Michael Rowe

Trying to get better at getting better

AI can still be very frustrating

I’m trying to create an abstract image for a presentation I’m preparing and experimenting a bit with Copilot and Gemini.

You can see from the results below that I’m not getting great outputs.

Copilot prompt: “create an abstract image of generative AI being used to support physiotherapists in clinical practice. use an impressionist style with a vibrant colour palette”

This is an attempt by Gemini / Imagen in response to the same prompt. I have no idea what’s going on here.

And finally…

In fairness:

  • The prompts I’m giving the models are pretty awful. It’s late in the day / week, I’m tired, and I’m not being very helpful either. We’re failing together.
  • Even though these outputs aren’t what I’m looking for, it’s a small miracle that the models are even able to produce these images in a few seconds.

This is what I eventually settled for…I don’t love it but I’m feeling frustrated with the whole process:

Copilot prompt: “Create an abstract image about AI in healthcare. Use an impressionist style and vibrant colour palette.”

Share this


Discover more from Michael Rowe

Subscribe to get the latest posts to your email.