Tested out GPT's image processing feature a bit. Not sure if I’m using it the right way? Pics 1234 are the original image, GPT, Grok, and Gemini under the same keyword comparison. Based on the results, I personally feel GPT is noticeably more powerful, especially with depth of field processing. Gemini’s banana model isn’t bad either, and while Grok’s preview looks a bit rough, the full image looks much better. Plus, I’m using the completely free version of GPT, while Grok and Gemini are paid versions. The keywords also came from GPT’s 'preset templates,' just a simple description can help the model refine further. Is it time for Prompt Engineers to start thinking about new career paths?