What Would Concept Artists Want From An AI Tool? 

I think you are right about the fact that it can become a powerful tool for artists to support them and do the tedious work. On my side I think I will continue to use it for moodboards as part of the research stage for now, see what future brings.

I'm an artist as well, I'm not sure if you've tried it yet but several of the features you've wanted to see explored are currently available in stable diffusion. I've actually used it already for some of these purposes. You can already do a very simple under 10 minute painting and through running separate parts of it through stable diffusion's image to image feature you can collage them into a profesional looking completed piece within an hour. The image to image feature also does in fact have a slider which determines how much the image generation is affected by the image you put in versus the prompt. If you set it so that it's focusing more on the prompt it will give you iterations that are very different from yours. I've already had some luck using it to generate some weapon concepts in this way. While if you set it very high in addition to words describing the level of finish and kind of brush work you're after it can very quickly complete a rough painting. It will also preserve the style and proportions present in your original sketch. Stable diffusion also has the advantage of running locally on your own hardware and does not require an internet connection. It is available completely for free. And a number of people have offered graphic user interfaces that can run it with completely automatic installation just like any other program. It it does run off of your GPU so you do need a fairly recent gaming PC or workstation to run it It's first one of these to give me some optimism, that these can be used as tools. As the results you can get through this are substantially better than the results you can get through just using a prompt It can't yet do orthos though haha

Great ideas here Neil 👍 Side note, I've heard the main guy at Midjourney state specifically that he is not interested in making a 'tool with specific features for professional artists'. However, I'm sure there are other devs out there who might want to do it and this is a great wish list you've put together.

Thank you for sharing this. This is exactly how I have been trying to use it. As a tool rather than a replacement for me altogether.

I think this tech will ultimately evolve into like a full 3d holodeck where you will be able to generate anything you want and have the ability to fine-tune everything from the smallest detail to the overall style. That said, the human factor will play a major role, and at the end of the day it's still about how creative and visually sophisticated you can be, whatever tool you use.

Sure, there's a lot of fear regarding this new tech and how this impact a whole market, it is similar to what happened with photography and with music decades ago. Again, there were a lot of beginners and mediocre pros that were left behind during this changes, I guess it's almost some sort of survival of the fittest scenario, but at the same time it gives power to people who don't got the skills or the time to develop it, it enhances the power of those who already are into art, so I agree with you Neil, embrace it and use it at its best. One of my main frustration on the late years was the expense of having to trade too much time to make personal art, mostly because of some technical stuff like rendering and painting, so when you come down with an solution like midjourney, it's like giving me half the job done and allowing more time to have ideas than to waste with stuff that is purely technical.

It would be useful for me to compile folders of my own images and illustrations to train some tool with. Probably number one for me.

lots of food for thought, thanks for sharing!

This is such an interesting video and I m definitely gonna be watching your other content on this subject!!!

transparent background, to paint fill in spots were AI missed , to have the program on my computer and I do the rendering my self and render what I want, things are getting better and quick, soon AI will do everything .

It's sad that this doesn't have more views. I really wanna see more thought out responses to AI in regards to art, not just straight up fear and rejection. I really feel like AI has the potential on cutting down work so artists can benefit and focus on the details they like

I'd like to see small automated addons into major packages, like 3ds max, Rizomuv, Topogun etc, so they would automate some functions

excelent video. Thank you :)

I would love some of these features. I keep seeing AI advancements in ‘two minute papers’ channel but no real advancements in user access to these projects.

You should really look into Stable Diffusion. Open source AI. You would be able to make these plugins for Photoshop ect. Look up Alpaca AI as an example of what is possible.

I have experimenting a bit with dalle, midjourney and stable diffusion, while it's great at generating more abstract things such as vegetation, rocks for environment, it really isn't good at coherent design, in fact it cannot to do coherent design, anything like line art, vehicle design, ships, weapons etc it cannot do well at all. I think artists will be fine, these AIs can generate some neat fantasy images, but they're very far from the level of complexity of a "hand crafted" art style that an artist has taken years to develop, these algorithms cannot achieve coherent nor creative composition, lighting, texture, etc it's very far from matching the art style of artists who have devoted years to develop their style. I am sure AI will improve and have more refinements and fine tuning to get there, but I think great artists will be just fine, most algorithms right now are a one trick pony, that said they can serve as a nice tool for mood boards, inspiration and iteration.

ai that can detect anatomical errors or shading errors to tell me what I'm doing wrong would be cool

hard surface texturing by AI instead of Neil 😀

Imagine how 3d artists will feel in a company where 2d artists and concepters work 100 times less and get paid the same or even more?

Buggy whips are still buggy whips.

You can't adapt to something that changes every day... can only comfort yourself with the fact that knowledge itself will not become outdated, but the usage scenarios derived from it will change.

i hate A.I art