Raw LLM Responses

@kaotiskhund Exactly right, I had many bad iterations, as in possibly hundreds. I needed to interact with Photoshop's UI to select the average colors of the color target image to create a soft-light gradient layer with the colors and luminosity of the target image. However, there are heavy limiations with Photosho's scripting language with some features available only via selection in the UI; I had to find a creative way to get the average three colors of the target image. I reduced the target image to three colors via Photoshop's save feature, I just needed to save it as a GIF with only three colors. At first ChatGPT thought it was impossible to automate a process to extract those colors, since scripting a pixel-scanning feature would be location-dependent, and every image is different in sizes, resolution, etc... I figured out that I can create a photoshop action to reduce color image to one pixel (thus eliminating the location issue ) three times for each color. Then I scanned the pixel which effectively updated the MRU color list (temp memory for recently used colors in Photoshop's UI), then I wrote another code to extract the information from the MRU color list for last three colors recorded, meaning, the last three pixels that were scanned via the photoshop action. The scanning MRU color list feature is a code I borrowed from a random post that shared how to do this. Solving this allowed me to create (with no other input) a soft-light gradient adjustment layer with the average target colors and correct corresponding luminosity, thus automating a normally necessary step of hand-selecting a color via UI. There are many other processes in this action-script, but that was the biggest challenge! One example of the many times I had to examing the code was when I noticed that the program wasn't mapping luminosity correctly in the gradient layer. I realized that though in the UI the colors are based on 0-100, the raw data uses a different scale, I think up to 1600. I had to analyze the code to understand the differences and correct the code with simple arithmetic to get the numbers converted correctly.