350 million users. That’s a lot of people, but is the quoted total number of Facebook users that was announced last Wednesday.

Although objected to by many, Facebook has undergone a set of significant changes across the site over the past year, perhaps the most prominent being the real-time news feed built from activity by users on your friends list, which is updated as things happen.

This is really cool, but the problem lies in the way that the Facebook community works, or rather how human nature works, in that being connected with more people or being seen as “popular” is something desired.

The main consequence of having an ecosystem that promotes connecting with more and more users in this context is that it creates a huge scalability problem. Building feeds from data from lots of users is computationally expensive, and requiring more resources to do so becomes economically expensive. And this problem is only going to increase as time goes on, because no-one has the time nor inclination to be constantly tweaking their friends list to remove users they no longer communicate with, for the sake of a number on their profile.

But the real irony is that Facebook are shooting themselves in the foot in a sense, by building a community where having more users is promoted. This comes in the form of them actively tempting you to connect with other users, with “friend suggestions” or suggestions for joining groups that all your friends are joining.

Users being users, they’re going to join groups that they’re not even mildly passionate about, or connect with people even if they’re loose acquaintances because the cost is only one click, which only contributes to the problem.

Jason Calacanis noted the scalability problem on episode #56 of the Stack Overflow podcast:

Like Facebook with the social graph. I have 5,000 people on Facebook. When I log in, it takes 90 seconds for me to log in, if it works.
…
Because – it’s true! And I talked to Zuckerberg about it, I’ve talked to people about it, and they’re like “Yeah when you log on you realise like 18 servers have to go to work to like..” – when I log into Facebook I I think it would put about 20lbs of Carbon into the air
…

That’s building feeds from 5,000 users, which is fairly uncommon. But these kinds of problems have been noted by users with only close to 1,000 friends. And of course, why wouldn’t it? You’re trying to build a feed based on data from 1,000 users, for one user, in a few seconds.

Factor in 350 million users, and you have an extremely complex computational problem, that won’t go away.