In his talk The Future of Programming, in the third section of the talk about "programming using goals", Bret Victor says this:
What won't work, what would be a total disaster, is---I'm gonna make up a term here---an API. This notion that you have a human programmer that writes against a fixed interface, that's exposed by some remote program. First of all, this requires the programs to already know about each other, right? And when you are writing a program in this one's language, now they're tied together, so the first program can go out and hunt and find other programs that implement the same service they're tied together. If this one's language changes it breaks this one. That is really brittle, it doesn't scale and worst of all you have... It is basically the machine code problem. You have a human doing low-level details that should be taken care of by the machine. So I'm pretty confident this will never happen. We are not gonna have APIs in the future. We are gonna have is programs that know how to figure out how to talk to each other and that's going to require programming in goals.
The talk is themed as a 1970's scientists view of the idyllic future, building upon and extending contemporary research ideas. Yet I found no references to research regarding this API bashing part. There is an accompanying website to the talk, which, for some reason, has absolutely no further references either.
I wish to know whether the idea of replacing APIs with "something better that does not take so much human time to work with" is something grounded in actual research, or it is an improbable wish, or if there are reasons it is impossible (I think it is).