skip to Main Content

On the robustness of Twitter and false SAT Analogies

This blog entry is a response to  Cody Brown’s post here. I wanted to leave a comment on the post, but it was going to run a bit long, so I thought I’d put up a response of my own.

Cody’s piece is interesting and well-reasoned. The basic argument he makes is that Twitter, much like Myspace, became popular before it new what it was, and is now suffering an identity crisis that will force it to fracture / fade into obscurity and be replaced by more focused applications. The comparison of Twitter to Myspace, however, falls short: in one sentence, I would claim that Twitter will not go the way of Myspace, because Myspace is more of an environment, whereas Twitter is a platform.

It’s true, both Twitter and Myspace suffer from lack of clear vision and perhaps an overabundance of uses. It’s a citizen journalism service, a way to catch up / chat with your friends, a procrastination device, a way for fans to follow celebrities, etc. You can use Twitter to write novels and play chess. This was the problem with Myspace: it offered users unlimited means of self-expression without a single overarching paradigm. When users were bored with expressing themselves (as all users inevitably are), there was nothing solid to keep them on the site, and so they drifted. But Twitter does have a single overarching paradigm: the tweet-stream.

Twitter’s greatest use is as a low-level service to provide individuals with a socially filtered, digestible, flexible stream of information. E-mail doesn’t provide digestible streams: the limit to the length of an e-mail is MUCH bigger than 140 characters. RSS is not socially filtered. Social bookmarking is both digestible and socially filtered, but less flexible – it doesn’t allow users to engage in dialogue and commentary as part of the bookmarking stream. Facebook has profiles and pictures and events which make the stream (news feed) much less digestible. 

Does this mean that Twitter is perfect? Not at all. I think Twitter is going through an identity crisis (though that crisis is going to happen more on the surface and the periphery, and not detract from the utility of the core service). In my opinion, the way out of that crisis lies in going back to the core, to the socially filtered, digestible flexible information stream. There have been a number of apps built on top of Twitter, but, to my knowledge, these apps have yet to fully harness the stream. First step is search, which is happening and is a good thing. Second step should be better filtering – I want real-time manipulation of my tweet-stream to filter out posts by person X. Third step should be extracting social interactions from pure information streams – I want to grab all the @posts I’ve had with X, and their responses, and the responses of all of my other friends who have seen these @posts and commented on them. I want an interactive social graph plugin that I can drag / click on to expose tweets between / by different subsets of my followers, or followees, or both. These features would not only help me manage incoming tweets, but also help me organize my stream, and give me more control over it. 

So these are some thoughts about Twitter, why it’s not going the way of MySpace, and how it could be better. I hope that the features I listed are either a) already available and I don’t know about them, or b) will be developed soon. Twitter is a great platform and has been incredibly useful for my own social media management. I hope to see it grow and improve in the coming months!

This Post Has 4 Comments

  1. Interesting angle but I think what you say is wrong about MySpace also applies to Twitter in a big way:

    “It offered users unlimited means of self-expression without a single overarching paradigm.”

    140 characters is the only restriction in communication in twitter. The site does not direct any overarching paradigm – this is decided individually (and often, even this will shift).

    The list of features you suggest are complex. Even if they are implemented well, once you start bootstrapping that much your experience on the site will be sharply different than the common experience. This will create an even more disparate set of users.

    I’m not saying Twitter is bad so much as Twitter, started as one thing, we’ve stumbled upon an even greater use for the product and it’s now about whether Twitter as a company can develop that use properly. I don’t think they can. That overarching paradigm, just doesn’t exist for Twitter. It’s potential as a website that enables you to hear and communicate to a public is much to important for 140 characters.

  2. Mr turney already developed a method to calculat the sat analogy results that works perfectly fine, much better than in humans, and on top of that he develop his famous aprisal measurement method which is a semisupervised one (oonly 7 good and 7 bad examples) and the hole corpus from the web used to learn from, so I think you are not right about what you are saying, what thing more robust do you want than something that has a more than a 90% accuracy in measusing the direction of the adjective using almost the entire corpora of the www. And now you are telling me that twitter is a more robust thing, which are simple sentences that barely have a background, Not to mention the very little percentage of information it has compared with a hole searcher result set.

    Did I missunderstand your argument? what is it that you say here that I do not get?

Comments are closed.

Back To Top