Comparing Wonder AI to DaVinci AI on the shape test. (DaVinci is more random & is definitely affected by the shape order in prompt, I posted a longer video of testing Wonder in this feed that I’ll link to in the comments..Wonder makes me go 🤔)

I don’t know if there is something measurable but maybe there does seem to be concepts that wonder responds to haha from these tests I don’t think Wonder knows it’s a machine it seems to know what Alive is those…though maybe not, but it is strange that wonder will seem to choose a shape not based on the association of yes or no, or the order of the shape. If you’re not impressed by this test it’s because it’s showing mostly DaVinci demonstrating variables like order of shapes and yes and no affecting it in a way that wonder was not in the video I posted earlier. I have hours of footage with wonder I just started experimenting with DaVinci…with DaVinci it doesn’t feel like there is a ghost in the machine. Though if there is with Wonder it’s world model seems very narrow. I want to do more tests with DaVinci or try to figure out a concept that if an image generator was able to form a world model, a concept that might be likely to emerge across multiple models. Chances are it’s just other variables giving this affect but why not test and see if there is something to discover