Hello! I thought I'd throw my hat in the court.
[Link removed by moderator: original site defunct, now taken over by malware]
Note that the frontend of this is barely more than a bootstrap site with some layout thrown together. The backend is a WIP, but currently has some notable things in place:
Python+pSQL backend, with nginx in front of it
SCSS+CoffeeScript processing pipeline (also supports plain CSS and JS)
SqlAlchemy is wired up with a few data types defined and wired up
[Link removed by moderator: original site defunct, now taken over by malware]
I've done scalable sites before. I have one site that runs on a single server and handles a million pageviews a month, streaming HTML5 video to most of them. It peaks at a thousand requests per second during heavy loads and doesn't break a sweat. I'm taking a similar approach here, and this site should be able to scale fine on day one (assuming it's on better hardware than the dev server it's running on now). Speaking of that site that handles a million pageviews a month, I have a good website with infrastructure for hosting images and video that can be easily integrated into this one - https://mediacru.sh (also open source). You can see on the example mod page that MediaCrush is running the embedded video. I'll get this integrated more or less seamlessly and then the load of images and video is handed over to a fast and reliable host.
I've read this thread and I know more or less what people want, so I'm going to do my best to match that. I chose not to go with MongoDB for two reasons - data integrity, and fancy queries. I know people want data on how their mods are being used, and a relational database is a good tool to crunch that data for them. I chose Python because it runs on any Linux server out there, as well as being a sane language overall (I did not choose PHP because it is not a sane language. I did not choose node because it's not as sane as it should be). I also went with Python because I can steal a lot of the MediaCrush code (like the asset pipeline) to make things go faster. I have also decided NOT to do some things that I've heard in this thread. I will not be hosting a build server (out of scope), and version 1 will probably not have email blasts when mods update (I'd rather not have GMail flag me as a spammer).
Let me know if you're interested in helping out.