Definitely not an attack or a skeptic, indeed, it may play into your assessment of the worth of the book...
But 12 months, 6 languages - how well do you feel you've mastered any/all of those?
I mean, it teaches scheme for Pete's sake. Playing Devil's advocate here, who honestly uses that in pratcital everyday systems? Nobody reads the book and immediately goes "Wow! Now that I learned Scheme, which is totally the best language, I'll use my Scheme powers to create the coolest software in the world!"
The reason why this book is so cherished is because it treats scheme as a building block for implementing the most interesting ideas and patterns from systems you use every day.
Rusty on recursion? That's chapter one. You'll be thinking with recursion and will understand most of its implications by the first few dozen pages.
Think you know about OO? Build your own object-orientation system in Chapter 2. Yes, on top of scheme.
Write your own programming language (well, implement eval anyway) in a few short pages in Chapter 4. (The result is about thirty lines).
How do Von Neuman register machines work? If you don't already know, you'll be writing a simple compiler in Chapter 5.
This up-close-and-personal whirlwind tour of these important CS concepts was very enlightening to me when I read it. This book isn't about how to learn languages; rather, it's about working with ideas: why abstractions are important, how to reason with them, and how to implement them using whatever bare tools you've got.
(Schemers: before commenting, realize that i wasn't poking fun at our favorite language; rather, i'm merely asserting that scheme isn't the point of the book)
The 'superpower' that the book gave me was a set of tools to think about computation in a language-independent way. Once I had that, then every new language was "oh, here's a nice syntax for doing something I already know how to do in Scheme".
The new languages I learned were from a variety of paradigms - functional (Haskell, OCaml, R), object-oriented (Java, Python) and array-based (Q, Matlab). I haven't learned a logic-based language yet, but I've written an evaluator for one (section 4.4 of SICP) so I don't expect picking up Prolog to be difficult, if I ever choose to do it.
That's a very nice line. I like that very much.
It's pretty ridiculous, but I think it puts the person back into the comfortable mindset of control. Rather than learning a big, scary thing that is beyond you, you are merely manipulating the world with your powers to accomplish tasks. For whatever reason, that paradigm is more intuitive for certain people.
In 2010 I started a job that was far, far ahead of my skillset. I'd written fewer than 10,000 lines of code in my life, and I was completely ill-prepared for the work I was supposed to be doing. I picked up SICP about a month or two into that job.
Fast forward a year, and I felt that I knew more about programming than half the people I worked with, and I'd moved to a smaller, more awesome company doing work that I think is much more fun.
A year since then and I've picked up half a dozen new languages, given talks at user groups on some of them. Next year I'm aiming for conferences.
Pick up this book. It gives you superpowers.