AI Generated

[TorNorth]

part 4

[BlirGimp]

AI has promise, but I don't know how people can produce any coherent images with it. Yesterday, I struggled with the AI call screener used by Experian to simply change my address, phone and e-mail address without the AI engine able to understand the request "I need to change my address" and, to add insult to injury, the AI could not transfer me to a live human without being able to understand what i wanted. The AI chatbot gave similar non-help. And this is Experian, which should be employing state of the art AI.

How people can produce any images at all given the state of AI I've encountered seems magical to me.

[cluseb]

-1289

AI has promise, but I don't know how people can produce any coherent images with it. Yesterday, I struggled with the AI call screener used by Experian to simply change my address, phone and e-mail address without the AI engine able to understand the request "I need to change my address" and, to add insult to injury, the AI could not transfer me to a live human without being able to understand what i wanted. The AI chatbot gave similar non-help. And this is Experian, which should be employing state of the art AI.

How people can produce any images at all given the state of AI I've encountered seems magical to me.

A.I. DELIVERS
In Defense of A.I.
100% A.I. (except for the signature )

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

100% A.I. (except for the signature and the Hamas watermark)

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

100% A.I. (except for the signature )

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

THE DEFENSE RESTS.

[wulf]

['/quote]100% A.I. (except for the signature )

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

THE DEFENSE RESTS.
[/quote]

The bottom one is especially well done. Very much a turn-on, and thank you for sharing.

[LLL]

-1289

AI has promise, but I don't know how people can produce any coherent images with it. Yesterday, I struggled with the AI call screener used by Experian to simply change my address, phone and e-mail address without the AI engine able to understand the request "I need to change my address" and, to add insult to injury, the AI could not transfer me to a live human without being able to understand what i wanted. The AI chatbot gave similar non-help. And this is Experian, which should be employing state of the art AI.

How people can produce any images at all given the state of AI I've encountered seems magical to me.

A.I. DELIVERS
In Defense of A.I.
100% A.I. (except for the signature )

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

100% A.I. (except for the signature and the Hamas watermark)

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

100% A.I. (except for the signature )

To expand, click blue ---> https://ralphus.net/thegimpforum/downlo ... &mode=view

THE DEFENSE RESTS.

I dabbled a bit in AI myself, and it's great fun. Plus, if even someone who's a klutz at photoshop like I can get half-decent results, it indeed delivers. However, BlirGimp has a point with "coherent". A single image, no problem. But would it be possible to make a coherent series? To tell the AI to remake a scene as viewed from a different angle, keeping the details? Or, even harder, taking your middle eastern girl in suspension as example: could you get the AI to make scenes with the same woman, in the same dungeon, with the same torturers, but in different stages? One where she's zapped with a cattle prod, for example, legs kicking, mouth wide open in a desperate scream... or one where she has been lowered and is now curled up in foetal position on the floor, but still recognizably the same woman?

[BlirGimp]

I admit to being a complete dunderhead when it comes to AI renderings.

Can one of the AI wizards on this site produce an AI rendered interpretation of even a short complete Dolcett comic? For example, use one of the super-smart AI engines to render the short 15-page Housewife Hangup comic which has been around for eons as a line drawn comic? I've seen it colorized using Photopaint/Photoshop or the like and even produced my own 3d rendering of the comic using Poser and Comic Lite tools, but how about the super-duper AI engines that people seem to be singing the praises of these days?

Given my limited experience with AIs used for call screening and the like, I simply cannot imagine the time, sweat and effort that would be required to instruct an AI to simply copy and produce a decent rendering of a line-drawn image much less a complete comic.

[doe.1971]

Given my limited experience with AIs used for call screening and the like, I simply cannot imagine the time, sweat and effort that would be required to instruct an AI to simply copy and produce a decent rendering of a line-drawn image much less a complete comic.

AI could certainly do what you describe. The problem is that no one yet wrote an interface that would do what you want. Look at comics, identify panels, replace imagery with AI generated and reassemble the panels in a similar structure. Given time, we'll get more and more specialized and who knows, in a few years maybe someone will write an AI method to do exactly that. There's only a handful of people in the world like Lvmin Zhang or Manuel Schmid that can operate on such level.

[TorNorth]

GIFs of videos

[TorNorth]

Witches

[Bill8816]

I would love to see you Generate some torture images. Maybe some Inquisition images?