Trying to use #chatgpt to generate a #graphic of two flags. It actually did a lot better than I would have expected, but it consistently gets the left flag not quite right. When I confronted the #chatbot, it agreed it was wrong, and eventually blamed "the #generator" for not doing the right thing. It then asked if it should "edit the #image."
So... the descriptive image generation of ChatGPT is a different application? #ai #llm