If you've been following this argument from the beginning (I'm very sorry to say that I have), you'll notice that Ted's point is not that you WILL do CPU bound tasks in Node, but that Ryan Dahl's misleading statements on Node.js's front page is absurd. Of course no one with any sense will put CPU bound jobs on the same process/thread/event queue as the one processing the request. But you do have to realize, it is exactly that no one will ever do the above that renders Node's existence rather pointless. Threads/processes or events, you'll still pass off the heavily computational stuff to a different processing queue. Events or threads, when that single process starts topping out, you'll still have to launch another process. What makes Node.js different from any other technology that have come before? Nothing except that it comes with a server and it's written in JS. If you like the Node REPL and JS, that's fine, just don't propagate the urban myth that's on Node's front page. I think this is Ted's point IMHO.
