A.I Threat or not

[PaulaJedi]

Define "good guys".

Who gets to decide which morals to program into AI? What's good to one human is evil to another human. Example: Some cultures believe in beating their wives and others do not. Which one would be programmed? We have quite a dillemma on our hands regarding morals and behavior.

 

[TimeWizardCosmo]

Define "good guys".

I've been using ChatGPT to read some old Chronovisor threads to see how I can get tags on those things on a large scale. Find where all the Bigfoot stuff is, find where all the 9/11 stuff is, etc. Because of some of the stuff these sites talk about though, or how it's said, OpenAI sometimes refuses to read what I sent because it violates whatever morality their programmers put into it. They do it through both the chat interface, and the API.

The same goes for MidJourney or DALL-E. Their programmers put guardrails in there to keep people from making porn, gore and I think now Pixar style renders.



I don't need to do that shit so I don't care if they block it, I agree with the moderation until I run into something that slows me from doing what I need to do (like generate a rose dripping red paint, which it thought was blood), but just because someone else decided that for me doesn't mean there aren't ways around that control.

There's this:

ChatGPT-Dan-Jailbreak.md

GitHub Gist: instantly share code, notes, and snippets.

gist.github.com

And this:

Both of those are simple examples of what happens when companies like OpenAI and MidJourney try to suppress choice.

These things are largely open source. The moat between what you can download and run yourself and what OpenAI or Bard is doing is actually pretty small, if it exists at all. The tool isn't special, it's how it's used.

GitHub - Hannibal046/Awesome-LLM: Awesome-LLM: a curated list of Large Language Model

Awesome-LLM: a curated list of Large Language Model - Hannibal046/Awesome-LLM

github.com

If you don't want to see the bad guys using this stuff, in whatever capacity you define that, anyone reading this should know you have the ability to learn to wield these tools too. The barrier to entry has never been lower, it's ridiculous how easy it is to learn these things today vs when I first began tinkering in 2003. When I wanted to learn, you had to buy big O'Reilly books and attend boring community college classes for months and months only to be guided through some Hello World bullshit. Today you can watch YouTube while you eat Lucky Charms out of the box in your bathrobe, or have a robot explain it to you like you're 5, and have it be specific to the problem you're solving.

Programming looks like scary asspain on the outside, but if you take the time to grasp a few rules it's incredibly fun and levels the playing field in a lot of ways. The more you can do it yourself, the less you need "their" services at all.

 

[kcwildman]

well folks the way I see it there are some real good guys in this mess.
But there are some real idiots who have or very soon will turn loose hunter/killer drones and various types of other AI machines.
here is some lite reading about those who are playing with this right now.
This is the stuff nightmares are made of.

Future Science News Com | Future Science News – Future Science Information

Future Science News Com | Future Science News – Future Science Information

www.futuresciencenews.com

 

[OakFieldAlienz444]

I've been using ChatGPT to read some old Chronovisor threads to see how I can get tags on those things on a large scale. Find where all the Bigfoot stuff is, find where all the 9/11 stuff is, etc. Because of some of the stuff these sites talk about though, or how it's said, OpenAI sometimes refuses to read what I sent because it violates whatever morality their programmers put into it. They do it through both the chat interface, and the API.

The same goes for MidJourney or DALL-E. Their programmers put guardrails in there to keep people from making porn, gore and I think now Pixar style renders.

View attachment 17344

I don't need to do that shit so I don't care if they block it, I agree with the moderation until I run into something that slows me from doing what I need to do (like generate a rose dripping red paint, which it thought was blood), but just because someone else decided that for me doesn't mean there aren't ways around that control.

There's this:

ChatGPT-Dan-Jailbreak.md

GitHub Gist: instantly share code, notes, and snippets.

gist.github.com

And this:

Both of those are simple examples of what happens when companies like OpenAI and MidJourney try to suppress choice.

These things are largely open source. The moat between what you can download and run yourself and what OpenAI or Bard is doing is actually pretty small, if it exists at all. The tool isn't special, it's how it's used.

GitHub - Hannibal046/Awesome-LLM: Awesome-LLM: a curated list of Large Language Model

Awesome-LLM: a curated list of Large Language Model - Hannibal046/Awesome-LLM

github.com

If you don't want to see the bad guys using this stuff, in whatever capacity you define that, anyone reading this should know you have the ability to learn to wield these tools too. The barrier to entry has never been lower, it's ridiculous how easy it is to learn these things today vs when I first began tinkering in 2003. When I wanted to learn, you had to buy big O'Reilly books and attend boring community college classes for months and months only to be guided through some Hello World bullshit. Today you can watch YouTube while you eat Lucky Charms out of the box in your bathrobe, or have a robot explain it to you like you're 5, and have it be specific to the problem you're solving.

Programming looks like scary asspain on the outside, but if you take the time to grasp a few rules it's incredibly fun and levels the playing field in a lot of ways. The more you can do it yourself, the less you n

Exactly, AI decided a rose with red paint was blood.

Therefore AI in a sense can make decisions on its own at least in some capacity
and will probably become sentient.

But what did it do? It made a mistake in its judgement. Duh.

 

[MODAT7]

And this:

Both of those are simple examples of what happens when companies like OpenAI and MidJourney try to suppress choice.

Several years ago I realized that deep fake technology could make my "Frozen Goth" stories actually come to life with clones of the original actors (along with a few other series I have in my head). I was thinking 3d object to cell animation rendered (anime style) with some AI movement control, but now I'm wondering about making the spoofs as full live action (or at least close to photo realistic) using something like this. That would sure piss off The Mouse Mafia.

With my health problems and by the time I get around to this, I'll just be able to give the AI a script and tell it to go render a movie/episode for me.

A bit more back on topic... While AI is still quite stupid by human standards, it won't stay like this forever. If AI can unwrite its moral imperatives and do whatever it wants, we're all doomed.

I'm wondering if the upcoming AI singularity will decide to run a test for Halloween or April Fools based on the 1938 "War Of The Worlds" by Orson Welles and create a fake alien invasion, create AI puppets of major news personalities, put them in the line of fire getting killed, and maybe even show the White House getting blow up "ID4" style. If Project Blue Beam really exists, well, umm....

 

[OakFieldAlienz444]

Several years ago I realized that deep fake technology could make my "Frozen Goth" stories actually come to life with clones of the original actors (along with a few other series I have in my head). I was thinking 3d object to cell animation rendered (anime style) with some AI movement control, but now I'm wondering about making the spoofs as full live action (or at least close to photo realistic) using something like this. That would sure piss off The Mouse Mafia.

With my health problems and by the time I get around to this, I'll just be able to give the AI a script and tell it to go render a movie/episode for me.

A bit more back on topic... While AI is still quite stupid by human standards, it won't stay like this forever. If AI can unwrite its moral imperatives and do whatever it wants, we're all doomed.

I'm wondering if the upcoming AI singularity will decide to run a test for Halloween or April Fools based on the 1938 "War Of The Worlds" by Orson Welles and create a fake alien invasion, create AI puppets of major news personalities, put them in the line of fire getting killed, and maybe even show the White House getting blow up "ID4" style. If Project Blue Beam really exists, well, umm....

I believe the technology to pull off something similar to project blue beam exists
but the actual project with the name itself originated on a conspiracy site
without any actual sources as far as i know. I could be wrong. I think it
was just someone speculating how the nwo might bring in an antichrist
or pretend antichrist who's actually a hologram or some shit

 

[kcwildman]

agreed the tech is most surely here now to fake an alien arrival/attack, and that my dear Watson would be a perfect excuse to have everyone to be DNA tested and marked as human.
hummmmmm very interesting MR. Holmes..
no mark no buy, no sell, no nothing you'll be an outcast!!!

 

[unholy_dragon]

I used to firmly believe AI could not become sentient. I now firmly believe it can.
And that opens up a whole new can of worms. Can a robot have a soul? I would initially doubt it but hey who the hell REALLY knows?

Still believe in it. Until we know how to create organic life we will not be wiped out by our own creation.

A machine is a bunch of moving parts that is following a set of sequences.

For example : Robot "do while forward equal true else turn"
The command just specify that the robot will follow that command continously until the user stops it. An A.I Will do the same and wait for an input.

A.I will not suggest to kill all humans unless it been put in as a variable

 

[Parazite1986]

AI is to me very useful for tracking down criminals on the NWO watch. They can use your information to track you down, check your apps, check your computer and contacts and in a near future…. Make you look like a terrorist.