Hacker News Reader

Comment by danielvaughn | Hacker News Reader

I'm sure the novelty wears off. But I'm already coming up with several applications for it.

On the personal side, I've been getting into game development, but the biggest roadblock is creating concept art. I'm an artist but it takes a huge amount of time to get the ideas on paper. Using DALLE will be a massive benefit and will let me expedite that process.

It's important to note that this is not replacing my entire creative process. But it solves the issue I have, where I'm lying in bed imagining a scene in my mind, but don't have the time or energy to sketch it out myself.

View on HN

>I'm an artist but it takes a huge amount of time to get the ideas on paper.

this is what I really like about DALLE-mini, it's ability to create pretty good basic outlines for a scene. it's low resolution enough that there's room for your own creativity while giving you a good template to spring off from. things like poses, composition of multiple people, etc.

I've used AI to try out different composition/layout possibilities. Sometimes it comes up with an arrangement of objects I hadn't considered. Sometimes it uses colors in really interesting ways. Great jumping-off point for drafting.

I'm in exactly the same boat. I got tired of waiting around for openai to take me off their waitlist and used DALLE-mini (now craiyon) to generate large batches of concept art for a project I was working on. I picked the ones that, despite being low-res blobs, conveyed the right mood or had an interesting composition of elements. I then layered my favorite elements of those and painted over, adding details wherever I wanted, and came out with something much better than I would've been able to make alone.

I've been having a blast using it in my dungeons and dragons games. If you type in, say, "dnd village battlemap" it's really pretty usable. Not to mention the wild magic weapons and monsters it can come up with.

I did notice it is very good at making small pixel art icons/sprites.

Do you have any tips or some prompt examples?

I’ve been using generative models as an art form in and of themselves since the mid/late 2010s. I like generating mundane things that bump right up along the edge of the uncanny valley and finding categories of images that challenge the model (e.g. for CLIP, phrases that have a clear meaning but are infrequently annotated).

Generating itself can be art. I’m not going to win a Pulitzer here, it’s for the personal joy of it, but I will certainly never get tired of it.

Reply on news.ycombinator.com