#2 - Embellishment and instructions in one prompt
Multi-Modal models are flexible when it comes to prompt specificity. You can mix detailed, exact instructions along with providing high level guidance on what you want it to think about and come up with on its own.
The prompt in this creation gives examples of:
- Giving instructions to generate a funny idea for a tweet
- Generating an image based on a specified layout
- Formatting one section as the output of the earlier instruction with appropriate features
- Formatting another section of the layout as a tweet with specific instructions
These models are still not perfect at writing text. Further, they still might add things you didn't ask for. Not every creation will be perfect - the technology is still advancing, but what an advance these new models are!