It seems like no matter how much neural networks accomplish some will always still take the Searle 'not real intelligence' stance.
Why are standards for what counts as intelligent with programs so much higher than what counts for people or other biological organisms?
I personally think programs where the programmer writes how to learn to accomplish a task rather than how to directly accomplish a task ought to count as intelligent. There are ways in which organisms still have advantages over AI but.vice verse is also true e.g. the greatest human chess players are outperformed by the greatest chess engines today and it's not even close.
Pic related: I find the program generated art and music nowadays to be really interesting and deeper than the more direct formula approach from past programmatic art projects.