Arbital: information hub

Larry Page, founder of Google, said: “Our goal is to get you the right in­for­ma­tion at just the right time … with­out you hav­ing to ask first.” This is Ar­bital’s goal as well. While there is a ba­sic sub­scrip­tion sys­tem in place already, most of the fea­tures de­scribed be­low are not yet im­ple­mented.

Right information

Whether or not in­for­ma­tion is right for any given user de­pends on how rele­vant it is to their in­ter­ests. The sim­plest way for peo­ple to ex­press their in­ter­est is to sub­scribe to top­ics or users they find in­ter­est­ing. To take a step fur­ther, the user should be able to spec­ify, if they want, how strong their in­ter­est is, or that, in fact, they are not in­ter­est­ing in any­thing in that area. For ex­am­ple, on Face­book you can say that you want to see some users’ posts first. But you can’t say that you don’t want any posts about poli­tics. You are ei­ther sub­scribed to ev­ery­thing a user posts or none of it, and there is no sup­port for top­ics.

On Ar­bital, you will be able to sub­scribe to any par­tic­u­lar topic, rang­ing from as broad as Math or as nar­row as Roboto Font, and to any user. You can vary the strength of the sub­scrip­tion, in­clud­ing say­ing that you are ex­tremely dis­in­ter­ested in a given topic. Ar­bital will ag­gre­gate your in­ter­ests and use them to pre­dict whether any given piece of in­for­ma­tion is rele­vant to you. For ex­am­ple, let’s say there is a re­cent, pop­u­lar blog post dis­cussing a new cryp­tocur­rency, Stel­lar. A few peo­ple you fol­low liked it and/​or com­mented on it. Even though you are not sub­scribed to that blog, as long as you are sub­scribed to a few rele­vant top­ics e.g. Bit­coin, Ar­bital will still show the post in your feed. It’ll be able to use all the sig­nals: post’s pop­u­lar­ity, friend’s likes and com­ments, and your re­lated sub­scrip­tions, to pre­dict that you’ll likely en­joy read­ing this post.

You might have seen that Ar­bital already has a func­tional, in­stant, fuzzy search. Mov­ing for­ward, we will make search­ing Ar­bital eas­ier and more pow­er­ful. The search will be able to lev­er­age all the meta­data that Ar­bital has about the page’s con­tent and how the page is con­nected to ev­ery­thing else. For ex­am­ple, you’d be able to search for “all pa­pers at the in­ter­sec­tion of ‘Vi­tamin D’ and ‘Cancer’ top­ics, which also have p<0.05.”

Struc­tured data

To be able to ex­e­cute a search query like the one above, the sys­tem needs to have the struc­tured data about pa­pers and other ob­jects. A com­mon com­plaint about Wikipe­dia is that while it has a lot of data, most of it is not struc­tured, which makes it very hard to use in com­pu­ta­tion. On Ar­bital, we plan to ad­dress this prob­lem early on.


Ar­bital: in­traweb solution

Many com­pa­nies have an in­ter­nal wiki sys­tem be­cause it works well for link­ing, search­ing, and cat­e­go­riz­ing in­for­ma­tion. (Much bet­ter than, say, Google Docs.) Ar­bital is bet­ter than any wiki, and it has pri­vate sub­do­mains that give any group their own pri­vate in­stance of Ar­bital. (You can still lev­er­age Ar­bital’s pub­lic con­tent.) If you are in­ter­ested in us­ing Ar­bital for your com­pany, send us an email: alexei@ar­ <div>