Written by Alex Iskold and edited by Richard MacManus

Last year at DEMOfall a company called Vapor Stream generated a lot of controversial press because it introduced an email system capable of completely erasing any trace of your communication. This year, one of the contenders that may generate a lot of controversy is Me.dium - a Colorado based startup that meshes up attention, web surfing and chatting to deliver extreme, real-time social networking, right in the browser.

The idea behind me.dium is to let you see and communicate in real-time with your friends and other people online who are "close" to you, based on what sites you are visiting on the Internet. Implemented as a browser add-on, me.dium continuously captures your locations on the Web (i.e. which sites or services you are using), sends this data to the centralized server and computes a map of other sites and users that it determines is relevant, based on your context. Here is a screenshot of it in action:

The map above is displayed using a network-like visualization, where distance between your site and other sites represents the inferred level of relevance.

The users are displayed at the sites - and the more popular the site, the more users you will see. Your friends are color coded differently from other people. Since these displayed users are actual people, me.dium takes the next logical step and embeds a chat window which allows you to chat with these users in real time.

The thinking behind me.dium

This product was inspired by real-life examples. According to the company web site, often our decisions are influenced by the decisions of others. Whether shopping for a car or choosing a movie or a restaurant, we often rely on the judgement and recommendations of others. Geographical proximity plays an important role, as we often strike conversations with complete strangers in order to find the "best deal". So the founders of me.dium asked the natural question: Why not do the same online?

Can this really work?

The idea is intruiging and so is the product, but can it really work? The key question is: Is it useful to be able to interact with other people who doing/looking at similar things? We think that yes, it is useful - but not all the time and not intensely. It is important to remember the lessons of the Uncertainty Principle in Quantum Mechanics - if you interact with the system, you change it. While sometimes it is interesting to ask people around you what they are thinking, you can't do that all the time; otherwise, instead of doing what you are doing, you are just talking to other people about it.

Also the current implementation, particularly UI and user experience, need to be improved when letting the user request to see the sites and people nearby. Particularly, a control that would say: just show me my friends (vs. everyone else) would be really useful. Also, to facilitate the right interactions, it is important to reveal more details about the other sites without much drilling. Perhaps a list of sites with descriptions would be a good alternative user interface, as the moving network of nodes is just too distracting.

Privacy concerns

The success of me.dium does not just hedge on the level of the user interest and the interface. Handling of privacy is a big part of the acceptance equation too. Right now, the tool immediately starts collecting all the private browsing/attention data and storing it on a centralized server. The company claims it is safe, because they spent a lot of time thinking about security. But a lot of privacy advocates will be upset when they hear about their browsing data being collected and stored somewhere. Overall, it seems to me that the tool takes a fairly aggressive approach to data collection. The key counter point from the company is that they need all this data in order to deliver the service and utility.

Conclusion

In its current re-incaranation, me.dium appears quite intrusive - both from privacy and activity point of view. It does not mean, however, that this idea is bad or that it cannot succeed. On the contrary, I believe that it has a lot of potential and with a bit of tweaking can take on and spread virally. The key things that need to fall into place are: user experience and (perhaps more importantly) full, fine-grained control of the user information by the actual user. As it stands, me.dium is bound to generate a lot of controversy and discussion. So please jump right in and tell us what you think.