fossil kraken
Yesterday, I had a bit of trouble when creating an image because DALL-E3 divided the screen into different frames. Harajuku and Shibuya in Japan are close enough to walk or cycle between, and of course, there are trains and cars too. They're not that far apart, so there was no need to split the screen, but DALL-E3 did it anyway. Even when I explained, 'No, they're close enough to walk to,' ChatGPT wasn't convinced.
Have you ever had such an experience?
So, I had no choice but to go to the Japanese Wikipedia, copy the explanation that Harajuku belongs to Shibuya Ward, and tell ChatGPT, 'This is how it's explained on Wikipedia.' Finally, ChatGPT understood.
Based on this experience, I had an interesting conversation about how the relationship between the user, ChatGPT, and DALL-E3 is like a team, and I wanted to share it here.
Since you use generative AI, you might find this interesting. Please take a look if you like.
Note:
This is just a conversation between me and ChatGPT, so it's not professional, and there might be some mistakes.
