A co-developer (@Ilya) and I worked on a recent project that was generating a QR code onto an AI generated image of a city that was requested by the user. Sounds simple right? Not exactly.
We used DALL-E by OpenAPI and while we had high expectations we realized that DALL-E was kind of cooked ๐
. Even with clear-cut directions it would generate an image of the city from far away even when we would say "zoomed in view please". This resulted in multiple hours of testing and prompt engineering to find the best prompt to get DALL-E to understand us.
Do you know any better AI-image generators with APIs that might trump DALL-E? Be sure to link them below :)