We say "practice makes perfect," so they asked the AI, tried it out themselves, and gained valuable experience.
They certainly won't be asking the AI about this again.

I wish people would not only share their experiences but also the lessons learned. We learn from our mistakes.

 

My sense of humour wanted to say "this is why I have been getting stomach upsets, I have been eating silicon chips instead of potato chips..." Silliness aside...
I'd agree in business now definitely and speaking to each other from obviously different countries here, also definitely, but not everything, thank God.

 

This is why every AI answer still needs to be check, it´s not like god, like some assume..
It can have some serious issues by trusting everything it says, I really hope people don´t use it for electronics, 30 mA can kill you..

 

Yes. As always: it depends.
I use AI only in fields I really know sth. about. Not for basic knowledge aquisition, where I rely on paper books (yes, old white man here)

But in the cases of physics and chemistry, LLM are really good in helping on e.g. solving standard equations, computing molar solutions et cetera. There it works like the good old calculator. This way it is only a tool like every other in my toolbox.

BUT. I would never recommend using AI for educating children or lettimng chiuldren play with it as a toy.
The risk is too great that children will become addicted or stop learning because they think they have AI to do their work, as has been already the case with Wikipedia in recent years.
The use of all modern technology, even cell phones and tablets, should only take place under strict adult supervision.


My personal experience:
When I let my granddaughter/son use my tablet to watch videos or play games on the internet, it's like throwing them into a deep, dark well with no water. They can't get out again without help.
Therefore, if it were up to me: no AI, no anti-social media, no YouTube, no mobile phones before the age of 16.

Instead, my wife and I play with our grandchildren indoors and outdoors and always come up with new adventures. The dishwasher is an intergalactic spaceship when we load it. Cardboard boxes are converted into houses, etc.

 

Last edited: Dec 6, 2025 at 3:30 PM

Yes, you're absolutely right. There's probably not much more you can do than warn people.

When I was a child, I accidentally touched the socket while trying to change the lightbulb in the ceiling lamp, and 220 volts went all the way up to my big toe and the tip of my tongue. It took me a few minutes to process what had happened and recover from the shock. I hope nothing was damaged in my brain. Since then, I've always been careful around electricity.

Back then, there were no instructions on how to change a lightbulb, and it wasn't even on any packaging.
People were simply assumed to know how to change a lightbulb.

If you blindly trust AI, you've already lost. I was surprised when smartphones first appeared, wondering where the warning labels were. There weren't any. Now, smartphone use is being gradually restricted in schools because it severely impairs concentration and learning, and this will only get worse with AI.

Back then, there were no instructions on how to change a lightbulb. I know some people who have only a superficial understanding of things, many myths and legends they simply believed, and then spread untruths with conviction. In a forum, you have the chance to correct falsehoods.

Recently, a young man touched a railing and a damaged lamp at the train station at the same time and got electrocuted. His friend, remaining alert, pulled him away from the danger by his backpack. The young man survived, and his friend received a government award. Life can end faster than you think; sometimes you're even saved by other people.

What does that mean? We should all do our jobs as well as possible, and we also have to limit stress because people make many mistakes under stress, to the detriment of others.

 

Like my teacher always said, do your homework..
Learnd it thru the hard way.. still careful with the AI

 

As so often happens, I have something to say regarding the subject, I read the thread, then find I've used up all the spare time I had to write a post. I guess I should go ask an AI to compose an answer on what it is that I think, now that I've no time to do it myself...

 

Certainly not, otherwise the image would have already gone viral

 

LLMs are not vast, reliable information databases like in sci-fi movies, it is more like advanced form of auto complete in your phone, xd, the knowledge is always mostly hallucinations, that's why always is ask it to search info online and to give links as proofs.

 

So it's ok to believe everything else you read on the internet? Or should everything be fact checked? Maybe I will start with all 200 of the posts written by the thread starter.

 

That take is stuck in 2020. Calling modern language models an advanced autocomplete is the kind of thing people say when they stopped learning right after hearing one headline and decided it was enough to sound informed. It is like calling a full studio rig a cassette recorder because both can capture sound.

The claim that the knowledge is mostly hallucinations just shows they never actually used the thing for anything serious. Models hallucinate when you treat them like a magic answer box. When you give them structure or real tasks, they produce consistent work at a level a phone keyboard could never touch.

And the part about always needing online searches and links as proof is the funniest bit. They do not trust the model but apparently trust whatever random web page shows up. The internet is where half of the bad information came from in the first place.

If anything, these models expose how often people repeat nonsense they picked up online without thinking. The model is not the weak spot here. That comment is.

Even worse, I propose, is creating a "random number username" is some attempt to have any kind of information posted under such an alias look like it is from either a "hacker" or a "cracker". Which look are you shooting for? Because it's probably neither anyway. All I need to see is that "autocomplete" comment. King of the Randos.

This is a fair take, like they are moving towards in Australia. The only problem is with being something relatively close in age to you, that I still remember all of the other forms of trouble available to those of such age without computers. I'm sure part of that was just New York in the 80s, but there was always more than enough outdoors variety trouble available if someone wanted to find it. If said younger people were anything like most users of this forum probably were even at our "advanced ages"; you'd need to go a lot further than just blocking their internet access. You will end up taking computer security lessons from them.

 

Last edited: Dec 6, 2025 at 8:12 PM

Yes, always read what you write or copy/paste, and be sure about it

 

You need to properly understand how AI works. Theres far too many misinformed, half-baked assumptions flying around and it's not helpful. LLMs are not like autocomplete at all - thats missing the point entirely. Really, read about how an LLM (data and model) are constructed and trained, and learn about how the technology uses the dataset and then, with that knowledge in hand, learn about prompt engineering. Once you have a proper grasp of what is going on, how to harness it productively, then you'll start to benefit from the absolutely transformative tool that AI can be.

If you just poke it about a bit and form your opinion based on the rather large body of ignorance, then you'll miss out on something that can be useful to you. You'll fall behind and join a band of credulous luddites. Up to you really.

You know, maybe actually find out for yourself by grabbing something like Comfy UI and building some toys with it. It's all there, generally open source and waiting for you to get familiar with it.

 

and don't push until you're 100% sure the info is legit
that should be the must

 

If this system meant anything it would not be in favour to those who spread, copy-paste not verified and incorrect information

 

I'm with Foster on this whether he had an opinion or not. REMEMBER THE NOTES!!!