http://www.paulgraham.com/lisp.html which you're probably already aware of
Paul Gragam's Lisp books were my second lisp books after a class using SICP, so I'm very fond of them, but his view of lisp is seriously idiosyncratic and wouldn't have gotten nearly as much traction as it did if he hadn't lucked out in the first dotcom boom, been a celebrity VC in the second and been one of the few authors of lisp books outside of the AI ghetto while AI was still a ghetto. Pete Seiibel's book and Edi Weitz's are a better looks at Lisp as a practical language without the vcistan bullshit. Paul Graham draws a ton of people to lisp, but no one who writes a lot of lisp writes lisp like Paul Graham. This is a microcosm of Silicon Valley's influence on the craft of programming.
Does this article put forward an assumption that the Lisp is less abstract than "Turing machines"? If so, I'd like someone to explain to me how, because this article confused me completely. And it's the fourth time I read this thing today. lm? Other CS mofos? ;) Arse shelter/background: Yes, I know who's Alan Kay. I'm not questioning his authority or contributions, just don't get what he's on about. My own experience with Lisp can be boiled down to "finished this thing and remained indifferent". I'm no expert and know it, but I'm also extremely sceptical about stuff proclaimed by Lispers.
I don't think he does? Perhaps one could say that Lisp is less abstract than lambda calculus in the sense that it has primitive operations on integers and whatnot. I also don't think he's arguing that it's more abstract, just that the notation makes it easier for humans to reason about programs, much in the same way that arithmetic is easier with Arabic numbers than Roman numerals. Or how currying and partial function application is common in Haskell because the syntax makes those operations painless.