The reasoning
The person making the claim will first set certain standards for AI's capabilities in a number of domains, stating it can never reach them. Then, when counterexamples are provided, they raise the standards further and demand they also be met. They continue this, ignoring the pattern of success, so that AI output is conveniently always considered slop.
Example
“AI images will always be slop.”
“Here are some examples of really good AI images.”
“Yes, but it won't be able to make video that isn't slop.”
