Solving the filter problem: Bottlenose introduces "assistants"

By | December 22, 2011
I got beta access to Bottlenose last week. There seems to be some problems (the most annoying one is in refresh rates), and some other smaller issues – keep in mind that this is a beta testing phase – but in general leaves a great first impression: this is a user-friendly solution to the filter problem that those of us following a lot of people have on social networks.

How it works
1. You log in and grant Bottlenose access to your account (typical OAuth token exchange).
2. Bottlenose crawls your profile and posts and takes a guess at your interests. There's a semantic natural language ontology hiding behind this interest categorization, so they can pick up similar topics and aggregate. They did pick up a lot of mine, but missed a few – not a big deal, since you can add them to your settings to 'fix' this.
3. Bottlenose now sets to work providing you with information streams. They provide three types of streams: Main, Assistants, and Types. Main is what you'd get on any native social network. Types are filtered according to content: Videos, Pictures, News, etc. Assistants are the interesting part. By default, you get a "Suggested for you" and a "Popular" stream. Popular are the incoming posts which are most popular. "Suggested for you" is a filter across the incoming posts based on your interests (defined in step 2). On top of that, you can create additional streams with your own criteria.

What's a pity is that this is currently only available for Facebook and Twitter, and only available in English. I don't use Facebook actively, and Twitter has been lurking in the background since all the great discussion takes place on Google+, but it does do a good job of helping me find the 'golden bits' in all the twitter noise without having to monitor the stream. (I asked: Google+ is in planning, they're just waiting for the API to be released to provide them access to non-public posts. Support for other languages is also planned for the future).

Why this solves the filter problem, and Google's sliders don't
I haven't been able to confirm whether the sliders cut non-popular posts or if a semantic assistant is running in the background. But there are a few key differences that make Bottlenose's solution, from a user perspective, a superior one to Google's sliders:
1. It does all the keyword categorization adaptively behind-the-scenes. No need to force hashtags on mainstream users (though it does intelligently take hashtags into consideration during post categorization), and no need to brainstorm all possible keywords you may want to read.
2. It has default settings that 'just work' and only require minor tweaks.
3. It offers users enough insight on how it calculates interests (via sonar and keywords) that you feel in control of what is getting filtered.
4. It makes it really obvious which stream is full and which one is filtered, and makes it really easy to toggle between the two.
5. It allows you to define your own streams in addition to the standard ones.

Want an invite?
In case you are interested in trying, I have 8 invites left – send me a message with your email, first come first serve.

/cc +Stefan Münz – you may be interested in taking a look.

Embedded Link

Bottlenose – A Smarter Way To Surf The Stream
Bottlenose is a new social assistance tool that helps you engage with your streams more intelligently.

One thought on “Solving the filter problem: Bottlenose introduces "assistants"

Leave a Reply

Your email address will not be published.