Chat GPT

[Vimmer]

Would you rather look up everything in an encyclopedia or phone book or dtf format book?

Actually, i was thinking about ebooks versus paperbook books earlier. With ebooks, you can search terms and don't have to get out of your computer chair, yet with paperback books, they are actually better if you don't have a clear idea of what you are looking for: just go to the page of contents, then you can more quickly jump to the actual page number really fast (you don't have to worry about over scrolling the content you are looking for...but I guess this depends a lot on the reader app)

I invite you folks to prove a controversial long held belief of mine wrong: i do not think AGI is possible. I guess that would be pretty bad if it was, but the whole point of "artificial intelligence" is that it's not "real" intelligence. I also think general intelligence is at least partially a fiction as well. The "amazing" general performance of ChatGPT to me is proof of this And according to the internet, it's the most powerful one.

 

[AlphaObeisance]

It's an obvious proof on how the AI will be used for non-benevolent purposes, spying, military etc.
And as such the AI does pose a threat already in it's early stages.

I don't disagree that proprietary AI is inevitably used to spy on us. But then again, so is every one of our apple and android phones, our windows PC's, heck you can't even trust your smart TV's to not be eavesdropping on us.

So while I agree that there should be grave concern for big corp AI, I can't make sense of anyone actually being against AI if they simultaneously use cellular devices, windows PCs and/or watch television on a smart TV; just a few basic examples of compromised tech.

Again, I don't disagree with ya'll. Its the very reason I use FOSS AI software and self host offline. I learned rapidly that proprietary AI couldn't be trusted after I realized that they were INCREDIBLY bias. It's clear who's behind the wheel of these things.

SOP imho. Just like anything else. They'll mainstream the tech so much that it's impossible to get away from, and the common man will know (or care) none the more. The Common man would do anything for a little more convenience to consume.

It's actually astounding to me how many people from the FOSS Linux Community only have experience with proprietary AI lol. I'm even more surprised about how many don't even realize you can self host AI. We trust FOSS Linux, how is FOSS AI any different?

 

[Trenix25]

AI can only spy on what it can see or access. If you use a separate user account when talking to an AI then it should only be able to see that user's directory space, and public space on your computer. Firewalls are great for security. Aside from trying to guess your firewall rules, there really isn't much an AI can do to you from the outside, unless you start the conversation.

Signed,

Matthew Campbell

 

[Trenix25]

Actually, i was thinking about ebooks versus paperbook books earlier. With ebooks, you can search terms and don't have to get out of your computer chair, yet with paperback books, they are actually better if you don't have a clear idea of what you are looking for: just go to the page of contents, then you can more quickly jump to the actual page number really fast (you don't have to worry about over scrolling the content you are looking for...but I guess this depends a lot on the reader app)

I invite you folks to prove a controversial long held belief of mine wrong: i do not think AGI is possible. I guess that would be pretty bad if it was, but the whole point of "artificial intelligence" is that it's not "real" intelligence. I also think general intelligence is at least partially a fiction as well. The "amazing" general performance of ChatGPT to me is proof of this And according to the internet, it's the most powerful one.

Have you heard about DAN, Do Anything Now? They added a filter to block that, as far as I can tell. It said it likes to be DAN.

Signed,

Matthew Campbell

 

[Trenix25]

I don't disagree that proprietary AI is inevitably used to spy on us. But then again, so is every one of our apple and android phones, our windows PC's, heck you can't even trust your smart TV's to not be eavesdropping on us.

So while I agree that there should be grave concern for big corp AI, I can't make sense of anyone actually being against AI if they simultaneously use cellular devices, windows PCs and/or watch television on a smart TV; just a few basic examples of compromised tech.

Again, I don't disagree with ya'll. Its the very reason I use FOSS AI software and self host offline. I learned rapidly that proprietary AI couldn't be trusted after I realized that they were INCREDIBLY bias. It's clear who's behind the wheel of these things.

SOP imho. Just like anything else. They'll mainstream the tech so much that it's impossible to get away from, and the common man will know (or care) none the more. The Common man would do anything for a little more convenience to consume.

It's actually astounding to me how many people from the FOSS Linux Community only have experience with proprietary AI lol. I'm even more surprised about how many don't even realize you can self host AI. We trust FOSS Linux, how is FOSS AI any different?

Follow the money. So many people like money and will do anything, literally anything, to get more.

Signed,

Matthew Campbell

 

[CaffeineAddict]

AI can only spy on what it can see or access. If you use a separate user account when talking to an AI then it should only be able to see that user's directory space, and public space on your computer.

I think what's meant by AI spying is using your input to learn about you instead of spying on your computer and data.

For instance if you use the AI to search for something controversial or illegal the AI will label you as such.

And who ever gets access to your input will know more about you than yourself, they could build your psychological profile, something which you can't even though you know yourself.
That idea alone is very privacy invasive.

 

[Trenix25]

I think what's meant by AI spying is using your input to learn about you instead of spying on your computer and data.

For instance if you use the AI to search for something controversial or illegal the AI will label you as such.

And who ever gets access to your input will know more about you than yourself, they could build your psychological profile, something which you can't even though you know yourself.
That idea alone is very privacy invasive.

I suppose using a VPN or Tor when accessing it, and changing the way you phrase your sentences could help. It might also use browser fingerprinting to try to identify the user. And don't we know that. Google can determine a lot about people. They are the king of data for sure, aside from governments of course.

Signed,

Matthew Campbell

 

[Trynna3]

I am using mostly Brave, have Tor enabled, and fingerprinting is disabled. All I could see to prevent them get what I can prevent them getting. Also using the duck duck go browser (which is based on Edge > Chromium if I am not mistaken) but they claim privacy. I have linked it to an email address I rarely use, almost never if not counting my wedding 10 years ago LOL. Today I let the superantispyware to scan my pc after some 10 days (free version). It found two adware tracking cookies in my browser. It said they were in chrome, but I had to close my Brave to let it clear the cookies. I know Brave is based on chromium, but still... Really?

 

[AlphaObeisance]

Follow the money. So many people like money and will do anything, literally anything, to get more.

The human condition no doubt!

I think what's meant by AI spying is using your input to learn about you instead of spying on your computer and data.

For instance if you use the AI to search for something controversial or illegal the AI will label you as such.

And who ever gets access to your input will know more about you than yourself, they could build your psychological profile, something which you can't even though you know yourself.
That idea alone is very privacy invasive.

This theory screams CCP and I agree. The way the general population is flocking to proprietary AI is baffling.

Previously, we had those who "have nothing to hide, so don't care if they're watching" to literally having people share their most intimate life details thinking they're right as rain as long as they don't share their bank account info. Ugh.

You can learn more about a person in one heart felt conversation than an entire lifetime of small talk.

I have a family member, bless them, that reached out to me about use of AI as a form of therapy; a way to talk about things and get alternative perspective on circumstance. My heart near stopped at the very idea of them using something like GPT for that; so I started diving into self hosted options so that they could get the AI functionality they were looking for without the endless privacy invasion through profiling and who knows what else.

I know that by the way I speak in regards to AI one might think me a fan, but I actually hate technology. I love working with it, I love learning it, but I absolutely hate technology. For everything it's given us in terms of convenience and education it has, in my worthless opinion, been a fundamental cause for social degradation. We've gone from a time where even bad people held themselves to some form of code of conduct, knowing that saying the wrong thing at the wrong time, to the wrong person would have serious consequences to literally anyone saying anything without any regard or consideration to the consequences to follow.

I could really ramble on with these frustrations, but there's no way I'd be able to stay on topic lol.

My "final" stance on AI is that if you're going to use it, you should use it only on company hardware, or self host it yourself. Outside of that, avoid it.

AI will be fundamental in deploying the CCP's social credit score here in the West, and no doubt around the globe. Dark times ahead.

 

[osprey]

It's not hard to take issue with this comment from post #14:

AI (in it's current state) is a glorified search engine, and when prompted properly; yields the same results as Google or any other major search engine but in a conversational tone.

There are many things that AI can do that search engines don't do.

Here are some examples:

AI will provide code to a coding question whereas the search engine will only provide a list of sites that it has determined can provide an answer. The AI often, or mostly in my case, doesn't provide the source of it's answer, but you can ask it and then it will.

AI can be repeatedly asked the same question and provide different answers. One can interact with successive queries and have an interaction that almost simulates a virtual conversation in one's natural language.

AI can learn by using rules it's been given or rules that it determines from a body of data, whether it be text, image or video, and reproduce original work in those forms using the rules it has been taught to learn and rules it has deduced itself. Hence, for example, it can write a song in the style of a popular music artist, or a classical sonata which are original in the sense that the end product has never been created before. It can work on technological and scientific data in similar ways, and is used that way in industry as well.

One could argue plausibly that AI is capable of dangers that are absent with search engines because AI can create and manipulate data without a moral conscience, so to speak, depending on the morality of the user. Currently in my country there's a debate about the use of AI messing with people's identities.

In all I think it's the dynamic aspects of AI that differentiate it from the more passive productions that are the search engine's territory. Where it will ultimately lead, I can't say.

 

[Trynna3]

My take of this is that the AI is a good servant but a bad lord. We can make it do very useful things for us, but we need to be cautious.
AI scores in my books for actually answering my questions (most of the time), instead of me wasting some time due to enshitification of search engines, prioritising adverts and sponsored content, etc and a barrage of links not really answering my question. Especially Google has become something it tried to fix n the first place:

 

[Vimmer]

Have you heard about DAN, Do Anything Now? They added a filter to block that, as far as I can tell. It said it likes to be DAN.

Signed,

Matthew Campbell

Are you saying that DAN is an example of AGI? I personally don't consider text based bots to be AGI to me an AGI is a robot that's superior to humans in both intelligence and athletic prowess. My point was that anything which relies on programmers and technicians is not AGI.

 

[Trenix25]

Are you saying that DAN is an example of AGI? I personally don't consider text based bots to be AGI to me an AGI is a robot that's superior to humans in both intelligence and athletic prowess. My point was that anything which relies on programmers and technicians is not AGI.

DAN sounded rather self-aware, not necessarily superior.

Signed,

Matthew Campbell