That's exactly it! People are acting like DALL-E is some monolithic thing that can't be fixed or tweaked.

There's no reason to throw up our hands and say "Well, the world sucks, so of course this thing sucks!"

No, the output of a model like DALL-E is based on the biases in the data on which it was trained. That's something that humans can readily modify.

When we see an example of inputs that need to be modified to avoid bias towards any group, we need to call those out. It's a programming fix, not a societal one.