Experience the ChatGPT ‘Make It More’ Trend and Create Outlandish AI-Generated Images
AI art generators occupy a unique niche. They’re capable of rendering nearly any concept you can imagine, ranging from a skateboarding dog in outer space to a cup of coffee adrift in the ocean. Setting aside the ethical considerations of AI-generated art, it’s often the case that these initial creations don’t quite meet expectations. To refine the outcomes, it’s necessary to provide the AI with specific prompts for adjustments, tailoring the final images to suit your preferences.
What if your aim isn’t to craft a high-quality AI artwork, but rather to create something wildly imaginative?
This is the essence of the “make it more” trend. Users of ChatGPT are experimenting with DALL-E, initially requesting it to generate an image, and then challenging the bot to amplify a certain aspect of it. Consider this instance from Justine Moore: DALL-E was first asked to produce an image of a bowl of ramen. Subsequently, Moore requested it to be made ‘spicier’. The AI responded, primarily by adding an abundance of peppers. When prompted again to increase the spiciness, DALL-E escalated things by igniting the bowl, seemingly transporting it to a fiery, pepper-laden realm. By the end of this creative journey, the ramen bowl was depicted emitting fiery beams into the cosmos, embodying the epitome of ‘spicy’.
Numerous instances of this trend can be found online for your enjoyment, ranging from Mashable editor Stan Schroeder’s experiment with an oversized water bottle to a bodybuilder becoming increasingly muscular. Should you wish to explore this trend personally, it’s important to be mindful of certain limitations.
Guidelines for Using the “Make It More” Trend with ChatGPT and DALL-E:
1. Be aware of prompt limitations: DALL-E, as part of the GPT-4 suite, has a cap on the number of prompts you can issue at a given time. OpenAI’s notification system for reaching this limit isn’t always clear, so it’s wise to moderate the frequency of your experiments to avoid a temporary lockout.
2. Understand DALL-E’s response patterns: DALL-E can be somewhat unpredictable when handling these specific types of requests. It’s possible that OpenAI has made adjustments following the popularity of this trend. For example, in one experiment, I successfully prompted DALL-E to generate an image of a dog running through a field and then to make the dog faster. However, when I requested to speed up the dog again, DALL-E declined, indicating that it had already fulfilled the request for a fast dog. This suggests a certain level of specificity and limitation in how DALL-E interprets and responds to iterative requests for modification.
I experimented once more with the dog scenario, initially requesting the creation of the world’s fastest dog and then challenging it to be even faster. DALL-E declined the second request, reasoning that it had already depicted the fastest dog possible. A bit of an oversight on my part.
However, I found better success with a different approach. I asked the AI to generate an image of a cup of coffee and then repeatedly requested it to make the coffee hotter. Initially, the AI struggled after a few tries, but eventually, it managed to produce around five versions of increasingly “hotter” coffee cups. The experiment reached its peak when the AI couldn’t visualize heat any further, and the final cup’s appearance was reminiscent of the Trinity nuclear test.
Thank you for reading this article so far, you can also get the free prompts from here.
https://www.buymeacoffee.com/yukitaylorw
What Will You Get?
- Access to my Premium Prompts Library.
- Access our News Letters to get help along your journey.
- Access to our Upcoming Premium Tools for free.
Also, Check out https://solan-ai.com
Subscribe My NewsLetter now!
