I have a genuine fondness for lisp, but this article lost my interest pretty quickly with the thought experiment. "Adding object orientation to C" doesn't require "the programming chops of Bjarne Stroustrup"; you can write object-oriented code in unmodified C (indeed, there are people who do exactly that). Nor is C++ the only way to extend the language to add convenience to doing so. Objective-C, for example, takes a completely different and much simpler approach. C++ doesn't even satisfy the stated requirement of "keeping them backward-compatible with the original languages, modulo some edge cases", unless "some edge cases" means "large parts of the language" (semantics of storage class specifiers and type qualifiers, what is and isn't an lvalue, rules for implicit type conversion, ...)
From a certain perspective, these are silly little details, but if you don't understand the silly little details of the other languages you write about, why should I listen to what you have to say about the virtues of lisp?
On he other hand, I didn't really need the C/objects piece of technical evidence to support the "Lisp makes metaprogramming trivial in ways that other languages don't" argument. That one kind of stands on it's own, so I could move on to the rest of the article.
The article wasn't about Lisp as much as it was about Not Invented Here syndrome. As others have said, the social and organizational issues become the major obstacle when your working with the greybeards and technical wizards that make the actual CS + CE questions a breeze
From a certain perspective, these are silly little details, but if you don't understand the silly little details of the other languages you write about, why should I listen to what you have to say about the virtues of lisp?