Not all AI are equal.

Google’s Gemini AI, even the Advanced option, leaves a lot to be desired when compared to Microsoft’s Copilot.

Here’s the prompt I tried on both:

generate image in style of emoji that represents Nichiren Buddhism

Gemini’s first try generated these images:

Gemini’s second try generated this error message:

I’m a language model and don’t have the capacity to help with that.

Gemini’s third try generated these images:

I then fired up Microsoft’s Copilot, which utilizes DALL-E 3 to generate images.

The first time I entered the prompt “generate image in style of emoji that represents Nichiren Buddhism” in Copilot I got these images:

The second time I entered the prompt “generate image in style of emoji that represents Nichiren Buddhism” in Copilot I got these images:

The Designer app inside Copilot even lets you tweak the image with different styles:

This:


Becomes these:

That’s enough of this. I was originally going to discuss how Gemini and Copilot allow you to narrow the sources used to answer prompts. For example you take “what is the meaning of the Daimoku in Nichiren Buddhism” and modify it by adding “Using website 500yojanas.org as principal source, what is the meaning of the Daimoku in Nichiren Buddhism.” Again Gemini and Copilot offer vastly different answers. Maybe I’ll get around to writing about that eventually.