Yes, this is it. We have very uneven abilities to make things look extraordinarily good with computers while our abilities on some other fronts have hardly progressed at all in the past 20 years, while others have had much more believable curves. Speech synthesis, for instance, is terrible by this standard. We have absolutely no ability to synthesize this voice with a specification much smaller than someone feeding it a full voice performance anyhow.
Personally I think it's a grave error to measure our progress on these fronts by getting distracted by the glitzy graphics... we are sooooo much better at those than anything else that they provide a very skewed view of where we are. You can see the same effect in video games; graphics that are orders of magnitude better than ten years ago, game engines themselves are only incrementally better in terms of capabilities and games themselves only incrementally more complicated.
(And to be clear, I'm not saying we have made significant progress in a wide variety of fields, because we have. I'm saying measuring our general progress by the subfield that has been an extreme outlier doesn't give a good view of that progress.)
Personally I think it's a grave error to measure our progress on these fronts by getting distracted by the glitzy graphics... we are sooooo much better at those than anything else that they provide a very skewed view of where we are. You can see the same effect in video games; graphics that are orders of magnitude better than ten years ago, game engines themselves are only incrementally better in terms of capabilities and games themselves only incrementally more complicated.
(And to be clear, I'm not saying we have made significant progress in a wide variety of fields, because we have. I'm saying measuring our general progress by the subfield that has been an extreme outlier doesn't give a good view of that progress.)