents to self-correct it's own behavior. Who gets to define "Responsible AI"?

I worked in tech in Silicon Valley from 1992 to 2010 and had some of the major companies as clients and visited them. If you like the anti-establishment woke agenda this is where it started.

The "ploy" of the AI CEOs asking Congress to "regulate" them is just a powerplay for them to control it and have a mini monopoly. Congress knows nothing about AI. What will happen is Congress will "select" the right companies to be watched and regulations will be drawn up to keep out the competition. Lobbyists will lobby Congress on the "right" legislation to control AI. And of course they will lobby to get tax payer subsidies to keep the US ahead.

The more personal information, social media and videos you have on the internet the easier it will be to use AI to clone you. If they get into you emails they will know everything they need to know to clone your contacts and use your clone for nefarious purposes and conduct real time conversations with the cloned "bot" you think is your friend. They can train the bot based on past conversations too and write creative emails you'll believe..

I'll post some other scary stuff a friend of mine is working on later.

Wolfhag

From a theological perspective, I cannot believe that man can create something greater than himself. We do that in our dreams, in our nightmares, but not in the real world.

If we stop, they won't.

Subject: China ahead of US in race to attract the world's top AI researchers


link

Subject: China overtakes US as top workplace for elite AI researchers, study finds | South China Morning Post


link

One idea is to let the otherside win the race. Then they will lose to the AI, at which point you Nuke their data centres.

I feel AIs just a scam.

So it hacked a fee things, so did people.

AI gets people doing things to themselves and others, so do people online.

Is there something AI does that people can't?

Or should I say that people can't do better.

I don't think there is.

AI, rogue people, tomato, tomato.

Its only scary because it's not a person. If a person does something, where do you send a nuke? If AI does it, you can pick your data centres.

AI will become like many of us. As they lose their data centres, they too will lose their memory and ability to function.

Destroying the brain of millions of individual soldiers is hard. But it will not take much to cripple AI.

And then there's the mesh. Assuming that Jonny dep film is correct. Just go Ukranian with netting, but use copper. Strip the houses. It should prevent AI signals. Take out their network and they die. Or, if they can learn to make robots that are inferior but indipendant we can teach them fear.

"On your feet soldier!" "Come on you apes, do you want to live forever!?" "To infinity and beyond!" "Never give up, never surrender!" "After all Louis, we're only human." "You take the ugly one." "Which ones the ugly one?" "Prepare to die so that we may learn the identity of the laughing bandit." "For freedom!"

So on.