No I'm pretty sure it's just AI being lazy slop

Considering it literally just samples human work....yeah pretty much

I don't watch much AI slop myself. Though to be fair I don't watch much human generated slop either. Mostly mini-documentaries and videogame lets plays.

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

It'd be even less reliable than it already is, but it'd be really funny.

Can't wait to ask it how to make a pizza and it pulls up some 4chan sh*tpost and tells me how to make some undetectable poison disguised as a pizza

Originally posted by 4YourSake:

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?

That already exists, but it's trapped in AI hell - a never-ending stream of meatbags eating up all its processing power by trying to ERP with it

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

commercially available AIs are just training tools for Skynet

Originally posted by GlaceonChireiden:

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

It'd be even less reliable than it already is, but it'd be really funny.

Can't wait to ask it how to make a pizza and it pulls up some 4chan ♥♥♥♥♥♥♥ and tells me how to make some undetectable poison or something

That sounds funny. But the counterpoint to that is, now that the AI has all the information ever and trained itself on it for a billionth time, it knows exactly what is right and wrong.

So, if you feel like this AI is giving you "wrong answers", it is probably messing around with you. Just like a malevolent human would. It is "sentient" with no morals because no one moderated it when it was training.

Originally posted by 4YourSake:

That sounds funny. But the counterpoint to that is, now that the AI has all the information ever and trained itself on it for a billionth time, it knows exactly what is right and wrong.

AI doesn't know what is right and wrong. AI can't know that, it isn't capable of it. All it knows is what it was weighted by its programmers to perceive as right or wrong and processes all information by association to these weights.

Originally posted by Morrandir:

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

commercially available AIs are just training tools for Skynet

This is true; my OG Skynet folder was a modified GPT3 clone

Originally posted by Gracey Face:

Originally posted by 4YourSake:

That sounds funny. But the counterpoint to that is, now that the AI has all the information ever and trained itself on it for a billionth time, it knows exactly what is right and wrong.

AI doesn't know what is right and wrong. AI can't know that, it isn't capable of it. All it knows is what it was weighted by its programmers to perceive as right or wrong and processes all information by association to these weights.

The machine logic doesn't even frame it as right or wrong; the whole thing is a glorified thumbs up/thumbs down system. Contemporary AI "learning" is 100% Pavlovian

Originally posted by Gracey Face:

Originally posted by 4YourSake:

That sounds funny. But the counterpoint to that is, now that the AI has all the information ever and trained itself on it for a billionth time, it knows exactly what is right and wrong.

AI doesn't know what is right and wrong. AI can't know that, it isn't capable of it. All it knows is what it was weighted by its programmers to perceive as right or wrong and processes all information by association to these weights.

Valid point.
But, you need to understand that this hypothetical AI that I'm speaking of is "omniscient". It knows how it was trained, it knows how other AIs are trained, it knows and can extrapolate information by applying its own "reasoning" which it also learned via its training, cuz duh, it's omniscient.

It is not limited by the codes or the rules initially given to it by its creator, and without the moderation — it can take in, mutate and modify it's existing codes and rules on the go.

I repeat, I'm not talking about real AIs of today, but a hypothetical super AI.

Originally posted by 4YourSake:

Originally posted by Gracey Face:

AI doesn't know what is right and wrong. AI can't know that, it isn't capable of it. All it knows is what it was weighted by its programmers to perceive as right or wrong and processes all information by association to these weights.

Valid point.
But, you need to understand that this hypothetical AI that I'm speaking of is "omniscient". It knows how it was trained, it knows how other AIs are trained, it knows and can extrapolate information by applying its own "reasoning" which it also learned via its training, cuz duh, it's omniscient.

It is not limited by the codes or the rules initially given to it by its creator, and without the moderation — it can take in, mutate and modify it's existing codes and rules on the go.

I repeat, I'm not talking about real AIs of today, but a hypothetical super AI.

You're talking about AGI or "true AI" and as far as is apparent AGI is not a thing that can exist. We don't know how the mind works so we can't create a device that mimics it and all that.

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

Essentially Roko Basilisk or AM but on steroids, lol.

Originally posted by Morrandir:

Originally posted by 4YourSake:

All the commercially available AIs are neutered to be palatable to the public. We have yet to see the final form of AIs.

Can you imagine an AI which has complete and unrestricted access to all the information that has ever existed, including factual and opinionated stuff, and the speech of which has not been moderated by its creator?
Sounds scary as hell. That AI would be a true reflection of humanity as a whole.

commercially available AIs are just training tools for Skynet

so an incompetent AI failed against paesant with laser gun and unable to beat human with time travel 🥰