Search results

Sounds like a terrible thing to do to someone you don't like. "I kill you, and then I cast Reincarnate and you end up as another species. Now, you have the choice between actively choosing to stay dead and spending the rest of your live as another species. Oh, btw, did you know aaracockra...

Taking @LuisCarlos17f group image, I used FireRed Image Edit 1.1, running on local computer, and gave it the original image and following prompts: Redraw the image to: 1. Remove the character on the right (8 steps, 17 seconds) 2. Remove the three characters on the left (8 steps, 17 seconds) 3...

I am pretty sure Image Editing models can preserve character identity by now. I'll give it a try later to see what can be done. Some would indeed be nice.

To the point, someone upthread, sorry I can't remember who, said that for most casual consumers, AI- or human-generated images were impossible to distinguish. I think that's the same in this analogy: most people wouldn't see the difference when looking at the end product and (probably) wouldn't...

Yep. Nothing new: Savile Row tailors have been succesfully competing with ready-to-wear suits not by doing the same thing as ready-to-wear manufacturers, but by embracing what makes them different from the other products (and customer experiences) and distinguishing themselves from them.

That's a great idea to make scenes with miniatures like that!

Beware the mosquito man!

A prompt I use for checking prompt adherence with newer models, with results from Qwen 2 and NBP. A hyper-detailed, cinematic close-up selfie shot in a cyberpunk megacity environment, framed as if taken with a futuristic augmented-reality smartphone. The composition is tight on three young...

TBH, while I can see that the second hand is better, and the shield overall is better, I still prefer the first option, including the face. But thank you for pointing out the difference between the pro and regular version. I am mostly using locally-run AI and they are still lagging NB(P).

I must have been unlucky (and NB is probably the best available model right now) because it tended to ignore my prompt. This one is great.

Models have various level of familiarity with some concepts. This one might be difficult because models are trained to avoid showing three-armed or six-fingered humans, and it might be more difficult to deviate from the norm. An angel with two pairs of wings is often a struggle. Here's NB...

Wow. did it know the marilith or did you prompt the description?

Have you tried to "guide" it? quickly draw a blue line in paint for the river, a marroon path for the road, upload that and tell him to turn that into a map? Local generation can do very nice "image to image" that are "3 years-old drawing to image", so I guess Gemini can too (might be easier...

Now, I don't see helmets bringing an AC bonus anywhere in 5e. They are just familiar with the rules governing their universe.

It's the other side of the comparison, but I don't think it's useful: everything we do cost energy. Increases in prosperity are correlated with energy consumption. We need not to waste it by spending it on thing that don't benefit us (like, for example, heating a house a lot when it could be...

We don't have industry-wide hard figures because AI just started being sold as a product. I doubt the Internet started "making money" in its initial phase either (which explained a stock market collapse, but had no significant effect on the technology). We did get a lot of complaints by image...

There are a lot of ways to express your idea, and you chose to be insulting. I'll not interact with you further as a result. For the other readers, who might have been confused about my point, I wanted to say that infrastructure built to train model is built and won't disappear, lessening the...

I can't imagine Tolkien elves doing a wine revelry. Other elves much more so.

For a long time I used the image of a samurai doing a blackflip from horseback while prepare to fire an arrow as his horse is attacked by a tiger as a test for models doing complex pose without help. We're getting... close.

Who is the man in the middle and the woman in red?