Not sure though. But top of it you would probably use backbone.js in both server and client. Another advantage is we have UI developers who are now writing server code to be much more. To some level it is giving a path way for one language everywhere paradigm. However, server side coding required much more capabilities like performant code, scaling etc., but for normal websites it is easy to go with node as it kills some amount learning curve.
I mean to say, suppose they have to finish a document. Now due to the time factor they are "making-up" something to achieve the deadline. They know they are "intentionally" missing something. Due to the fact they know I will catch the missing thing or "improper" document they say to other peers. "Don't miss it, Steve will catch you". Now this is okay that the document ends up to be expected one. But the mindset of someone will catch me if I miss it -- is this is normal? What if I fail to attend such daily meetings in future. At some point, once they realize, I am no more there to review them they will no longer producing "perfect" ones. Or I am doing lot of baby-sitting that made my team so dump they just follow my lines. Kind of have some thoughts around how to manage such team would be great to hear.
"I was miserable. I couldn't stand San Francisco. I couldn't stand office life. I couldn't stand Wired. I took a long Christmas vacation. I got sick. I thought of suicide. I ran from the police. And when I got back on Monday morning, I was asked to resign."
A person who moved from developer to CTO to CEO I am warning you don't MISS college. I have found myself in strange situations without knowing Mathematics that I have to go back to basics to do my job. If you take it easily by looking at low hanging fruit you will suffer later. The memories and learning you will have from college will be always there with you like swimming -- no matter if you dont swim for 10yrs still you won't loose that talent.
Until or unless you don't face scaling problems you won't e able to learn it. Luckily my project's had scaling problems in its core. Every design discussion we had was always having some "scaling caution" because we know we will be making a technical liability of we had to just solve the business problem. Essentialy, you need such projects or projects thats would be anticipating scaling issues in long run would be right projects to work for. Every learning you carry from projects and companies on scaling never fits to another scaling problem. In other words, you just can't use some tools to solve the problem. For example, just by knowing to use HBase or Cassandara or scalable technologies you would never be able to call you as a "scaling expert". It is much more beyond tools knowledge like trying to understand cost around a scalable idea, maintainability etc.,. I just scratched the surface, I can talk another twenty pages.
Another way to answer your question would be - think how FB might have implemented "people you may know" widget. Do they need to traverse thru 900m users to check whether you are related them in some way? Is that possible to do so everytime page gets rendered? Certainly not. There is some other way to do this. So taking up such questions and trying to find answers helps to get motivated on getting a feel of how scaling could be done? May be trying to make such POC would lead you understand technologies in problem specific approach.
Just my thoughts. Can't go back and review my typing as iPhone and HN doesn't go well. Will edit this for more clarity when am near to desk.