Explicit vs Implicit Social Nets

For years many industry observers including me have stated that email is the ultimate social network.

If you take all the people I email with regularly, you’ll quickly figure out my top relationships. Then you can build a social network out of that.

And that’s exactly what Google did with Buzz. But what I learned from that experience is that email is not one social network, it is many. And it may be a mistake to combine them all together into one network.

I’m writing this on a plane on my blackberry. I can’t easily link to to Miguel Heft’s piece in today’s NY Times, but I can quote from it. He says:

“Email, as it turns out, can hold many secrets, from the names of personal physicians and illicit lovers to the identities of whistle blowers and antigovernment activists.”

That was my experience with Buzz too. Many of the people that Buzz suggested I follow are people that I would never think of following on Twitter or friending on Facebook.

One of the more recent lessons for me in the world of social media is the value of the explicit connection. At one time, I thought that automatically building my networks would be useful.

But if you compare Google Latitude, where you broadcast your location, to Foursquare, where you explicitly checkin to a location, I think you’ll agree that the explicit gesture is better.

And so it turns out that implicity deriving social relationships is tricky and potentially dangerous. That doesn’t mean the idea isn’t powerful. So was Facebook’s Beacon. It just means you have to roll it out very carefully.


Comments (Archived):

  1. chipcorrera

    I agree that everyone maintains different networks with various purposes. However, I don’t think that it is a mistake to combine them – in fact, I believe that you can achieve some powerful results when combining. I think the point is that you should be in control of explicit combination, and that this should not be usurped by the enabling technology – especially unknowingly.

    1. maverickny

      Integration of tools and platforms can indeed be very powerful, but when you allow geeks and nerds to play with algorithms without thinking through the human element and potential consequences, you have a recipe for disaster in the making. It seemed as though privacy mattered very little and the social tool was rushed to market.There were some sad stories in the news and blogs yesterday about women exposed to abusive exs via Buzz, for what should have been a private not public matter.

      1. Tereza

        Yeah to me it triggers not just a murky creepy feeling but actually a flight response. You might be surprised how common stalking of the ‘fairer sex’ is. Thanks Buzz for making it even easier.

    2. raycote

      That was shot and concise. When you put it that way it makes me think of Google Voice. I’ve not used it and am therefore talking through my hat here. But is Google Voice not all explicit in grouping the way you control why connect to what and when.

  2. laurentk

    That’s your experience, but there’s a good chance that people I email the most are people i’ll follow/friend on socnets. *But* they should be suggested more than automatically added.However, for mainstream usage (which is what Buzz is aiming at), email might be the fastest way to build your buzz community.

    1. fredwilson

      Maybe. But I email a lot with my private banker and real estate contractor. Not likely that I want them in any of my social nets

  3. Simon Cast

    I see this as where developers value the algorithm over the system. Social networks exist only within broader context. Without understanding what that broader context mistakes will be made. And increasingly dangerous mistakes.I’m not sure that algorithms will be able to adequately determine the context of a social connection to make a valid decision on a connection. Social networks change, evolve over time connections form & fade sometimes very quickly. None of the social network services capture the dynamic nature of users networks.The danger with this implicit friending/following is life is messy and as you pointed out Fred, a maze of interconnected networks where the only join may be you. I don’t think it is potentially dangerous, implicit networking generation is very dangerous. Particularly with something as personal as email.

    1. fredwilson

      Yup. That’s the big insight I got from all of this

  4. kidmercury

    they just need to get the default settings right, IMHO that’s the trickiest part. the killer thing about email is that it not only maps out the connections but provides the most meaningful text to analyze and classify the nature of the relationship.and of course buzz is not really geared towards a nation of early adopters like fredland, in much the same way RSS readers are not for mainstream web surfers.

    1. ErikSchwartz

      I really agree with kid here.They messed up the defaults badly.The data wrapped around email (especially the 2nd or 3rd derivative of that data) is a gold mine of contextual value. The reason I am far more bullish on Facebook than Twitter is while twitter knows who I am connected to, Facebook know WHY I am connected to that person. Facebook knows if it’s someone I worked with, or someone who went to school with my wife, or the spouse of a friend. That context in my mind is incredibly valuable.Buzz isn’t there yet, but they have the tools and data to get there.

      1. fredwilson

        I think its instructive to explain how my daughter in college uses fb and twitterFb is for social networkingTwitter is for following celebsThey are very different use cases. The twitter use case is more like broadcast media

        1. Tereza


        2. NICCAI

          I think that’s how most twitter users use FB – for “friends.” The friending phenomenon in general is an interesting dichotomy – especially for FB. On one hand, there is your count and connections (call it your popularity). On the other, there is your privacy. I’ve been saying for a while that anonymity will be the wealth of the chosen few 50 years from now.Moving forward, FB’s achilles heel is noise. Hotmail had the same problem – spam. I know I can’t be the only one on this list that is dangerously close to unfriending a number of lonely farmers. The one thing I will say for Buzz is that I trust Google has an even different end game. I think they are feeling out the sweet spot by running data – social rank.

    2. ShanaC

      I do not early adopt a lot stuff- I don’t understand everything. And a bunch of stuff turns me off. Other stuff scares me, and other stuff I would adopt instantaneously I think a lot of people here are on a spectrum depending on what the item in question is.I think we have to be more careful when talk “crossing the chasm” I think there are many chasms to be crossed for each individual depending on the product at hand.

    3. fredwilson

      I don’t use a rss reader and never have. I couldn’t get friendfeed as hard as I triedMaybe I don’t belong here in fredland

      1. kulsingh

        I wonder if you did not get into either of these products because neither are topic-based? Personally I separate people and websites into categories (work, social, etc)…and I give attention to each when in the appropriate mindset. But when you put it all together in a single stream the value decreases. In some ways, this is the complaint with Gmail / Buzz. Just wondering since I have the same reaction to rss readers, friendfeed, and buzz.

        1. Tereza

          I have a lot of crossover between work and friends.Sometimes I’m in a ‘topic’ mood (though there are multiple cuts of topics, some overlapping) and there are also geo cuts, time series cuts and others. So silos if ill-defined might hide people I should be talking to, even of they’re right under my nose.More times than I can count I’ve targeted a business contact and relied on my business network which was say 3 degrees out when in fact the personal network was 1 degree and very warm, and I didn’t even know it. The real goodness is when you discover those gems you didn’t know were right there because you were looking the other way.For me personally work and life is better and better when there is crossover/multiple touchpoints. More productive and more pleasurable too.For example another mom whose kid is in my kids’ school, we share a committee, and she is an exec w hot startup and we can trade contacts and ideas at karate pickup.Not quite sure where I’m going with this except to say that relationship categorizing is highly personal, dynamic and organic. And we may not have the bandwidth to stay on top of how they ebb and flow, especially at a high volume of connections.I feel like a place to start is simply to ask:–what are my relationship goals right now? (ie expanding to new, nurturing/deepening existing, or pruning too many). And then also….–how many can I personally handle? (there is a broad range of preferance on this)–do rigid or loose categories work for me–what is my preference to manage it personally, vs. do I prefer to largely automate it.Golden rule: Start from the user and work backwards.

          1. kulsingh

            Very fair points. My specific context was in curating / aggregating the web as I expect that is the reason most people use rss readers and friendfeed. I say this since I use (and have heard others state the same) Twitter for work related web curation and FB for social web curation. Put it all together in one stream and it gets frustrating. Your points in terms on the network are valid but it is also why Buzz is controversial. We each use our networks differently and to mingle as a default your bank loan officer with your friends into one network because you email them both is unintuitive.

          2. Tereza

            I’m totally with you on that.

  5. David Semeria

    The key question is how come the big G didn’t see this coming…

    1. mrcai

      I find it amazing that Google managed to get this wrong. Someone, somewhere, should have said no.. this isn’t what people want.

    2. zerobeta

      Methinks they did but decided it was a necessary evil. Social Networks need people there and to appear active for people to start using them. Thats why they did the auto-adding. It was more to suit their needs than yours (which usually results in a negative user reaction).

      1. fredwilson

        Twitter has struggled with this too. The ill conceived sul is how they approached it. That wasn’t so good either but at least it didn’t create a privacy backlash

        1. zerobeta

          Twitter needs to be able to suggest users to follow based on who you follow and what you tweet about. Ex. If you follow @fredwilson you may also like @howardlindzon or @bfeld. Or, if someone whom you communicate with regularly just added a new follower, it throws them on the list as well. I believe there are external apps that do this but they need to internalize it because only a small percentage of Twitter users will ever seek out an external app.

          1. Laurent Boncenne

            Not good, what happens if you have no interest in who your friends follow ? I wouldn’t want to get spam about bfeld if I’m not interested in him. Alternatively, my friends wouldn’t want to have suggestions based on who I follow. None of them have any interest in VC or design or even entrepreneurship. And I certainly don’t want to have suggestions about accountants.

          2. zerobeta

            It wouldn’t be spam, but just something on the sidebar that updates every time you click on something (like Facebook does). I don’t see that as too intrusive.

          3. Tereza

            LinkedIn too. I’ve picked up some great contacts that way. Just have to be able to opt in.

          4. fredwilson

            Yes. That’s how it should be done. The way facebook does it

          5. fredwilson

            Yes. That’s how it should be done. The way facebook does it

          6. Tereza

            I’d really like that, so long as the workflow doesn’t overwhelm (ie it’s batched) and also opt-in (both the capability and the people selecting). Way more useful and less creepy than the email triangulation.For someone who is in sales/marketing or an entrepreneur Fundraising, ID’ing targets you don’t already know is the name of the game. And quite inefficient. You could pick up some gems which might take you months of in person/analog networking, which can be painfully linear.But again opt-in is key bc the people who want it really want it (and possibly would even pay for it) while the ones who don’t may well run in the other direction.

          7. fredwilson


          8. fredwilson


    3. rick gregory

      There are three separate issues, only one of which is the defaults:1) Having an engineering culture drive a social product…. And I’m not just trying to be snarky. Social isn’t a technical problem it’s a people one and Google’s culture doesn’t seem geared that way.2) Layering a social sharing product on top of a private communication product. Why would anyone assume that I’d want to do this? I might email business partners, clients, doctors (mine), lovers… why would you ever assume that frequency of email connections maps to ‘want to share with the world?”3) The settings. I despair of us as an industry sometimes. Ever since the earliest concerns over privacy a decade ago, people insist on ignoring obvious things – you don’t opt people in just to build an audience and force them to opt out. You don’t auto-subscribe them to a bunch of followers then make them removed those people one by one. You don’t suddenly violate customer trust by changing the nature of a familiar product to suit senior management. These aren’t hard or obscure lessons… but time and again otherwise bright people screw up by ignoring them.

      1. Mark Essel

        Maybe they just want to see which people will use it.

        1. Carl Rahn Griffith

          That just about sums it up, Mark. The thing that gets me is: are we all supposed to be perpetually trying/testing out the (possibly, maybe, could be, if) ‘next big thing’ for the ‘net?We wonder why Corporates, Government, etc, is slow in take-up on our evangelising of all things Web2+ (or whatever you may wish to call it) yet they can see the constant stream of often irrelevant new ‘products’ and trends, 99% of which never stick because they are ‘solutions’ or services looking for a problem or need that often doesn’t exist.Little wonder.Google just seems to be replaying exactly what Yahoo did, throwing lots of stuff out there and seeing what sticks – if anything?I love our industry (most of the time!) yet its immaturity and unprofessionalism – even at Google’s level – frustrates me at times. Rant over 🙂

          1. David Semeria

            Good rant Carl – but something tells me there’s a lot more where that came from….

          2. Carl Rahn Griffith

            Lol. I had a good long discussion regarding this topic with @octane (a very cool, clever chap, Wayne) about it the other day … Just how many bloody web sites are we supposed to visit every day? How much free time is it assumed we all have to play with all these tools?How many ‘Social Networking/Media’ (urgh! to that damn terminology and the whole pseudo sub-‘industry’ it’s spawned – I use the term ‘industry’ very loosely) new tools are we supposed to adopt and embed into our lives? It all smacks of desperation. Lotus 1-2-3, Mozilla, et al, weren’t borne from a “let’s do this and see what happens” approach. They were created to address a known, quantified need. The sooner a lot of talented developers focus on the basics of the real-world, and address some real-world opportunities and needs, the better.And don’t even start me on the likes of TED!Anyway – it’s Saturday evening – time to ‘socially network’ at the village pub, in the real world, lol ;-)(Sorry to go off-topic a little, Fred!). Have a great weekend, one and all!

          3. David Semeria

            Aye, O tempora O mores…I knew there was more – as the Italians like to say “I know my chickens…” :)Have a pint of Tetley’s and think of those poor souls who have no access to the nectar.

          4. Carl Rahn Griffith

            I’ve been reading too much Charlie Brooker vitriol and angst today (@charltonbrooker), and also watched some recordings of his TV shows (the superb newswipe and screenwipe), back to back. Lord knows what he’d make of our profession.Tetley’s? Shock, horror. This is firmly John Smiths territory, David.Anyway, I am in a Strongbow mood tonight. Makes summer seem closer. 🙂

          5. David Semeria

            A Yorkshire cider drinker? Thread stops here.

          6. Carl Rahn Griffith

            Mea Culpa.But, in my defence, I am a ponce from Warwickshire, David. Although Warwickshire is firmly in the Midlands and at the very heart of England, many Yorkshire folk consider this region to be the ‘Soft South’ so my being a cider drinker is to be expected. My better half, Helen, is of true Yorkshire stock, though. Which appeases most of the locals. Most of the time. 😉

          7. David Semeria

            Um… the jury has retired to consider its verdict.It may well be irrelevant, since we both risk banning for being so far off topic 🙂

          8. Carl Rahn Griffith

            Please be gentle on us, Kid’ – we’re first time offenders.Anyway, David started it! 🙂

          9. David Semeria

            Kid, dude, we’re pals right?…

          10. Mark Essel

            Ha, epic thread. I love the OT (perhaps a little lament at my old drinking days…)

          11. Mark Essel

            Appreciate all your perspective brings on the “new for new sake” tools. Maybe Googlers just really loved friendfeed and wanted to emulate it?

          12. Tereza

            Totally agree Carl.I love trying new things. But it exhausts me the degree to which the vast majority are a colossal mismatch with my native environment.Improved gender and age diversity would go a long way.

          13. fredwilson

            We should do a road trip to your village pub. Anyone who has left at least 100 comments here in the past year can come. Would be fun. Maybe catch a swfc match too

          14. fredwilson

            What’s the best time of year weather wise and swfc wise?

          15. Mark Essel

            I can be designated driver.

          16. fredwilson

            There is

          17. Aviah Laor

            “are we all supposed to be perpetually trying/testing out the (possibly, maybe, could be, if) ‘next big thing’ for the ‘net? “. Yes 😀

        2. rick gregory

          Maybe, but “Hey, let’s toss out this half-baked idea, violate peoples’ trust, expose connections that are assumed to be private because, what the heck, we wanna see if it flies” doesn’t seem even remotely professional. It sounds like someone who either doesn’t get or doesn’t care about social norms and what their customers want.As I said, I despair that we’ll every get the really basic stuff right. People want to control what the expose, they want to feel in control of their world in general. Offer me Buzz? Sure. FORCE it on? No. Suggest people that I might want to follow? Yep. Autofollow them and publicly announce connections? Um no.What’s so hard about this? Really, I’d like to know what the hell is so hard to get about the fact that people will always be more engaged if they opt-in, that if your product doesn’t compel them to opt-in that it might be your product that needs work and that my data should be under my control. You don’t toss all that aside because some engineer wants to see what will happen. If some tiny startup run by a 22 year old kid who’d not lived through the growth of the web had done this I’d chalk it up to youth, inexperience and perhaps geekdom, etc. But for Google to do this shows either contempt or cluelessness.

          1. Mark Essel

            They certainly f’ed up the release, but are working on the correction fast. I don’t expect any more or less out of Buzz even though it’s coming out of Goog (Buzz could be coming from a startup for all I care). As long as they listen to users and correct course I’m satisfied.

          2. rick gregory

            I do. I expect better out of everyone as we go along *especially* from people who’ve been around a bit. Frankly, had this come out of a startup run by 2 20 year old kids it would be more allowable – they don’t have the same experience and knowledge. Google *does* have that knowledge – they’ve been around long enough to have seen others do these privacy mistakes and should know better by now.I don’t expect perfection, but I expect that people don’t repeat the same, obvious screwups time and again, especially if there were around to witness them before. It’s called learning – and I expect any decent entrepreneur to do it and not simply tread the same ground that’s been trod before. In short, I’m not satisfied simly because something’s shiny and new – I expect things to be better than what’s come before and to mostly avoid the mistakes of the past. Buzz fails that test.

      2. David Semeria

        To quote a recent @timoreilly tweet on the same subject:Samuel Johnson: “Nothing will ever be attempted if all possible objections must first be overcome.”

        1. rick gregory

          Oh a pithy quote! Well that makes it all better. Oh wait, no it doesn’t …Neither I nor the others who dislike Buzz are arguing that all possible objections be overcome – we’re arguing that it’s foolish to make the same mistakes over and over. That some principles are well-known (forced optin and lack of control over one’s environment are bad choices) and it’s foolish to violate them, especially when you then act surprised about objections.We shouldn’t have to keep relearning the same damn lessons over and over and over. Or, since we’re into quotes:”Whatever is worth doing at all, is worth doing well.” — Philip Dormer Stanhope, Earl of Chesterfield

          1. David Semeria

            Touché.Clearly, the idea to piggy-back on Gmail was commercially astute – but it’s also clear it could have been executed much more intelligently.

          2. rick gregory

            pppfffttttt agreeing is SO boring :)I’m still unsure about layering social sharing and conversation over a private messaging product. Even had they done all of the opt-in, let people follow as they want stuff, I wonder if the two mix. The Friendfeed-like experience is great for some people, but others just want to share quickly and not engage in a conversation all of the time.

          3. David Semeria

            Take it as a compliment!Many people have pointed out that Buzz takes one type of social network and tries to transform it into another. It’s like walking into a bar with your boss and your boss’ boss only to find your old college buddies at the bar. Embarrassment guaranteed.Another common comment is that all this serves to reinforce the “quality” – and perhaps monetizability, if such a word exists – of FB’s and, to a lesser extent, Twitter’s maps.Sorry for agreeing, it doesn’t happen often…

          4. Mark Essel

            I think we’re doomed to repeat ourselves ad nauseum. The timing for various stews of technology part of their magic. I’m excited about all the unnecessary mashups we see fired out by developers with increasing speed. As a gent who love’s choices, another social channel doesn’t get me the least bit ticked off. Even though they just may be swallowing the space I’m developing in.

          5. rick gregory

            No we’re not. There’s no Fates forcing our choices – we choose our fates. All of the Google mistakes on Buzz are annoying *because* they’re so obvious. I’m not excited about unnecessary mashups because the mere fact of new and shiny doesn’t cut it anymore. Things should be new, shiny and *compelling* not just iterative me too products.

          6. Mark Essel

            Copycat products are merely a starting point. The real question has been, and will remain to be “what will you do now?”. I agree we control our fates, there’s no market lock in so we choose what suits us best. You can choose to use or not use Buzz, Facebook, Twitter, a Phone, etc. I think things get really exciting when tools are crafted with end user customization in mind.Social media popularity contests (ebb and flow of new social nets) have been the norm for one good reason, users don’t retain and control their networks and identity.

          7. raycote

            The disservice of a food fight is that a lot of food gets wasted.In case of software development however, the food fight serve to accelerate creative synthesis which hastens the evolution of mature products that embody optimal feature sets and interfaces.I known I’m coming a little shot on this post – sorry 🙁

          8. rick gregory

            I’ve been in 2 startups and around tech for 25 years… I get all that. But that is NOT an excuse for making basic mistakes over and over. Forcing optin, forcing followers etc are all mistakes that anyone who’s been paying attention knows will get strong negative reactions. This isn’t new and it’s silly to see people acting like it is – sorry guys, but all of the new creative destruction talk isn’t new. It’s old. Not learning from the past isn’t cool, cutting edge or web x.x… it’s just stupid. Don’t let the past limit you…. don’t figure something can’t be done because it hasn’t. But don’t act like you’re the first people to ever develop software either. Especially not if you’re Google and your company’s been around for a lot of this.I *DO* have to give them credit for reacting fast and redoing some of the t hings like following – that’s good to see. But virtually all of the negative reaction about privacy could easily have been avoided. The didn’t either out of a cynical ploy to get a lot of users quickly or because they genuinely don’t get social software past the feature set level.

          9. raycote

            I agree with —> zerobeta <— above ^^^^^^^^Google is not learning the same damn lessons over and overGoogle is making the same damn cynical calculation over and overPrime the pimps, sorry I mean the pumps, then clean up the mess later. There is also the “no advertising is bad advertising” win by winding up the BUZZ!It all reminds me of a quote from my dad to me and now that I’ve turned into my dadI’ll pass it on from me to Google “YO…YOU MUST BE STUPID TO THINK WE’RE THAT STUPID” less the angry tone usually accompanying my dad’s delivery style, a tone that was most often richly deserved.C: Social is presently at the DOS COMMAND LINE levelI have no pressing need as of yet, so I’m kinda dragging my feet waiting for the mouse/windowing version interface.I think Google should have been a little more social about picking the name. OK…OK… I Know… it is there product. But they could have picked BUZZ and then let us all chime in with suggestions for variations on the BUZZ theme.My suggestion would have been:TWEETY BUZZBUZZ TWEETSIn hind sight maybe:BUZZ KILLBUZZ CUT

          10. Wavelengths

            So, “Google is making the same damn cynical calculation over and over . . .”What does “cynical” have to do with real social connection?As Fred said, a couple of days ago, “currency is trust.” And I see a breach of trust, which is social currency.Where is the heart, and the soul? We don’t generally choose our close social network for their cynicism and greed.I wonder why Google didn’t “get it”?

          11. raycote

            I think we agree. I meant cynical as in compromising their integrity in order to quickly priming the user pump for competitive commercial advantage.

          12. Wavelengths

            Yep.And how can you trust someone who clearly violates integrity, and puts people at risk — of embarrassment, if not literal physical risk.With the magnitude of the reach of Google, people should be paying very close attention to the way they have abused that access to data.I think consideration for privacy may be a whole new area for developers to explore. Allow people to have safe ground, and then encourage them to open up to new connections.This event makes me a little nervous about my Android phone and my gmail account.

        2. Mark Essel

          Quality quote. We do what we must. I tick off a good fraction of the folks who read my posts. It’s the nature of opinions, and their external realization in design.

        3. Aviah Laor

          small mistake, sorry, you probably meant @samueljohnson

        4. Aviah Laor

          @macbeth: the witches are killing me

          1. ShanaC

            It’s the scottish play!!!! NU.

          2. Aviah Laor

            @darth: I’m your father Luke

          3. ShanaC

            @Luke “Nooo”

      3. fredwilson

        Great critique

      4. Keenan

        rickg, spot on! They needed a “webcologist”The list of companies screwing up because they don’t understand the psyche of people is long. Especially in the online world.I remember back 1997, when CVS tried alerting it’s customers it was time to refill their prescriptions via email. Customers freaked because they didn’t want to risk others seeing what they were taking.I argue online/web based companies should create a position called a “webcologist” This person reviews the psychology behind a social, or interactive feature as it is being developed. I wrote about it here http://asalesguy.com/2009/1….If companies spent a bit more time understanding the human psyche and how it relates to their products they could avoid mistakes like this and create features/products we will be more inclined to use.It’s about knowing how we behave.

    4. Carl Rahn Griffith

      I see Google going exactly the same way as Yahoo.

  6. Matt Hunter

    Question unrelated to the topic: Did you leave your Nexus One for Blackberry? Or are you still using both?(I’m contemplating ditching my Blackberry for the Nexus, but no keyboard or BBM makes it a difficult).

    1. fredwilson

      I’m using both and happily. I am replying to this on bberry because I can type way faster on it. But I use the google phone more than bberry. I am waiting for a google phone with a killer keyboard

  7. Brownstoner

    I couldn’t agree more. I’ve probably opted out 75% of the people Buzz has suggested for me. I don’t want or need to follow all the people who’ve emailed me with tips, ad inquiries, etc; nor do I have the bandwidth for another major social media commitment to keep up with. I’ve settled nicely into a world where I use Facebook for my personal relationships and Twitter for professional. As of now, Buzz seems completely superfluous.

    1. fredwilson

      That’s exactly how I do it. I was on fb this morning with my son. I was on twitter with a bunch of professional contacts

  8. muratcannoyan

    Not everyone was going to be happy about the way Buzz was integrated and Google had to know this. They could have taken a more conservative approach but didn’t. The relationship I have with Google has changed.

    1. kidmercury

      that is true, goog did diss their userbase, willfully, and with profit as the only real motive, it seems to me. i wonder how much longer they will be able to get away with such behavior.

      1. muratcannoyan

        I guess it is possible it was just a poorly executed roll out and they really didn’t realize how it was going to turn out, but I doubt it. Their model is so dependent on trust that it’s very surprising they would consciously do anything to potentially damage it. Time will tell what they can get away with.

    2. fredwilson

      Well they rolled it back very quickly. I like how fast they moved

      1. muratcannoyan

        Heck I love Google’s products and the incredible speed that they iterate but the explicit/implicit social balance is personal and should be treated as such, at least for the time being.As a side note comments can at times be so complete and sharp that there is an impulse to share them.

  9. andyswan

    This may be one of those things that is true until it isn’t. i.e. when someone figures out how to make a social application as flexible and powerful as email. Seems like facebook is on its way.

    1. howardlindzon

      Exactly andy – email is where I work, not play so facebook seems the only one with a logical chance to win me. more here http://howardlindzon.com/my

      1. Aviah Laor

        but isn’t Stocktweets on the blur social/work area?

        1. ShanaC

          A lot of people have that problem- you work and socialize with the same people or variations of the same people…

    2. fredwilson

      Yup. Facebook is headed there. But they may lose the youth crowd to something hipper or freakier. We talked about one such candidate here last week

      1. Tereza

        Yeah. I’m also keeping my eye on Lockerz.Tereza

  10. steve cheney

    People want to “separate their worlds”. There is a reason Facebook and Twitter have totally different audiences and will both succeed independent of each other.Buzz also needs to have access points outside of gmail or it will fail (does API support this?). Though gmail is popular, not all folks access through a browser (IMAP etc). The appeal of Twitter is its rich variety of clients – e.g. Tweetdeck running separately vs web-based access etc.

    1. fredwilson

      Yes the api supports it. I suspect we’ll see all the twitter clients support buzz quickly

  11. zerobeta

    Google doesn’t get social the same way Microsoft didn’t get the internet. It’s deep, institutional, and I don’t think it’s changing any time soon. The “me-too” attitude by which they’re approaching it is eerily similarSure they know who you are and who you talk with but BIG DEAL – so do you. In fact, you know this information much better than Google. As you point out, they think this implicit way of driving relationships is the value they *think* they add, but there is very little value in that in general.I could go on forever, perhaps a blog post in the making.

    1. fredwilson

      If you write it please alert me to it

  12. maverickny

    The concept of Buzz was a good one, poorly executed and ill thought out. I was appalled to find that Google decided to allow Buzz to follow various people by default that I wouldn’t want to be connected to on social networks (my accountant, my ex, my clients etc).Blocking spammers also failed to execute, despite repeated attempts. Unfollowing people in Buzz also unfollowed their Google Reader shares, which I did not want. The whole thing was a mess. After 3 days I was thoroughly fed up with it despite enjoying the science and tech convos.Allowing people to choose who to opt in, rather than opt out, is always a smarter idea.

  13. scyphers

    Implicitly deriving relationships is exceedingly powerful; in forensics/intelligence/etc, traffic analysis of even encrypted messages often reveals more than the contents of the messages themselves.

    1. raycote

      forensics/intelligence – those would be the past tense dangers associated with implicitly derived relationships

    2. fredwilson

      Yup. Powerful for sure

  14. luisa elena vidaurre

    There is a big power too enhanced in facebook, it is definitely a tool to fight for freedom and human rights in countries who uses law to disminished and oppress citizens. It is also very usefull to generate awareness and a ” conscience of our rights”. Take for example -Million voices against Farc- in Colombia, who was able to activate the biggest march ever, or A World withot Censorship- Un Mundo Sin Mordaza- who contributed greatly to stop the Venezuelan Goverment to close a TV channel: Globovisión, and the gag is used all over to fight against represion of human rights.

  15. Mo Koyfman

    as it relates to buzz, i do not think the implied network is particularly useful.i email people i work with, entrepreneurs, etc. and those folks are not necessarily the ones i want to connect with in buzz.it’s certainly not as intuitive as one might think.

  16. Jan Schultink

    Excellent post. Comparing to real-life networks I wonder what will happen online:1) A clever system to assume different social modes: work blog friends family, etc.2) Equalizing of modes: you share a bit more with work connection and a bit less with close family all across the same networkI think it will be 2)…

    1. Mark Essel

      Since I’ve gone Net loopy in late 2008 I’ve done my best to think about one genuine identity for better or worse. It does mean some slight self filtering (2) but ideally the people I deal with can be comfortable with my professional/personal mix and trust that there’s only one “me” on the web.

    1. fredwilson

      Not yet. I’ll check it out

  17. petekazanjy

    I don’t think explicit is necessarily better all the time. It can be, depending on the use case. However, explicitness often lends itself to being gamed by bad actors. I think what we’re talking about here is control, and when implicit data is surfaced, it deemphasizes control on the part of the ‘speaking’ user. That said, it can often be in the better interest of the information consumer…who can’t be gamed as easily.A straw man example is how one of my friend’s uses FourSquare. He checks in cool places that makes it look like he’s a happening man about town. But he may or may not actually be there, have ever been there. But consumers of that information can lose out if they thought it was real. Whether because they walk over to the place where he was supposed to be, and boom, he’s not there, or because they draw judgments about him that are unwarranted (wow! He got in at *that* club?). It’s good for him. It’s not as good for them.

    1. fredwilson

      Better yet, gps is going to rear its head and out him automatically

      1. Tereza

        That will be good on the other side. I can start getting credit for the many thwarted check-ins when I “uh-oh can’t connect to the Internets”.

        1. petekazanjy

          You still won’t get credit unless they cache checkins. GPS won’t help with that. Your app will know where you are, but won’t be able to kick that to FS’s server’s…

          1. Tereza

            Awwww, man!!! You burst my bubble. :-). I was so excited.

          2. Tereza

            I’ve also been disappointed that I can’t check in thru wifi? We have optimum wifi in our little downtown, no cell.They’ve been trying to put in a tower 9+ years but major NIMBY issues. I am sceptical it will ever happen.

      2. petekazanjy

        Nah, because he can check in as he walks by out in front / directionally near (FS already does this, kinda. But you can still be within like four blocks, and it will register).And plus, doesn’t that kinda prove our point: the service is going to pull implicit data without my explicit gesture? Ruh roh.Point is, explicit is good in certain use cases, but isn’t panacea. I mean, I know people who run around ‘friending’ as many internet-famous people on LinkedIn and Facebook in hopes that when someone looks at their profile, they’ll say “Wow, Pete’s friends with Fred and Jason Calacanis and iJustine!?!? Wow!” Problem is, that explicit gesture is taken by the world to mean way more than it really does. It’s explicit gesture spam.The result? Facebook specifically uses a ‘meta-graph’ internally in order to show you content about the people you are *actually* friends with (predicated on whose posts you *actually* comment on, and view their pictures, and so on? Implicit gestures.), because they, too, know that the explicit gestures are crufted up and getting more and more useless with time.Anyway, I still love to play me some FourSquare, but I think it’s important to recognize the value of implicit data. But just, you know, don’t go and show everyone on my FourSquare checkin page that when I was checking in at 24-hour Fitness, I was really at McDonald’s. ; )

        1. fredwilson

          When your friend gets a free espresso upon checking in out front, maybe he’ll come on in

  18. Bora Celik

    I don’t want to socialize with everyone. I even make an effort to keep my Twitter and Facebook connections separate, why should I all of a sudden start socializing with the random people in my email stream? Just doesn’t make sense. Buzz fail.

  19. Jon Knight

    Thanks Fred – I’ve long held the opinion that social networks nearly always need to be explicit, for the exact reasons you mentioned.I don’t begrudge brilliant coders their intelligence and efforts, but it does seem there’s an inordinate amount of effort put into implicit everything. From social networks to advertising, they try to guess who we are. More success might be found through simply asking.

    1. Mark Essel

      I wouldn’t mind a smart implicit default, that listens to user feedback and follows explicit direction. That’s how I’m building tools.

    2. fredwilson

      I agree

  20. Jay Leon

    Real people all have multiple social circles that overlap and interact in extremely complex ways. So far no one has figured our an elegant way to model the complexity of real relationships — and I suspect it will be a long time before such a thing comes about. A social media system fails when the makers think it actually replicates real life. So the bigger and more self assured makers of systems get, Google would be the biggest and most self assured, the less likely they will be to create a system of value.

  21. @davehonig

    I agree that the explicit gesture is excellent, but scaling will end up being based of implicity deriving social relationships. But as many of us know, many explicit gestures are not authentic. How easy is it to check in on Foursquare while not actually being at the venue. Sure this can be fixed by just matching up the check-in with the location. Lets discuss sell reported data. My thoughts is that self reported data for the most part is an avatar of how people want to be perceived not so much how they are. I think Google is on the right path if they can start scoring your audience (graph ) segments ( and dont assume you have one social graph ) score based on many variables ( sites visted, click stream data, email data, social data ) Look at everything horizontally. From day one, I believed the short url services like bit.ly would be the ones to figure it out with real engaging data. Cant help to think what would happen if they licensed this data ( a very strong social graph). There is so many other variables we can discuss. At the end of the day you are absolutely correct, Its is very powerful and yes, it would have to be rolled our very carefully. We shall see.@davehonig

    1. fredwilson

      Right. We will get there but slowly, not overnight

  22. kulsingh

    Fred, I see there being a critical differentiation between private implicit and public implicit. I think what you refer above is actually private implicit social graphs. Google went too far by taking users’ private implicit graphs and using it as their public graph. In this regard, you are absolutely correct.But public implicit social graphs are very powerful. The fact that people communicate on this blog suggests an implicit relationship. There was no explicit relationships identified between users until Shana was generous enough to create an explicit Twitter list. And from users’ reactions, it was clear there was demand and a benefit in identifying that implicit graph.In full disclosure, I say this because my company Tickreel leverages these public implicit graphs as part of its framework and I have seen how compelling that information can be. But I agree, if the same private implicit graphs were opened to the public, it would not be pretty. So I agree that Google went too far but suggest a delineation between private implicit and public implicit.

    1. Mark Essel

      I’m surprised you don’t link your Disqus avatar to Tickreel (or you blog). I’m interested in learning more about the value of public implicit networks, sounds related to the conversationlist that Kevin developed.

      1. Aviah Laor

        Victus Media – explicit or implicit?

        1. Mark Essel

          Explicit existing social connectivity, implicit topics edited/controlled by the user which can allow for new implicit connections.I hope that makes sense 🙂

          1. Aviah Laor

            It does. Seems like you mix the advantages from both alternatives. So when do you crush it?

          2. Mark Essel

            We have more work to do under the hood (identifying nodes user/tags/trending links) and plenty of interface work ensuring that the tool is both easy to use, intuitive, and looks good (we fail at pretty).We’ll continue to iterate over the next few months, hoping to have a sharper interface/backend in the next few months. Planning to meet with Robert Scoble and show it off in mid June (fiance and I are coming back early from our Honeymoon from Hawaii to SF/bay area).Last but not least, we can kick up an API but haven’t figured how to cover it’s cost yet (semantic api hits will cost us over a fixed level so we’d have to pass along their cost). The other alternative is hitting RDF databases ourselves and identifying objects in language (not looking forward to it but it could make more sense than paying $$ monthly forever for increased access to APIs).check out Victus Media if you’re curious. There’s a couple of how-to vids at our base site. Single users, and our ad widget is ready for prime time, but our lists are still under development.

          3. Aviah Laor

            Good luck! June is not too far, and you have a product and a honeymoon to launch! Take a look at 99designs.com, they are really affordable. I know nothing about semantic API or their costs though.

      2. kulsingh

        Mark, since we are still pre-launch have not been actively seeking to drive traffic to Tickreel since the site is still rather sparse…but very soon and appreciate your interest. We are close to launching and at that point would love to get your feedback. High level, we leverage public implicit networks within our algorithms to make the web less of a black box as we see there being a critical missing piece to the web’s operating system given immense changes over last few years.

        1. Mark Essel

          Groovy, look forward to checking it out.

    2. fredwilson

      That’s a great insight. Thanks

  23. gorbachev

    It really is a shame the email marketers have successfully lobbied for opt-in to be a four letter word. Can you imagine a world where opt-in was the default? We wouldn’t be having these discussions.Another aspect of what you wrote is that how come we still can’t “partition” our social networks. It makes sense not to do it in LinkedIn, which is a one-purpose (more or less) professional social network, but Buzz is completely different. I don’t want my personal contacts co-mingled with my professional ones, and I might even want to keep my family separate from friends. Right now if I Buzz that my work sucks and that my boss is a bloodsucking vermin (he’s not), I have no way of doing that without sharing it with my colleagues (assuming they’re following me). If I was into “alternative” lifestyles, there isn’t a way to share that with my friends, but not with my colleagues.For Google to not understand or do anything about it is inexcusable.

    1. Wavelengths

      I absolutely agree. And, as one example, anyone who has every been through a bad divorce or romantic breakup would understand why you don’t want this serious invasion of privacy.

      1. raycote

        So WAVELENGTHS your agreeing that we should be able to group our email contacts into separate Buzz channels so we can speak to different groups of friends on differ WAVELENGTHS

        1. Wavelengths


    2. fredwilson


  24. Aaron Klein

    One interesting thing is that they limited the social defaults to the people you e-mail through the gmail interface. As a result, they only recommended three people for me to follow – despite the fact that I use their SMTP server to send large volumes of gmail from outlook.I already had a “Google Profile” so it was annoying to instantly display those three people to the world. The bottom line of this post is dead on. If you can’t build a tool that makes it easy and fun for people to proactively engage, building it where they “passively engage without knowing it” doesn’t make for a vibrant product anyway.

    1. fredwilson

      Suggesting follows is touchy territory

  25. Mark Essel

    Since gmail has been scanning my mail for relevant ads forever I never treated it as a sacred and secure communication medium. It’s a channel like any other, that’s privacy lasts only as long as the financial security of the business that manages it.Am the only one that really loves Buzz? It’s another great communication channel. Heck I was a big friendfeed fan (when they were independent) and still admire the dev team. Buzz feels like gmail/friendfeed offspring. Maybe I just love tech mashups, this one feels dangerous.It’s pretty easy to edit your network, and Buzz is changing so you can be more protective of your contacts.As with most social networks, the value will come from the ability of the community to self organize and utilize the information externally (apis ftw).

    1. kidmercury

      i hear ya mark. i’ll avoid buzz for as long as i can due to my opposition to google dominance, but i gotta give props where props are due. IMHO this is a great first release that will lead to significantly better iterations. also the haters here are not the mainstream users that buzz will succeed most with IMHO, our perception is a bit warped here in fredland (although i think even many fredlanders will be impressed at what happens when you combine buzz, android, and google maps).

      1. Jason

        no matter how great it works with maps, places, mobile, the moon, etc, it’s not right they made us their social experiment, especially the mainstream non first adopters. i’m surprised you’re giving props Kid, they’ll be succeeding with mainstream by changing behavior, not because it went viral, like things going mainstream should. IMHO

        1. kidmercury

          i hear what you are saying, and i criticized google elsewhere in this thread for basically going about this in a rather rude manner by forcing it on much of their user base. but i also think platforms need to have certain freedoms, and the line between platform rights and user rights is blurry and subjective — perhaps similar to the rights of govts vs the rights of people. ultimately i think it would have been far more respectful if google rolled this out carefully and slowly rather than smacking everyone with it.but the competition is no better. they are google wannabes but with a much worse strategy and far inferior in terms of execution. so, i do give props to google for their ability to integrate their apps and deliver a solution that is far superior to what their competitors can do.to get to the point where platforms respect users and cannot abuse us the way google did with buzz, we need niche platforms, IMHO. i don’t think that’s a typical VC opportunty and so the mechanism for financing these niche platforms may be an issue, IMHO

          1. Mark Essel

            Fredland is one such niche community. All it took was Fred blogging for a bunch of years, and Disqus to get us cooking. Fredland doesn’t need VC backing ;)All it needs is an API, a virtual currency, and…a couple of diehard fans.

          2. Jason

            you call it default settings, they call it auto-traction.

        2. Aviah Laor

          FB was on the edge of doing the same thing, and they dropped after the angry responses

      2. fredwilson

        I’m with you kid and stick with what I wrote on friday. I also am impressed with how quickly they rolled back the most hated features

  26. Chris Phenner

    In 2003, Spoke and LinkedIn were my first professional social nets. I liked Spoke more than LinkedIn because it assigned an actual number to your relationships, based on frequency, response times and volume of email and IMs. It was message-based scrobbling before scrobbling had been popularized by Last.fm. My wife was a ’74.’ I loved the data that implicit offered. But I think LinkedIn won due to explicit-driven user experience factors.And I think online music offers a sense of history. Pandora beats Last.fm (in number of users) because Pandora relies on explicit input (your favorite artists, from which it derives ‘Stations’). But I’m sure the Last.fm crowd is savvier (and likely much younger).BTW: I recalled an SF Chronicle article from back then that I recall featured (a then-whopping 14) social networks. There are some *classic* Web2 quotes (including Mark Pincus).http://www.sfgate.com/cgi-b

    1. Mark Essel

      Thanks Chris for the share, will read it now :).*update*Very cool look back (2003 seems so long ago while reading that).

    2. fredwilson

      I love last.fm but that’s because I tell it so much about me every day automatically. That’s not a mainstream user, at least yet

  27. efliv

    When Buzz first dropped, many commented that Google finally “gets” social. After all the backlash, I’m left wondering if they ever did or if this bungle is a sign of the “strategy tax” forcing their hand.I’m hoping for the former because the implications of the latter are much more troublesome.

    1. fredwilson

      I said that. I still feel they are on their way but clearly we have to question that

  28. Umberto Milletti

    Fred, email traffic is one way to derive implicit connections, with the limitations you point out.But there are other ways to have implicit (business) connections:- Through your coworkers, company’s BOD (even if you don’t email with them)- Through your customers- Through your previous employersThese are harder to discover, but can be very valuable, especially for young workers who have not yet developed a large explicit network.An example of this technology: http://bit.ly/clhMjw

    1. fredwilson

      I’ll check it out

  29. Eric Leebow

    I disagree and sometimes agree to some extent about email being the ultimate social network, as sometimes the people you email are people who are repetitively seeking or selling something to you, yet do you, or would you share an experience with them in real life? Then there are friends who text message you, yet don’t spend much time emailing, or they’d rather call you or send you a message via chat. Email is a business network/social network. It’s interesting to think about though as email as the ultimate social network, yet everyone using different email services are on a different network, similar to mobile phones and instant messaging services. Google Buzz won’t be able to be used if your friend isn’t on Gmail, and they use Hotmail, AOL Mail, Yahoo Mail or some other e-mail service. If another mail service builds a Google Buzz like service, then they will not be compatible, unless there is a service that connects them all. If everything we do in our real lives can be sent back to email, then it can be the ultimate network. I think the ultimate of social networks is the convergence of media, whether it may be pictures, video, text messaging, geolocation, email, messaging, video messaging etc. rolled into one. If you can take a little bit of everything and put them together, then you’d have the ultimate social network, yet it sounds good in theory, it doesn’t always work in practice, as people tend to be stronger in some media areas than others, some people care about sharing their location, yet won’t share a comment, or vice versa. It seems like Buzz is doing a good job at combining things as a mashup. The person someone is having a hidden relationship with, may be in their social network, yet they are not detected on a service.

    1. Mark Essel

      Interoperability of social data is a big part of Buzz (I think it’s the end goal). They don’t want to own our social graph, they it to be available anywhere we want to use it as they believe, and I think correctly, that it keep us engaged on the web more.Here’s where I got that perspective Eric: DeWitt Clinton’s post

  30. Aviah Laor

    Great insight. The tech and social aspects of social networks explained in two succinct terms. Before you judge something you have to understand what it is. these terms were really missing from yesterday discussion.Anyway, B2B/B2C out, Explicit/Implicit in. You will get a Wikipedia item for that “The term was first coined by …” I can clearly see the VC-founders talks: “Ah, it’s implicit network? well, we are investing only in explicit networks now” 😀

  31. Wavelengths

    Perhaps this Wikipedia description of “social network analysis” will be interesting to some people in this discussion. I wonder if anyone at Google thought about Buzz in this way.http://en.wikipedia.org/wik

  32. Aviah Laor

    Ah, I’m sorry to nag about DISQUS but now I have the correct words: DISQUS is a new type of Explicit social network. In Foursquare you Geo check-in, in DISQUS it’s comment “check-in”. It’s not derived, it’s Explicit, and it’s power that it’s VERY explicit: Where you comment, to whom you reply etc. It’s not just a tool but a different social pattern. What they should do now is to mine the DB and create the social graph.

    1. Mark Essel

      It’s publicly available information and it’d be great candidate for an API.

  33. markslater

    Fred – have you checked in using Buzz? its very different to latitude. Also the dynamic couponing is amazing, as is the “where you friends are” function

    1. fredwilson

      yes. it was the first thing i did on buzz. i mentioned it and linked to the checkin on my friday post. here’s the checkin. http://bit.ly/aQDFbw

  34. Aviah Laor

    from Ycom :”Automatically followed by abusive ex-husband (and his friends) on Google Buzz” (the link is blocked now).Implicit.

  35. Laurent Boncenne

    reading the comments on buzz , and on your blog, there is one thing that might be useful (at least to me). When you end up with a blog post with more than 30 or 40 comments(which happens nearly all the time), it ends up being nearly time-consuming rather than being a learning/sharing experience to read the comments.It would be nice if by default comments would be hidden (in buzz at least, having mashable or any other social media related site on your buzz timeline is a pain because there is too much action), and after explicitly watching those comments, you would pick the commenters who provide value to a post and allow them to be seen by you because these might add value.I don’t know if I make it easy to understand what I’m trying to say, but for instance, the comments about the superbowl didn’t add any value to me… (I’m French by the way, so yeah, I find it uninteresting to see anything related to that….) The ones really about Time Value of Money did.This would be interesting to see what actors in the field of commenting systems (disqus, echo or intense debate) do about it.I love your blog, the posts and many comments, but some aren’t as valuable to me as others. Seeing what is popular now is not enough when there are 130 comments to scroll through and find something valuable. Buzz makes it worse. Maybe that’s a field where it could play a role.

    1. Mark Essel

      You have echoed a need I’ve seen many times in my web wanderings, self controlled filters. You should have all the dials you wish for modifying the flow of content, comments, and reactions to your inbox. I wrote about it early last year and it’s a driving long term goal of mine. It’s one of the spaces I’d like to work towards with Victus Media, but it requires many earlier steps to build up an unstoppable business to fuel that type of work.1) The first step to providing you with this type of tool is getting a handle on meaning. Understanding relevancy and associating connections is the first major functional requirement. This is a big problem as you can imagine. There is so much context we can pick up on as humans that machines have trouble understanding. Basically we have huge association databases from a lifetime of learning, where machines may have only fixed methodic databases (DBPedia, RDF data, and the like). Cultural context is also a HUGE part of this problem. There’s not one answer, there’s thousands or millions of required interpretation engines.2) The second step is memory. Software needs to remember not only what content I like, but who I like in that context. Topics, people, and their associations to my dynamic interests are what the fading memory should be capable of retaining. The interconnection of relationships is what must be stored, maintained with real time updates, and be completely modifiable by me.3) Language. We should be able to easily extend a virtual search assistant to gather and process information independent of the base language it is written in.I could go on and on but you probably get a feel for how excited I am about this area Laurent.

      1. Tereza


        1. Mark Essel

          Thanks Tereza, always appreciate a good “Like.” 🙂

      2. Laurent Boncenne

        This might just be one of the toughest replies I’ve had to write so far.This subject is so wide and yet so specific about one concept of the “social web” of today that, narrowing it down to simple – yet understandable stuff – is probably too hard for me at this point.I’ll try anyway.I think that first; the database problem of today is probably not the hardest part as it is more a scalability issue rather than the ability to store millions if not billions of meanings. As the web evolves and the open source community grows, having providers of open source databases share seamlessly their content is something that might be doable.Getting a handle on meaning is just a further evolution of a complex db query. As far as I know (then again I’m not a software engineer or an AI specialist), for a computer to understand anything will always be bound to the laws of 1 and 0. This only leaves the possibility to rethink how a database is presented to us (as of today: rows and columns of cells). It might be interesting to look at it this way: the computer-brain side of the db would store the basic data it needs to perform a search (most likely keywords) in one row and would call for the following one which would store the data humans are used to interact with (semantic here ). The problem here is that it calls for a need of some sort of objects oriented database system where multiple rows could be equal to each other with a relationship between each value of its columns’ cells. (apple has many meaning for a human but on the language side and computer ones, it is clearly defined) You would have on the rows some sort of matching like a “da Vinci code cryptex” where each value has a meaning that is unlocked if it meets some conditions (here being the notions of languages, context, people, etc.)This might simplify somehow the problem to let meaning, context and people interact with the computer-side meaning of the data.The concept I believe you are talking about is somehow, how we could have both a computer-friendly brain and a human-friendly brain; allowing us to let computer process the data and use the human-friendly way of showing/interpreting it to feed us with the “real-time intel” we want/need/seek (which this very last sentence illustrate pretty well as it is probably to hard these days for a computer to chose between the 3 options for us). Do I want it? Need it or Seek for it?After reading your post ( for those who didn’t click through), I believe that the connection between those 3 parts you talk about above are – sadly – to this matter, completely related as it is probably impossible for any of those to work in the context of virtual assistants.Having any sort of database to store the data is one thing, plugging in the people on sure is a complicated part I believe, as it would seem to call for “opinion” regarding that user.The cultural context here, and the ability to sort through it is one powerful asset but it could also alter the initial meaning of what is intended to being said. There comes the processing of languages which I believe are not that complex if you look at it in some sort of a “pattern”. The difficulty is to enrich one language into another to reflect what is at stake. (I think it’s tough to give any viable example for this, but if we take Fred’s blog here, having the posts in French would probably alter the core meaning of what Fred intended to say in the first place)While memory is something computers have, there is not anything that can actually relate to how we (humans) function. While going through historical data and making sure that you don’t forget anything (actually the virtual assistant), we humans tend to forget things. And this might be for a good reason. Such software should be able to remember a few of the key things we need, because remembering it all would be wasting – somehow – the brain capacity of selection. By forgetting, we select. The software should perpetually be learning about your wanderings, instead of storing everything. What I do now is definitely not what I was doing last year. Another thing that might be also important to look out is the way this system would learn, by being autonomous (what you suggest), rules need to be clearly defined and I think that thanks to crazy movies (or books), we might have some nice user input for free. Weren’t authors like Asimov thinking of such a thing before? It is, I think, important to look forward and see how this can evolve in order to control the possible side effects of such a tool.I believe this is not about how it should process the data that now counts in terms of web standards (when you talk about page rank which might be wrong, given that you might find “serendipity” while looking for something completely unrelated. It is that serendipity effect that we can take advantage of that computers can’t do for ourselves now).All this seems quite complicated, and what I was suggesting was some sort of a way to search through what comments are now and sort it (visually) to let met focus on what is giving value to a blog post, and let me decide if I want to participate in an alternate conversation.Then again, as I was talking to a friend in front of some pints this afternoon, the discussion started from salary to driving to food to commenting on a blog to what the view of entrepreneurship in France was these days. A virtual assistant needs to be able to make connections through people because of their interests and their discussions, but also foster such connection when the action between two people is shifting from the initial subject. (I believe I’m doing so right now….)I’ve been to your company website to see what you were referring to, comments or ads (your field if I’ve understood it correctly, but I think you want more than just that) or even learning/teaching could benefit from “virtually enhanced browsing”.I hope my answer seems related to what you said (given the length, I’m pretty sure I got lost in the process lol), I’d be happy to talk more about it if you wish so.

        1. Mark Essel

          Wonderful comment Laurent. You’ve just driven home the point of how complicated the processes of a virtual assistant could be. At the company site our first tools are simple. They parse social status streams, look for meaning real people/place/things, and store them in a database which degrades over time (just like the fading memory you’d prefer). Users can add or delete words at their page: http://imm.victusmedia.com/…The Database at Victus Media:1) is the backbone of an association engine to relate users who share tags, show relevant trending links that are on topsy (also share the entity), and provides a rapid interface to a large follow list (working on this as I type)2) The database is also used by businesses to show relevant ads to you (in our case Amazon had the best utility for us now, Google requires 100k hits per day which is beyond our startup at this point). We share revenue 90% with the host, and 10% with our business. We think that can cover API costs as we scale but we’ll adjust it as makes sense.

          1. Laurent Boncenne

            thank, you ! I like that the database degrades over time, It makes more sense, especially for ads to be culturally sensitive to what a user is doing (and where and at what times he is living).But what if the db itself and the api to access to it would be free ? what if you would be a provider to a technology to either countries/companies and even unique user to store and share the database of “their virtual life”. There are many positive effects of doing so, scaling and handling of such gigantic (and also tiny in a way) db is done by third party partners wich saves on costs on your side. You just need to handle the processing of it.Covering API costs might do the job for the scaling of the business, but wouldn’t be taking a commission over time be a more fitted way to grow ? I believe these are the most complicated part of the business, and it would seem like there is a lot to do on the marketing side to reach a certain number of customers to actually cover the actual costs.Would I pay to get proper ads related to me? most likely not. Would I pay a monthly or yearly sum to have a “friend” who can help me dive into the gigantic web and sort it out how I would want to if I get the insurance that the non-intrusive ads are actually of value to me? yeah I think I would.This is just me though, but having access to the database, or keeping mine for myself is – I think – key.The field of ads doesn’t necessarily need to be in a predefined spot on a web page (heat maps and such). But if you take the web as an entire searching engine, I would tend to believe that user and advertiser would pay to provide links/contents etc. to what might bring in a 80% conversion rate.(number totally made up obviously, assuming that google adsense and the likes aren’t bringing as much)Like I said previously, I think there is a lot to be taken from movies of the last decade (and other science fiction books like asimov’s) where you get a possible user scenario of what a world driven by ads would be. Ads these days are viewed as a pollution somehow, if you can manage to provide a user with a seamless way to share and/or buy from a brand while browsing without feeling like that very brand is selling anything, I think there might be an interesting way to generate revenue in complete transparency.To summarize, wouldn’t it be better to think of a “virtual assistant” as a personal adSense / adWord tool ? one side of it would not only be searching for the web (with characteristics that define the virtually enhanced web as opposed to the web of today based on google page rank and services or networks of today and such) but also categorizing it to bring an ad at the time it feels right and plugging in the adSense / adWord part to bind it all together letting the social network of someone be controlled by himeself only, even if he’s playing a role in the revenue?

          2. Laurent Boncenne

            Ok I think I’ve used the wrong idea here…If you take google, they don’t charge you for using their API, they charge you for somehow being on this technology (adsense and the likes). And they bind ad sense to their google search.In Essence a virtual assistant is your own personal all powerful google, the ad system being one of the backbone of a virtual assistant is I think viable. Provided you take in account that those ads assist you instead of begging for you to click.

          3. Mark Essel

            The virtual assistant would serve as courtesan to the wealth of web information that best suits one’s interests. Semantic two way search is one application of such a tool. It could just as easily 1) find a relevant article2) a needed utility or product3) or a wonderful cafe at a conferenceby providing a potent default suggestion based on usage history.I’m not sure how much of this I can provide in the near term versus later, but I can do something now, even if it’s minor, and so I have (with the help of my brilliant web tech/cofounder Tyler Gillies). Appreciate your feedback Laurent! It was very encouraging.

  36. sigmaalgebra

    Fred,R. Bellman once said, “We often find that a good question is better than a good answer.”Okay: Here you have identified that a problem is “deriving social relationships”.Nice problem identification! Nice “question”!The identification alone is NICE: E.g., MySpace, Facebook, Twitter, Buzz appear to have for most of their importance something or other, poorly identified and articulated, beyond the surface, ‘face value’, functionality.So, you have identified, maybe accurately, that the real purpose is “deriving social relationships”.Or, maybe for one word better, “deriving relationships”, social, business, education, research, arts, political, etc., and maybe two words still slightly more clear, “deriving relationships among people” for various purposes.Now that the problem has been identified, on to a solution!So, (1) What data is, or might be made, available? (2) How can we be a little more clear on just what the desirable results might be? I.e., what is the form of the ‘solution’? I.e., what ‘results’ will ‘solve’ the problem? (3) Given the data, how to manipulate it to get the desired results? (4) Will such a ‘service’ really attract users? (5) What are the more obvious risks? (6) How to make a business out of it?In this, we will be trying to take data and estimate the ‘value’ of a ‘relationship’, between two people or between a person and a group. ‘Estimating a value’, hmm …. Sounds like some mathematics! Uh, it’s not necessarily very elementary mathematics!Or, you had a thread on “cohorts”. That was to ‘cross-tabulate’ on, say, some common point in time for some common event. Or, you were ‘controlling on cohort’. That is a special case of estimating a ‘conditional expectation’ which, quite generally, is the ‘mechanism’ of ‘information’. So if X is the cohort and Y is the information about it, then we want to estimate the expectation of Y given some value of X, say X = x. The notation is E[Y|X = x]. That is a function of x, say, f(x). Or can also write just f(X) = E[Y|X]. Then nicely E[Y] = E[E[Y|X]] (with thanks to John von Neumann). Also, E[Y|X} is the best least squares (likely non-linear) estimate of Y and, geometrically, is also a projection. So, if have X and want Y, then the best can do is f(X) = E[Y|X]. If X = g(Z), then E[Y|Z] is the same distance or closer to Y than E[Y|X] — so, want to exploit all the given information we can. Or, this is the best way to get coveted data Y from available data Z.So, here, thinking of Z instead of X, we want more ‘detail’ about the relationship. Or, if we wish to estimate the value of a relationship separately for each of social, business, education, research, arts, politics, then we can be more accurate than just estimating a relationship without this detail.Yes, can be more accurate with “many” social networks and not just one!”And it may be a mistake to combine them all together into one network.”, yes, that would be using X = g(Z) instead of Z, would be throwing away data, and would give a less good approximation to the desired Y.Gee, you’ve been “speaking prose” already!”One of the more recent lessons for me in the world of social media is the value of the explicit connection. At one time, I thought that automatically building my networks would be useful.”: It would still be better to be “automatic” but only if the “automatic” solution is accurate enough actually to work!”And so it turns out that implicitly deriving social relationships is tricky and potentially dangerous. That doesn’t mean the idea isn’t powerful.” — yup!”It just means you have to roll it out very carefully.” — yup, as in nearly all important things, roll out carefully. Or, as Heifetz said, “There are no small pieces”.Uh, still better …. We better have coffee for that one!For being much more clear on just what X, Y, and Z are, and we can be VERY clear, thank you A. Kolmogorov!Uh, this isn’t ‘computer science’!

  37. awaldstein

    Fred–I’m finding this post 96 comments later.I woke up with the refrain that ‘Social noise is not the same as community value”. I was trying to think through a different problem but it rings true for this discussion as well. http://bit.ly/68LiXR

  38. ShanaC

    Being late as usual though I found reading the comments sort of useful:The question is- what does the explicit and implicit data tell us?And where do you want to draw the line. Clearly you can make networks both ways (we did find this out right now with Buzz, which now I can’t figure out how to turn off)All this being love- there are kinds of implicit networks I would love to just have access to. And in many ways I do. Here is a good example: I have all the comments from the Rhizome streaming into my RSS feed. I’m passively part of that community. You could make a list of all the people who are actively and passively part of that community using similar means, because clearly these people share some sort of similar interest and make a social network out of that implicit data. And I doubt anyone would mind. It’s a voluntary association. As long as someone (or something) keeps updating who is doing what, do I really care if I am on some list with everyone else.Probably not.The issue with Buzz and using email in particular is that it doesn’t define what the relationship is. Or rather, what the association is. The number one thing people are complaining about: I don’t get the use of seeing all the content that xyz person is doing. That doesn’t mean you don’t want to share with them (nor via email). What people want it is to be highly organized by how they are associated with that person. And to socially network based on that thought.It’s pretty much how we function on a day to day basis. We associate on friendly terms with a variety of people for a variety of needs based on a variety of different situations. And that is the way the world works.Personally, I would love to have a social network service that keep track of my implicit relationships, the third degree implicit relationships based around topics and needs so I could better collaborate with those people.. Including yes, in my email. With fun stuff attached, like a wiki! I think that would be so useful…

  39. vruz

    roll out carefully, start as invitation-only beta, make it opt-in, don’t assume all people you emailed in the last 6 years are close friends.google fucked up big time. very visibly for hundreds of millions of people simultaneously. they didn’t even observe the basics. there’s no two ways to see it.this is extremely damaging to their ‘no evil’ creed. steve jobs said that was bullshit, coming from google. and buzz is the proof.

  40. vruz

    roll out carefully, start as invitation-only beta, make it opt-in, don’t assume all people you emailed in the last 6 years are close friends.google fucked up. very visibly for hundreds of millions of people simultaneously. they didn’t even observe the basics. there’s no two ways to see it.

  41. Jan Schultink

    I searched the 117 comments so far on Xobni, but did not find it.Xobni is an outlook add-on to help you search email: http://www.xobni.com/One of the most interesting features of it is that it creates subnetworks: i.e., it groups people together based on correspondence. As a result almost perfectly disjunct social circles are created.The big question is how to use this as a basis for a reliable auto-generated social networking tool.And: if you send a big broadcast happy new year email to your address book, your social networks are not that disjunct anymore…

  42. Andrew J Scott

    Didn’t Facebook get its first traction by implicitly dumping everyone on to the service from the university email directories?..that aside the bold move by Google represents a recognition of just how behind in the social space they are. They have HAD to do something drastic.I feel many underestimate the influence that peoples existing relationships with a brand has over their engagement decision to engage with or “like” a service.Google, people perceive as their “search provider” (and for many, their email provider and maps too). So in the same way I probably would not use a Vodafone social network (however good it is) because I dont WANT Vodafone to be my social network provider, I’m unlikely to use Facebook for my email in the future (I sense it might be coming). In the same way, it is a big leap for me to use Google for my social network or social aggregator – even though I “trust” Google, but in the context of search and maps.Many other factors play a role too of course, but Google has had to make this drastic leap, or face continued loss in the social networking space, and Gmail was it’s only real asset which could catalyse that process.

  43. Guest

    Beacon was an interesting beast. I think the “knowing what your friends purchased” model is useless unless there’s some sort of complementing information (your friend bought it— but would they recommend it?). The implicit nature of the service only compounded the problem. I mean, let’s say Beacon DID work… Now you have a stream full of purchases that mean absolutely nothing when they’re out of context.Implicitly choosing my social network is just as useless. It’s a bunch of noise from people that I don’t necessarily want to follow, or people following me that I’d rather not share certain information with.Implicit, IMO, is almost always bad.

  44. rimalovski

    I use FB, LI, Twitter and my work and personal e-mail accounts all for very different purposes. Would not like to see that all munged together. Agree Google messed up the defaults here, but if you will pardon the pun, it sure got them a lot of buzz no matter how you look at it. May have been part of their calculus?I too like the way LI & FB suggest people you may know based upon overlap in your connections’ social nets. Twitter could/should do this too, but liked LI & FB offer only a suggestion…they shouldn’t automagically link you to those people.

  45. JO

    hey, you guys remember Wave?

  46. Tyler Willis

    looking at the example of 4sq vs latitude, the 4sq data is better because it goes through an element of human curation. That human layer that makes social information work stays in tact when the average publisher knows his or her audience and has a motive/desire to please them.With regard to Buzz – explicit networks where you add your friends directly (and often critically) create a pretty good framework for understanding “who is my audience,” which allows users to share more effectively.

  47. andyswan

    Spot on….It’s the difference between drinking Pappy 23 and Woodford Reserve (it’s saturday, after noon….don’t judge me!!)

  48. Aviah Laor

    FB was born from other (more down to earth motives). To follow B.Bettelheim, Care is not enough 😀 see Myspace, Friendster.

  49. Mark Essel

    You should absolutely have that control Roman, and it’s something Buzz will have to conform to if informed users are going to adopt it. So far I’ve embraced Buzz but I’ve been broadcasting blog comments onto my poor Facebook friends for some time. Filtering is in the hands of the browser.