It's the year of AI, and I've been playing a lot with text-to-image generative AI tools. I've found Bing Image Creator to be the most accessible because it is free to use (you get 15 credits which renews each day). It also uses DALLE 3, which is pretty impressive, I am slow to say it's better than Stable Diffusion but there's a nice comparison at the Zapier blog if you're curious. Ultimately it depends on what you want to create and what aesthetic you are after.

Workflow

I discovered Krea.ai which has real-time generation, and a pretty good upscaler. Leonardo also has real-time generation, generally I think all tools should offer it as a way to get familiar with prompting.

I would usually sketch out a prompt with the various real-time generators and then copy, and refine the prompt to an on-demand generator like Leonardo AI or Nightcafe or Bing Image Creator and generate images there, tweaking it to more closely represent the image I want to generate.

I started an Instagram account where I post some of the better generated images. It acts like a showcase of what I've been trying out. I had a little scare today when Instagram sent an email about my account suspension. Thankfully, they just wanted to make sure I was not a bot, which is quite funny because I get a lot of 'send pics' bots every time I post a new picture.

At the moment I'm trying different styles and experimenting but I would like to focus on a particular style or a handful of styles and generate images exclusively in that style.

Thoughts

I really like the real-time generation of Krea and wish every other tool offered it as a feature. It really is great for sketching out an image.

More and more it's getting easier to create impressive images without knowing the right keywords to use, as these models mature, I expect the tools to generate images will become smarter.

What are you doing with AI?