Comparing Different Flavors of AI Image Generation
Same Prompt, Different Browsers, Different Images
You may know (based on my recurring “Greatest AI Images” posts) that I enjoy generating AI images. I’ve noticed that the images created differ based on whether I’m accessing the Open AI-based AI engine in the Google Chrome browser or the Microsoft Edge browser. I get to the respective AI image generating pages for each by entering in those browsers the search term “Bing Image Creator.”
To record a concrete example of the different results I obtain in these two browers, this morning I fed both of them the same exact prompt, “Three impossible things before breakfast” which, as you may know, is a quote from “Alice in Wonderland.” The fuller quote is something like, “Why, sometimes I’ve believed three impossible things before breakfast.” The AI “robots” obviously made the connection with Alice, as you can see in the images below, but the resulting sets of images differ from each other in “look and feel” (style).
Here are the images the Edge browser (“Designer Image Creator”) generated from that prompt:
…and here are the images that the Chrome browser (“Copilot Designer”) generated:
Why the difference? I don’t know. Maybe each of them are based on the same Open AI software, but Google and Microsoft have put their own spin/tweaks on the code? That’s doubtless it (after all, Open AI’s code is obviously open source), but how many different versions will there eventually be, and which ones will be “best” for which type of images? Time will tell, I guess, but as of now, I normally prefer the Google set of images (but in the test case shown above, it’s the opposite — I prefer the first four, from Edge, over the Google images.
How about you?