Programming is, and ought to be seen as, merely a means to an end – not an end in and of itself. A movement that gets programming into to schools by claiming code is itself a Good Thing will inevitably bring into existence a programming curriculum that ignores the important abstract stuff about coding-as-a-concept (the new ways in which computers can extend the mind) and focuses instead on the concrete "implementation details" of programming today (text files, the command line, and all that other baggage from 20 years ago with which programmers are still saddled).
Basically: the reasoning is important, the code isn't, and code.org is all about the code.
This is the essential piece I thought Bret made clear by highlighting the choice quotes from Seymour Papert.
Treating programming as a means to an end has very pragmatic ramifications for programming. It means we can question the value of our tools when they do not serve our goals. Otherwise we are the mercy of them (and entrench the status quo as state-of-the-art).
One question I've been grappling with is why it takes so many human-hours of work to produce conceptually simple programs.
Fair enough, but is Gates's claim that "learning to write programs stretches your mind" really that far off or that much worse? Or that "an understanding of computer science is becoming increasing essentially in today's world"? Maybe it's a trivial point, but is it as offensive as this link makes it out to be?