RIB Images: There’s No Thing Like the Real Thing
That said, my journey with AI image generation has been a fascinating ride of trial and error.
During my creative journey (I didn’t know I had one), I focused on RIB (Rigid Inflatable Boat) images.
I also worked on mug designs and birthday cards. Gradually, I expanded into video content which is only just starting to pick up.
The process of creating with AI is rarely straightforward—prompt refinement is essential to match the intend.
While ChatGPT offers helpful starting points, further tweaking is always needed.
I’ve explored several AI tools—Adobe Firefly, Google Gemini, Midjourney, DALL·E, Leonardo AI—each with unique strengths and quirks.
I know there are more tool but that’s not the focus of this blog post.
Currently, Gemini by Google provides the best RIB-related results.
However, its persistent square format requires post-editing in Photoshop. This editing is necessary to achieve a 16:9 aspect ratio.
I need to generate extra content to fill in those gaps between square and 16:9 which is a challenge.
DALL·E initially struggled significantly with RIB images. It often produced awkward, crazy, out of this world, unusable results.
This highlighted the complexity of the “RIB formula” and the insufficient training of the language model.
Still, it’s rewarding when everything aligns.
I must ad that he environmental impact of generative AI is something to be aware of.
An MIT article on this topic is particularly insightful. It’s worth reading.
Here are my funkperiments.
