AI is being forced down our throat like the new Meta, and VR world, but it’s not just a trend. It’s a tool, deployed so rapidly, it’s now a race between giants for profit: Google Duet/Bard, Microsoft 365 Copilot/ChatGPT/OpenAI/Stable Diffusion, and so many more AI companies/tools. ChatGPT/OpenAI is bigger than Linux, and it’s open source. WordPress and Android, are also open source, proving an open source model scales faster. We’re adapting to AI faster than the IPhone due to free onboarding, with AI in daily tools, like gmail (still in beta) and the Google browser, which provides you with more info for your search. eBay is releasing a generative AI tool so that when you upload an image, it creates a product listing which will make it easier for new sellers. Why are AI tools free to low cost? So they collect more user data to train AI faster, and like all good tools, will come with an increasing price tag so this is a good time to try many AI tools. Regarding images, not just the 2 billion images fed to OpenAI/ChatGPT, Shutterstock has a 6 yr contract, giving OpenAI access to train on their images and integrate AI tools on their platform. I do question that AI is keeping us online, on a computer platform, using services like Google, and Microsoft, just to compete, and scale business. COVID forced businesses to be online and social media. Now it’s AI.
Testing generative AI: As an artist/graphic designer, I started out with #promptengineering images. I asked #LinkedIn to add it to job skills. For image prompts, I started with #DiscoDiffusion and Google Colab, moving onto other Diffusions, MidJourney beta, and Dall-e. I was the 1st of 5 people to offer AI image prompts service on Fiverr, just to test the software. I couldn’t get AI to create a good hobit hole from Lord of the Rings, surprisingly, so niche categories are difficult, but after 20 hours, 2 days of AI learning, I could make a pattern design I was happy with. Just looking at my current AI designs and last year’s, it’s incredible how much better the AI software tools are in just 1 year, generating better images.
I was extremely successful using AI in pattern design contests, where customers buy print on demand cloth of your image. Last year, in one month, I generated 270 pattern design images an hour, based on contest themes. Just to label and file all these images needs a tool, due to the sheer volume of production output so I predict more software will appear geared for that. Using Generative AI is like creating a painting because as you change the image prompt, you see it become closer to what you’re trying to create, and then you pick the best images and build on that. This doesn’t replace artists. It’s just another tool. I still spent time to adjust prompts to get a good image, use another AI tool to reformat, and edit in Photoshop. It’s a process. Entering #Spoonful #patterncontest each month, was so much fun, creating completely outside the box, new patterns, like a ‘cat ear backpack with books,’ based on Dark Academia theme, or ‘3 Picasso cat faces.’ Seeing results, how many people liked my pattern, made it a good test environment. After testing the most complicated image programs, I ended up using an easier AI software for production that generates 9 images based on my prompt in 2 min, so that I could get better designs, faster. With many tools emerging, not just researching, but testing AI tools for what you need should be your focus. Also, #spoonflower no longer allows pattern designers to use AI in their contests.
sumanidesign 