Author mentions web browsers and extensions in the context of barriers to creating a successful malleable system.
Surely the web browser - considered in relation to web pages & technologies (HTML, Javascript etc), rather than extensions - is perhaps the most successful malleable system in history.
The core browser "functionality" has been and can be extended in an almost infinite number of ways via web "pages". You could consider his dichotomy between "foundational" features (unavailable to the user/tinkerer) and extendable features to be comparable between Emacs (C vs elisp) and the Browser (C++ [for example] vs HTML/Javascript/CSS).
I realise that to "extend" the browser by building a web page includes a great deal more friction than playing with elisp in scratch. Does this rule out the browser as a "malleable" system?