It's interesting to think that it's the increased complexity of modern programming that undid Hypercard, and with sufficient advancements we managed to build APIs which have now recreated some of that building block love. Interesting to see the U curve.
love it.
It might shake out that language models are overblown as an engine for machine _agency_
but wildly successful as a compatibility layer for machine communication.
Maybe the APIs of tomorrow will be plaintext correspondences - formally written requests from one language model to another.
It's interesting to think that it's the increased complexity of modern programming that undid Hypercard, and with sufficient advancements we managed to build APIs which have now recreated some of that building block love. Interesting to see the U curve.
So Hypercard was a failed attempt at Emacs.