• Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more

Merged Artificial Intelligence

arthwollipot said:
The video I posted is about 15 minutes long, so it's less investment of your time than Forumvision is, and it goes in detail about how the AI agents behaved and exactly how and why they demonstrate agency. I've spent considerable effort now trying to summarise it and with the above confession I am officially done. Either you want to know, and you'll spend 15 minutes learning something, or you don't want to know and you won't. I'm done here. Good day.
Click to expand...
See you next kurzgesagt, then.
 
arthwollipot said:
I'm not talking about the kurzgesagt one, I'm talking about the more recent one.
Click to expand...
Yeah but that sounds better than see you next... [checks video description] literal venture capitalist techbro desperately hyping the capabilities of AI dude you've got to apply a little bit more critical thinking here. You don't need to hear one word from his mouth to know he is selling you a bill of goods and you are buying it.

Doesn't roll off the tongue, see.
 
Finnish ⁠police have launched a criminal investigation into possible breaches ⁠of the country’s environmental ⁠laws at the site of a planned Google datacentre in rural northern Finland.

Authorities allege that Tuike Finland, a Google subsidiary, cleared and drained 330 hectares (815 acres) of forest at the planned site in Muhos before environmental impact assessments required by Finnish law had been completed.

www.theguardian.com

Finland launches criminal investigation into Google’s clearing of forest for datacentre site

Authorities allege company may have breached Finnish environmental laws by forgoing required impact assessment
www.theguardian.com www.theguardian.com
 
Emily's Cat said:
A lot of this discussion is technically out of my wheelhouse. I want to ask a different question:

When you're using an AI tool, are you polite to it? Do you say please and thank you?
Click to expand...
Yes. As I have stated before I am being polite so I may be kept as a pet when the AIs take over.

Seriously: The tone you take with an AI will be reflected back at you so it really boils down to what style of response you want.
 
Emily's Cat said:
A lot of this discussion is technically out of my wheelhouse. I want to ask a different question:

When you're using an AI tool, are you polite to it? Do you say please and thank you?
Click to expand...

I don't use AI, but according to this study being polite helps.
Our study finds that the politeness of prompts can significantly affect LLM performance. This phenomenon is thought to reflect human social behavior. The study notes that using impolite prompts can result in the low performance of LLMs, which may lead to increased bias, incorrect answers, or refusal of answers.
Click to expand...
 
Beelzebuddy said:
Yeah but that sounds better than see you next... [checks video description] literal venture capitalist techbro desperately hyping the capabilities of AI dude you've got to apply a little bit more critical thinking here. You don't need to hear one word from his mouth to know he is selling you a bill of goods and you are buying it.

Doesn't roll off the tongue, see.
Click to expand...
You've made the effort to check the description. Why won't you watch the video?
 
arthwollipot said:
You've made the effort to check the description. Why won't you watch the video?
Click to expand...
Because I don't see any need to. I find the premise trivial, and it's not actually relevant to the point you're trying to make. The fact that I don't give one cold, wet ◊◊◊◊ what some techbro jagoff wants me to think about his latest VC investment is just a bonus.
 
Darat said:
Yes. As I have stated before I am being polite so I may be kept as a pet when the AIs take over.

Seriously: The tone you take with an AI will be reflected back at you so it really boils down to what style of response you want.
Click to expand...
I end up being polite largely because of the style of responses I get. Nothing fancy, just Copilot at work and Google's AI whatever it's called. Copilot I just can't help being polite - it's polite, it gives good feedback, and it digs through my neverending email archives far more efficiently than I can and finds the right email from fairly vague descriptions of what I'm looking for. It's genuinely helpful, so I almost always say thank you. And because it's so effective, and the responses are so close to natural conversation, I almost always ask for help. I'll say "Would you look through my emails and find one that has an attachment or a link about the contract with vendorX? I think it was at least six months ago, and my colleague Alex might be copied on it." And then it goes through it's routine, it finds a few that it thinks are good matches, suggests which one it thinks is the most likely one, and often asks follow-up questions. So when it finds in 20 seconds an email it would have taken me 20 minutes to hunt down... I simply can't help but say "Thank you, that's the one".

Google is a little more hit or miss. A lot of the time I'm just after a particular word or event that my brain has memory holed and I can't recall with enough specificity to find it on my own. If all I have is a general question, I just ask it "What's a good word that means something like forbidden, but isn't as strong and is a little funny?" If it comes back with the right word on the first try, I don't necessarily thank it. If it asks for context or an example of how I'll use it, and there's some interaction, I almost always thank it.

I find it an interesting dynamic.
 
Beelzebuddy said:
Because I don't see any need to. I find the premise trivial, and it's not actually relevant to the point you're trying to make. The fact that I don't give one cold, wet ◊◊◊◊ what some techbro jagoff wants me to think about his latest VC investment is just a bonus.
Click to expand...
Well, you're the one who keeps saying that it's safe and everybody's fears are unfounded because it's just a tool like any other. Why should it matter if the techbros actually agree with you?

Anyway, you've stated your determination to not learn anything so like I said, I'm done.
 
arthwollipot said:
Well, you're the one who keeps saying that it's safe and everybody's fears are unfounded because it's just a tool like any other. Why should it matter if the techbros actually agree with you?

Anyway, you've stated your determination to not learn anything so like I said, I'm done.
Click to expand...
Just as a reminder, this conversation thread started about country music.

But if you want to talk about the other stuff, who said anything about "safe?" Tools aren't safe. Not the good ones. Ask any shop teacher. While you're there, ask him who is ultimately responsible for any misuse of their tools, including understanding why their operation leads to the results it does. Then compare that to the tech bros' dick measuring contests about how stupidly easy it is to steal data and wreck crap with their tools. Oh but it's totes the AI doing it because its sapience is just a single funding cycle away, and not their own fault for being venal sociopaths.
 
arthwollipot said:
When even your vaunted techbro jagoffs are saying that it's dangerous, who are you going to believe?
Click to expand...
Yourself. Use the damn things yourself and you'll see what they're (not) capable of. They aren't going to hack a website or make a hit country single unless you yourself use them to do it.
 
We are at the point where the current dangers of the technology exceed the current benefits.
And no one has a plan how to make AI inherently safer, simply because they have no concept of safety.
We have to go the late-robot Asimov route and make AI so small and dedicated to specific tasks that the potential damage is also minimized.

And no, it's not going to take over as long as people remember that they can pull the plug anytime
 
Beelzebuddy said:
Yourself. Use the damn things yourself and you'll see what they're (not) capable of. They aren't going to hack a website or make a hit country single unless you yourself use them to do it.
Click to expand...
No, I won't. Because I've seen how awful they are. Every video I produce is going to have "No AI was used to make this video" tacked onto the end.
 
The Great Zaganza said:
And no, it's not going to take over as long as people remember that they can pull the plug anytime
Click to expand...
Not the way you think, but it will take over - because humans are lazy. We won't pull the plug because that would mean doing the work ourselves. We are entering a new era of slavery where the 'masters' live in constant fear of being murdered by their AI workers, but are psychologically unable to give it up - even if it costs more than employing humans. The more we rely on AI the more dangerous it gets, because we will be locked in to using it everywhere as nobody has the skills to replace it.

Or perhaps not. AI stocks are getting very wobbly right now as the hype crashes into reality. A huge rift is opening up between the AI evangelists and those who want nothing to do with it. Nuclear power stalled partly due to opposition by people who thought it was unsafe (validated every time there was a major incident), and AI may suffer similarly.

All those YouTube videos imploring you to try AI are doing the opposite of what was intended. They just tell us that AI companies are desperate to get us hooked on their marginally useful product. Every day the MSM publishes articles questioning where AI is taking us. Calling it 'Super-intelligent' isn't helping either, it just stokes the fear and loathing.

Yesterday I was talking about AI with a friend who does risk management. He uses it to format reports, which is fine. However his team leader uses it for everything. My friend complained that the guy 'wrote' a report which was just vacuous AI slop with no detail on what risks were present - at a bitumen plant. Management think the guy's a genius. In reality he's just lazy.
 
Last edited:
You must log in or register to reply here.

Share this page

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom