• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

AI images we have created

I think if you publish it as an individual, no one will be bothered to go after you. If you try use it as part of a publication for a company you work for, legal will not let you. At least that's how it works around the company I work for.
Copyright and trademark infringement is more of an issue for a company that uses the Internet or with globalized market. AI images can otherwise be used by companies.
 
Before "computer" meant a machine, it meant a person; someone hired to crunch numbers by hand or with mechanical calculators. From the late 1800s through the 1950s, armies of human computers carried out the endless arithmetic that industries and governments required. Because calculation was seen as routine clerical labor rather than prestigious science, employers could pay women less than men for the same work and they did. So women filled these rooms.

Then came the shift. When programming emerged, it was treated as an extension of that same administrative work. The result? Many of the first programmers were women, hired at the same suppressed wages. Only later, once coding revealed itself as difficult and creative, did it become a high status engineering discipline and the field's demographics shifted along with its paychecks. Now, in an ironic way, the field is just going back to its origins.
 
AI is fun and all + a good search engine to find all those tricky little details you sometimes need, when you're a real DIY geek. But cats are nicer in real life... Women differ more. Purely statically that is, comparing to the amount of pictures of either cats or babes ;)
6723.png
 
The Krea 2 model is free to download and use. The AI models can be downloaded from huggingface.com. The software I'm using to run it, Forge-Neo is free and open source. It can be installed standalone, but it is easier to just install Stability Matrix first and then use that to download and manage it.

Yes you can specify lens types and f-stop to a certain extent even with the base models. For more advanced effects / lighting, etc. There are plugins called Loras, most of which are also free. There is a numbered series of Loras called Vision Vanguard that I'm trying out.

There are loras for different types of film grain, styles like Technicolor, 50s Scifi movie, etc. The best site for loras is civitai.com and most are also on huggingface.com. You can filter and search for almost anything. Those sites are also free, but require creating a free account.

As an example, here is the description for the Vision Vanguard 22 lora (most of the loras are around 218MB in size vs ~12GB for the AI models):

Vision Vanguard 22​

Film-grade realism. A masterly command of natural light and analogue texture.

The Aesthetic​

Vision Vanguard 22 is a masterclass in cinematic contrast and depth. Designed for those who demand uncompromising photorealism, this model specializes in the interplay between light and shadow. It excels at plunging scenes into moody, brooding blacks while highlighting textures with surgical precision. From pensive human portraits to vast, dramatic landscapes, VV22 grounds every generation in a world of quiet intensity and environmental storytelling.

Key Features​

  • Surgical Texture Rendering: Captures the physical reality of materials—wet stone, weathered skin, wrinkled fabric, and gnarled wood—with unparalleled detail.
  • Dynamic Lighting Mastery: Expertly handles high-contrast, low-key lighting, chiaroscuro effects, and dramatic artificial sources like stadium floodlights.
  • Cinematic Depth of Field: Delivers professional-grade focus management, from razor-sharp macro details to expansive, atmospheric wide shots.

Ideal Use Cases​

  • Editorial Portraits: Capturing the pensive and industrious expressions of human subjects with a "lived-in" feel.
  • Prestige Drama Stills: Creating scenes that look like high-budget film stills or brooding documentary photography.
  • Macro Nature & Industrial: Discovering the beauty in decay, from rusted gears to crystalline formations and bioluminescent life.

Getting Started​

To unlock the best results from Vision Vanguard 22, use the primary trigger word and descriptive natural language:

  • Trigger Word: VISION VANGUARD 22
  • Recommended Strength: 0.85 - 1.0
  • Core Strategy: Focus on material textures and specific lighting directions. Use the Specificity Stack (Subject -> Action -> Environment -> Lighting) to guide the model.t
I know huggingface, as I have downloaded some MLX models to my MacOS M1 with 32g. How much memory are using on your Mint Linux computer ?
 
As a guy who was a CGI/Photoshop artist for over a decade and over 15,000 professional hours as a Photoshop user for world-class clients, you have created nothing.
A person a century back would say the same to you with respect to use of Photoshop. Or an expert who retouches antique paintings with period paint and brushing techniques.

One of the bird photographers that I like, routinely adds cut off bird wings to complete the image using photoshop. I dislike that more than the AI images created by the OP.

Just because you use photoshop vs AI, to say nothing of CGI, doesn't mean what you are doing is honest. Effort put into a more crude tool doesn't make the work any more valuable.
 
Will you really choose a machine over a human ?
A machine is enabling us to have this communication. No human could do it.

I use a dough mixer for making Pizza. Should I be hiring a person instead?

AI will have an incredible effect on human race. Some of it is liable to be bad. But none of this calls for the argument you are making. Or the judgement call.
 
I have to say, these AI pictures are amazing. I can see the real life model industry coming to an end soon. The details are uncanny.
 
Her right heel appears to be totally uncompressed on the table with no displacement of the surrounding tissue

(I’m being critical ;))

The Krea 2 model can also do accurate text:

View attachment 544824

Not really a good render- lighting is off all over the place (shadows, face is slightly front-lit etc), her feet lack details, right hand is wonky, lamp should not be on in the day, tell tale AI coffee cup (not sure why they put coffee cups all the time - LOL).
 
A machine is enabling us to have this communication. No human could do it.

I use a dough mixer for making Pizza. Should I be hiring a person instead?

AI will have an incredible effect on human race. Some of it is liable to be bad. But none of this calls for the argument you are making. Or the judgement call.
The way we are communicating now via the internet is not AI its tools,code and computers designed by humans.
The internet as we know it is still 'old school'. The design of the internet is by humans. Maybe in the future an alternative internet could be designed by AI ?! An example of using AI here would have the AI system generating the discussion then responding with cold logic and inference, Using a bread machine, automatic coffee system, power lawn mower is still 'old school' tech.
 
Not really a good render- lighting is off all over the place (shadows, face is slightly front-lit etc), her feet lack details, right hand is wonky, lamp should not be on in the day, tell tale AI coffee cup (not sure why they put coffee cups all the time - LOL).

I asked for the coffee cup to be there, not an AI tell. There is nothing wrong with the right hand. I zoomed in on the original to make sure. The lamp was my fault. I had specified it was on because I had generated some night scenes and forgot to change it. I'm not an expert on feet details so I don't know what you're not seeing there.

I'd love to see a better render using local AI tools. Generate one and post it here.
 
I asked for the coffee cup to be there, not an AI tell. There is nothing wrong with the right hand. I zoomed in on the original to make sure. The lamp was my fault. I had specified it was on because I had generated some night scenes and forgot to change it. I'm not an expert on feet details so I don't know what you're not seeing there.

I'd love to see a better render using local AI tools. Generate one and post it here.

I can try - what was the prompt?

Also - I hope you didn't read that as a personal "attack" or anything - I like this stuff and have been playing with image models for quite a while and have formed opinions be them what they may, I take it all with a grain of salt - these are just my opinions, nothing personal and I hope it is not coming off as arrogant.
 
Back
Top Bottom