Why ‘Friending’ Will Be Obsolete

Every few days, (or hours) you probably get a friend request of some sort, the good news is, someday, this will not be relevant.

I just got finished watching this video of Renato of “E”, a device and software platform that allows you to phsyically gesture in the real world with people you meet that you are friends. Remember palm pilot back in 2001 that let you ‘beam’ contact info to each other? Similiar to that, but now with more ‘social’ context.

Thinking forward a few years, “friending people” whether in Facebook, Plaxo, or will no longer be an activity that we’ll have to do. Intelligent websites (and their data) will be able to determine who our friends are from our behaviors, context, and preferences, without us verbally (or physically) having to indicate so.

Those ‘bacn’ messages make it painfully obvious that the ‘system’ (the web, data, and beyond) isn’t quite intelligent, as it can’t determine who our own social contacts are yet.

Like a baby, we’re teaching the ‘system’ our language, how to walk, how to coexist in our real flesh and blood world, the ‘system’ is just starting to show intelligence. One primary example of this is the use of hashtags in Twitter. We use the # sign to tag content so it’s easily to organize and find. That one # character isn’t native to our tongue (unless when you recite your grocery list and say “hashtag”) it’s another example of us speaking machine language in order to teach the system.

For example, I started a social experiment on Sunday, where I encouraged folks to tweet related music artists using the tag “#relatedmusic” you can see the database form when you search for that term –If we had enough people do this in my –and your– network we’d be able to build a reference engine that other music reccomendations services could pull from.

Saturday, I posted my thoughts about what’s next after the social web, and some interesting comments are coming in, take the time to read about it. For now, I’m tagging these posts the ‘Intelligent Web” as I don’t really think we’re there yet, we’re having to input too much to teach the machine right now.

If you can’t see where this is headed, I’ll tell you: all of what we’re doing from our clicks, queries, wall posts and tweets is teaching the ‘system’. In the long run we’re creating a massive global computer, an artificial intelligence, and someday, a thinking being.

Therefore, when the ‘system’ is more mature, we won’t have to explicitly state who we’re friends with anyone –it’ll have learned and already know.

Phrases to know:
I’ll refine these over time, but for now, here’s some early stakes in the ground.

The System: The system is the combination of all websites combined, it’s a massive data base of content, clicks, search terms, time on site, shared posts, wall posts, links, and tweets.

Teaching the System: Humans are constantly speaking in machine language, from use of hashtags in twitter, or boolean searches in Google, or even from the act of friending folks in your social network. All of these behaviors are humans teaching the system how to understand us, so it can better serve us.

The Intelligent Web: Software that is able to collect and make sense of all the data in the system and is able to deliver meaningful content back to people in context –often without us saying or gesturing that we need it.

Update: This has spun off to more discussions in Friendfeed, I think a lot of folks missed the big idea I was trying to convey here –or perhaps more likely, I didn’t communicate it well.

47 Replies to “Why ‘Friending’ Will Be Obsolete”

  1. And just to be clear, I don’t think “E” or “Facebook” or any other social network is going to be obsolete, but it’s a necessary step to get to the intelligent web. We’ve got to teach our child first.

  2. If you haven’t already, you should read ‘Linked’ by Albert-Laszlo Barabasi. Published in 2002, it talks about the Internet as becoming a computer like skin and, at some point, it will become self-aware.

  3. Every few days, (or hours) you probably get a friend request of some sort,

    Hey Jeremiah, I would dare say that this is a problem quite unique to a set of bloggers and social hubs.

  4. This is some good stuff Jeremiah. I wonder if we’ll see a new employment category related to “Teaching the System.” In particular, what areas of human experience will we teach it and why?

    Obviously the nature of human relationships (“friending” and “business contacts” being the subsets of that we’re working on now). But others as well.

    Currently I’m most interested in locality or geography, the whole place/space thing. Using mobile technology for machines (GPS) we’ll probably get some of the data. Using mobile technology for humans (iPhone/Android/ec) we might gather some of the meaning.

    Maybe the Intelligent Web will be able to cross the street on it’s own someday. And then maybe learn about the street and where it goes, why we use streets, etc.

    Fascinating post. Can’t wait to read more, my gears are churning through it already.

  5. What we understood yesterday as “friend” is changing as we connect, participate and watch .. and what we understand today as “friend” will continue to morph into something quite different still.

    due to this Web thingy, people have been talking and writing for years about “trust”. Think for a moment about how long, often, it has taken you to really trust someone in “real life”, and then think, for a few seconds longer, about what it will take for you to “trust” someone you know mainly through online interactions.

    Yes, we have the public record of what they have said / say on the Web (and that is an important input to the process of trusting). But that process of trusting will take on new dimensions, and we have much yet to learn about it.

  6. G. Dewald

    Glad you’re churning, keep on giving me input, as I’m thinking outloud here in public, but enjoy the input from others.

    Jon Husband
    Agreed, in real life, we’ve many layers of friends, we can’t often label people accurately as they change over time. My former colleagues were very important to me when we worked together –but may not be as much now, social networks don’t currently reflect that change.

    At some point, the act of ‘friending’ needs to go away.

  7. I guess what you are saying is that the awkwardness of formally accepting someone as a relation will disappear. I don’t know if this mean physical-like contact will be the model, and I doubt it. Facebook and Google have been quite efficient in using your browsing history ” and I guess something along the lines of a list of suggested people you might want to call, write to, meet with, etc. based on your actions, and the possibility to explore those previous interactions would make sense. The fading formalism with help the current main problem that is hard-to-accept banning or un-friending ” but I’m afraid more pervasive recommendation won’t help with balkanisation.

  8. Am I alone in being just a little scared of this prospect? In hoping that J is wrong, that the web is not going to become intelligent? I like the web – I work through it every day of the week – but the idea of it becoming aware and telling me who my friends are is just a little spooky to me.

  9. Great article! It looks like we are moving towards arranged friendships. Will there be a competitive market for AI software which gives us a better digital social presence?

  10. Excellent post Jeremiah!

    Another aspect of our ‘teaching the system’ and ‘the system learning’ can be found in APML or Attention Profiling Mark-up Language. APML is an open standard and open format that encapsulates a summary of your interests (across multiple profiles) in a simple & portable way, to store/capture/gather (more) user information.

    SEE: http://www.apml.org

    “The idea is to compress all forms of Attention Data into a portable file format containing a description of ranked user interests.”

    Either Adoption of APML or friend source aggregation service integration into our daily social lives online, via our derived set of interests, could just possibly do the trick!

    This is about our Attention Data.

  11. There is a new research field dubbed “reality mining” which constructs social networks — relationships/friends — based on actual behavior, primarily proximity and duration of contact. Researchers at MIT have created automated tagging/logging sensors for just that purpose, and there are beginning to be mobile phone apps that accomplish almost the same thing. Of course, “proximity” doesn’t always equate to “affinity” – we don’t always like our coworkers for example.

    I haven’t heard anybody propose bridging “reality mining” and online social networking platforms… yet… but it’s a very intriguing idea.

    I’ve posted more on reality mining at:

  12. I definitely agree w the concept of the intelligent web as a system where you won’t have to overtly label all acquaintances as friends, but I wonder and hope that proactive choices and labeling still exist and grow. How will the intelligent web help me parse through the layers of friends to follow and track closely those that really matter and lightly follow those that I™m mildly curious about?

  13. Jeremiah, We have to give the system the intelligence it needs to serve us in this regard and not expect it to spring up organically. A lot of smart people are laboring to try to make it happen, like those working on the Friend of a Friend (FOAF) semantic web project.

  14. Interesting idea. Although I think the likelihood of a sentient Internet is quite some decades off, we might find the Internet becoming a really good assistant to each of us individually.

    Having an Internet sidekick to handle the mundane chores of a social relationship seems valuable (sending Jeremiah’s GPS unit the location of the restaurant where we agreed to meet after making lunch arrangements, recognizing a good friend is coming to town and fostering a quick IM/phone/email conversation, etc.). But the Internet as a being seems a ways off and not necessarily a positive action (maybe I’ve seen too many Outer Limits episodes). Remember, HAL was essentially “awake” and had emotions, but for all his knowledge he dealt with his emotions on the level of an 800 lb first-grader and the outcome was predictable.

    We have a long ways to go before a sentient Internet could be considered a positive contributor.

    I agree that FaceBook, etc., are not going away. If anything, they will grow even more powerful because they provide the entry drug to the social network, the trial waters. But, eventually, the social sites outlive their usefulness to the individual andthe more open web (intended) of blogs, Flickr, etc., will take hold. Friendfeed already provides what is essentially an un-walled version of FaceBook, et.al.version

  15. Re: time-based alterations of “friendness” this one is a biggie. Perhaps working hand in hand with the use-by date of information in general (I think had a link out to Saad somewhere that laid this out great).

    The social graph is just like any other bit of information, perhaps with shorter use-by dates than other kinds.

    And it isn’t just one-way in becoming less relevant when you change jobs/towns/social-scenes etc. You might suddenly reconnect with an old friend and then that contact becomes more important for a time etc.

    I think seeing beyond just the social aspects of teaching the system will get even more interesting.

  16. “Teaching the system” is a great way to think about all the things going on in the web world (and other places).

    How far will all this go? Could you imagine a world where you interact with people face to face and ‘the system’ is some how there, involved and acting on the information that is being transmitted in the real life social environment! (sounds like the Matrix)

    Thanks for challenging & expanding my thinking.


  17. Great post!

    I noticed Google is already doing a little bit on this respect. When I was playing with my Youtube account, I realized now it pulls my contact list from Gmail, which are people I have exchanged emails with previously.

    The web can’t know who we meet in real life yet. But knowing who we have exchanged conversation with is a good start! I also like this “loose definition” of “relationship/friendship” a lot, which is based on interaction instead of confirmation. If Google can take it one step further and sort my contacts based on the frequency I communication with them, including using other tools such as g-talk, things would be even more interesting!

  18. I feel like what you’re talking about can’t happen without significant efforts on the data portability front. Most teaching the system concepts I see in practice now are mostly superficial. Some recent research in this subject however is interesting. It’ll get really interesting when we can break out of the walled garden approach to social networks and figure out how to govern relationships in the cloud as a whole.

  19. I agree with Nigel and I don’t think this is going to happen.

    What you describe “Intelligent websites (and their data) will be able to determine who our friends are from our behaviors, context, and preferences, without us verbally (or physically) having to indicate so.” even when technology will make it possible it will fail because it goes against human nature.

    Related to our personal lifes, humans will not let anyone to make decisions for them and will dislike this possibility even if it could be useful. We will prefer to make mistakes than to trust a machine or ‘system’. That’s simply human.

    Personally, I like web recommendations based on my browsing history or past behaviour on the web but not to the extend that the web will recommend me friends. To illustrate my point, there are cases when users don’t mind the system using their info (amazon recommendations) but there are other cases when users are pretty vocal against it (remember gMail at launch).

    This post reminds me the article a couple of numbers ago in Business Week about IBM trying to model with numbers their consultants in such a way that ‘the system’ could optimize their output. Really interesting article that drove a huge reaction from readers saying that they will not like to work for IBM….curious 🙂

  20. Luis

    There are systems like this that already look at your email behavior (like Xobni) that can already track at who your real contacts are based on behavior and history.

    This trend is really inevitable.

  21. Jeremiah, at first blush I couldn™t disagree more. But then maybe I™m reading this wrong. Are you actually saying that a machine is going to be able to learn enough to know who my friends are, whether they are friends, business associates, or potential employees/partners and whether they are in my inner circle, what information I want to share with them and how I want to connect with them? What is going to endow that machine with something that most people don™t have?

  22. RE: “determine who our friends are from our behaviors, context, and preferences…”

    Unless the system plugs into a pair of “precogs” like in the Minority Report, it will never know who I am going to make friends with TOMORROW.

    So, IMHO, as long as humans are making new friends and random coincidences are happening (i.e. bumping into and meeting new people), then “friending” is not going anywhere.

    RE: Where’s it all going, a “thinking being”: Yes, you can just graph the human brain vs. the accelerating trend of machine intelligence and see the lines will cross in the not too distant future. If you assume continued and accelerating rate of technological advancement, then this is indeed, a no-brainer, if I may use the term.

    Just read “The Singluarity Is Near” by Ray Kurzweil.


  23. Bill

    This post is so much more than ‘friending’ it’s about the internet becoming a sentient being that like a butler, can anticipate what we need and want.

    To answer your specific question, ‘yes’ it will be able to identify who your relationships are with (in proximity from your wife to your business contacts) based on your previous behavior –and theirs.

    Now, back out of the weeds: Think bigger, everything we do online (even clicking) is teaching the system to crawl, walk, and someday talk and run next to us.

  24. BTW: When I see many folks agreeing, some going “hmmm” and some saying flat out “now way” I know I’m getting traction. I saw this same reaction in 2005 with social media.

  25. Here’s an example of an attempt to teach the system. It’s in relation to a recent official Google blog post about the Google preferring folks not rewrite their dynamic URLs (i.e. they now prefer /?p=1234 over /why-friending-will-be-obsolete/):


    The Google wants to teach its system more about how web sites work. Web site owners want the presentation of the data to be useful to real people.

    Some review of the old semantic vs visual arguments from back in the day will probably yield some good stuff to mine here.

  26. Fascinating stuff. I’m not sure I know enough to agree or disagree with where this is headed, but do any of us? That to me is the interesting part. It seems we can understand and predict trends in technology and markets and even behavior, but we’ll always be surprised.

  27. Great post! I wonder how this applies in a given culture. Will friending become more intuitive in societies where a social network already tips into the overall population? If you think about CyWorld, for example, 45 million Koreans (or 1/3 of the population of South Korea) are members. Does this mean they will converge faster and think more similarly about politics, religion or social issues?

  28. Hi,

    This is really interesting stuff.

    In my mind this relates to Project VRM (http://cyber.law.harvard.edu/projectvrm/Main_Page) and the notion that we can use The System to our advantage in many ways – taking our relationships with us wherever we go and also consuming information from companies, people etc based on our interests/ friendships that The System know’s about based on what we’ve been doing online.

  29. To build on this conversation – I think it is a mistake to get locked into arguments over whether individual examples will be acceptable based on current social norms – the overall direction of what Jeremiah calls “the intelligent web” and, unless I am mistaken, has long been called the “semantic web” is clear: our input into the web is instructional – and indicative of our behavior, preferences etc. As computing power grows this massive amount of data is being used to provide a richer and more predictive set of services. As these service offerings grow, we will take advantage of them. As we take advantage of them – our expectations will change regarding identity, privacy and our sense of independence from machine-mediated living etc.

    Also, it is a big mistake to argue from current cultural norms (esp. citing “human nature” since I am not sure anyone believes that cultural is not a flexible construct that evolves based on environmental factors). If anyone had told us in the 60s that our society would accept the lack of privacy inherent in the Internet today (At any moment the system knows what you have purchased, where you are (if you have a GPS phone, what you have watched on TV, not to mention every click you have made on the web, your friends on social networks, your job history etc.) I am sure we would have thought this was a description of Stalinist Russia…
    Social norms are shifting rapidly and we are becoming more accustomed to an Internet-intermediated life…

    Here is a great video that, if you haven’t seen it, tells the “we are teaching the machine” story very well. http://www.youtube.com/watch?v=6gmP4nk0EOE
    Stick with the video – the payoff is worth it.

  30. Interesting thoughts, but I think friending will always be around. How else can I get in touch with a friend I haven’t see in 10 yrs? Just like my hotmail account won’t be able to search for them and send them an email to say hi, no web site will be able to know my friends unless I already interact with them in the system.

  31. This is some good stuff Jeremiah. I wonder if we'll see a new employment category related to “Teaching the System.” In particular, what areas of human experience will we teach it and why?

    Obviously the nature of human relationships (“friending” and “business contacts” being the subsets of that we're working on now). But others as well.

    Currently I'm most interested in locality or geography, the whole place/space thing. Using mobile technology for machines (GPS) we'll probably get some of the data. Using mobile technology for humans (iPhone/Android/ec) we might gather some of the meaning.

    Maybe the Intelligent Web will be able to cross the street on it's own someday. And then maybe learn about the street and where it goes, why we use streets, etc.

    Fascinating post. Can't wait to read more, my gears are churning through it already.

Comments are closed.