The Gillmor Gang talks with FriendFeed co-founders Bret Taylor and Paul Buchheit. Recorded Friday, May 30, 2008.
Steve Gillmor: Paul, are you on the call?
Paul Buchheit: Yes.
Bret Taylor: Yes. Paul and Bret from FriendFeed.
Steve: OK. Well, thank you so much for appearing. I don’t know why you would put yourself in the lion’s den here with Harrington hating FriendFeed as much as I do.
But seriously, you’ve got a great service there. When are you going to take over from Twitter?
Bret: Yeah, a lot of people have been writing all these articles about FriendFeed killing service A or service B. But, honestly, from the beginning, we’ve really been about sort of making the services you already use more useful.
You know, when we first started FriendFeed, it was really about aggregating the activity that you do on other sites, and presenting it to your friends in a useful way. And what’s really been neat about it is it really has sort of increased the usage of these services for a lot of users.
So, just as an example, when we first started using FriendFeed, I think, one of us used BlastFM. And because all these BlastFM songs started showing up at our friends’ sites, really the office really adopted that service.
So, really, we sort of think of FriendFeed as almost the opposite of that. You know it’s not so much about displacing services, but really about making all the services you already use more [audio cutout].
You know your [audio cutout]
Steve: Something’s breaking up there. I don’t know what that is.
Bret: Yes, it’s a digital distortion. I’m not sure what’s going on.
Steve: Well, I didn’t hear the last thing you said, but it seems like we’ve learned from Twitter’s experience that what the founders of a service think is what they’re doing, and what actually happens, are two separate things.
Bret: That’s absolutely true. I mean not always. Certainly users define what a service is and how it should be used. And we’re not going to deny that’s the case.
But that said, I think, that in practice most people use FriendFeed as their content discovery pool. So, it’s sort of viewing the web through the eyes of people you know. The majority of the content on FriendFeed are links to articles, whether they’re techfriend’s blog posts, or a funny video on YouTube, or a funny picture on Flickr.
And I think, that most people, and there’s a lot of different uses of Twitter, but it’s largely a messaging services. So, I would say like most people who use FriendFeed also use Twitter and probably use them in somewhat different ways.
Steve: What you guys are doing in terms of being an aggregator for a bunch of different services, that’s terrific. There are some people, like myself, who don’t use a lot of services.
And so from that perspective, the issue of conversations that are being siloed inside of your environment is something that I think, over time is going to become limiting in the same way that when people like Robert Scoble talk about a FaceBook/Microsoft strategy, which I don’t think exists.
But, let’s say that he’s right. Where they are all going to go behind some firewall along with Yahoo Surge and create this alternate web, where they can control price points. You know I think, that that’s a fatally flawed strategy that everybody who’s on the web realizes won’t work.
Paul: No, I agree with you completely about that. And we’re not trying to silo anything here. In fact, we’ve made every effort to try to be open with everything that we are doing. So, we do have a full API that exposes basically everything that’s in our system, all of the data.
So, if someone wants to write a thirdparty application that pulls out the comments, or puts comments into it, or really whatever they want to do, we try to make that as easy as possible.
And to address your other points about people only using a handful of services, I think, you’re again correct about that. I think, most people probably use zero or one service, or something like that.
And part of what makes FriendFeed so powerful, though, for myself and a lot of people is that we all use different services. And so, for example, maybe I use Google Picasa Photos for my photos, and my friend uses Flickr, and another friend uses SmugMug. With FriendFeed I can see all of those different friends, even though we’re all using separate services.
Steve: Drilling down on the notion that you have to have, that somebody has to come up with an alternative service to Twitter, in order for there to be a true marketplace. Right now we’re basically held hostage by Twitter’s inability to be able to keep up and running.
In that context, it seems to me that it boils down to about four different capabilities that you have to provide in order to be able to provide an alternative to Twitter. Not that you necessarily want to, or that’s your business model.
But, the ability to be able to provide real time services is currently distributed via XMPP when it’s working on Twitter. Is that something that’s in your roadmap?
Paul: To make FriendFeed available via XMPP?
Paul: You know that’s something we’ve talked about. We’re not working on it at the moment. But, it’s not a bad idea. I think, it’s a good idea. It isn’t what we’re doing at the moment. But, I wouldn’t be surprised if we do it in the future.
And in fact, one thing that you can do using the APIs is expose it. Again, don’t I recall. There’s so many API apps now, I think, someone may have already implemented that.
Steve: All right. So, let’s say you then create essentially the ability to be able to export from an XMPP feed, and be able to import from somebody like Twitter, and then be able to continue to pass that along. That would create a certain velocity for your application, which would be significant.
Steve: That would be my suggestion.
Paul: To open up an XMPP panel?
Steve: Yes, absolutely.
Paul: Yes. It’s very much in line with our… What we want to do, really, is both be able to pull in data from anywhere, but also to make FriendFeed available anywhere. So, we already have a FaceBook app that brings FriendFeed into FaceBook. We have obviously the RSS feed. We have an iGoogle Gadgets for the Google home page.
We send out daily email summaries if you want to get your FriendFeed in email. And so, adding XMPP, I think, would fit in with that very logically.
Steve: OK. So, then again, what I’m operating under is regardless of what the original intentions of Twitter were, it has achieved a critical mass, which allows for a certain ontology, might be the correct word. A certain ability to be able to express your circle of influence and the people that you’re interested in following or tracking, in such a way that you can maintain a balance between serendipity or discovery and the ability to not just sit in front of the machine all day.
OK. So, the second part of that then would be some sort of follow architecture, which would mirror, in fact, clone the follow architecture of Twitter. Now, it seems to me that you’ve already got most of that at least potentially available.
Paul: The ability to follow other people?
Steve: To be able to follow in a coherent way. To be able to model the same kind of cloud, the follow cloud, that Twitter provides.
Paul: OK. I think, so, if I’m understanding you correctly. I mean, we have a very similar model to Twitter in terms of you subscribe to people. It isn’t necessarily a twoway relationship the way it is on a typical social network.
Steve: Follow is not a twoway model. Well, it’s a oneway model…
Paul: That’s exactly right. FriendFeed and Twitter both have the oneway follow model vs. the traditional Facebook, MySpace two way model.
Steve: OK. So, the capability of being able to follow in one direction is available. The ability to block is not available.
Steve: OK. Why is that?
Paul: We just haven’t gotten to it yet. But, I’m sure we will add that. We’re really, in many cases, just adding features kind of as users request them. So, that’s actually something I was just thinking about recently to see what the right way to implement that feature is for FriendFeed, since the interaction is a little bit more complex.
For example, I may not want these people to be able to comment on my entries if I blocked them and so forth.
Steve: Right. Well, you’re again, I don’t mean to be… well, I do mean to be very specific about what I am asking you about. What I’ve tried to do is to model an exact clone of Twitter, including the services that it currently provides.
What you’ve got here is an aggregator, which does a lot of things for a lot of different services, en masse and to some extent, individually. If you were to take the model of Twitter, break it down into its components and then provide a view that essentially was identical to Twitter that would be a very powerful move on your part.
So, I’m talking in that context.
Paul: I see. Possibly. I’m going to say, though, I really think that clones of other services are generally never as good or as successful. We really aren’t interested in cloning Twitter or any of the other services. We’re really…
Steve: I’ll be glad to argue about that point. I still want to ask the questions that I want to ask. And once we get through the feature list, then we can back off and perhaps have a dialog about why that might be intelligent to do, not only for you, but for your customers.
So, once you’ve got block enabled, which you suggest that you’re interested in perhaps implementing, then the track function. Any plans for track?
Paul: No, not at the moment, but we do have search, which I think, may be kind of an equivalent functionality in some ways. You can search over the entire public feed or over just your friends and also limit it to particular services, if you like.
Steve: Right. The problem with search, as it is currently implemented on your site, the difference is between a search methodology and track methodology; what is the degree of latency. And the reason I’ve attacked FriendFeed, from a humorous perspective, about the latency is that it’s not because I really believe that it takes admins for round trips, but mostly because there is a quality and a kind of communication possible on an interactive basis in Twitter when track is running, which is not available on any other server.
Paul: I see.
Steve: OK. In other words, I say something. Anybody who is in the user base says something, if they attach the appropriate signal, so that I pick it up via track, I immediately become aware of it. This creates a much larger surface area for communications on the network than any other service has at the moment.
It also, I believe, is responsible in large part for their scaling problem. I remain unconvinced that they can actually support that, because it requires that their entire user base be up and running simultaneously. That includes people like Scoble. And I don’t think that they can handle it.
So, it’s a pretty big bar, but one of the reasons that I think, Scoble has been talking about you guys and investing in your service, other than the fact that he likes the ability to aggregate these conversations in a silo, which I don’t think is an asset, but none the less, it’s a solvable problem to be able to unsilo them, I think.
But, the main reason, I think, is… or one of the significant reasons is because you guys come from Google and from an experience developing applications like Gmail that have a tremendous amount of scalability in precisely the area that we are talking about here.
Steve: All right. So, track, is it going to happen do you think?
Paul: You know what, I honestly haven’t given the feature a lot of thought. But, given what you said, I think, maybe we should talk about it and figure out what that means in the context of FriendFeed. As I said, we really haven’t actually talked about it much, but perhaps we should.
Steve: All right. So, Scoble will be on the call in a little bit and he may have a couple of other questions. We sort of mapped out part of this conversation at the Google IO conference. I may have forgotten some stuff. And I don’t want to hold off Hugh for much longer to ask his questions.
I do want to at this point, after we hear from Hugh and anybody else who wants to talk, I want to come back and start talking about your conversations, your comments and about the silo aspect of it.
Robert Anderson: Hey, Steve, I just want to say, Hi, this is Robert Anderson. I’m a little bit late.
Steve: OK, Hugh, go ahead.
Hugh: Yeah, I kind of notice a trend. I’ve been using Web 2.0 for like about seven years now, with my blog and now. The way people use Web 2.0 we have all these different services we use. Let’s say, Twitter and Seismic and FriendFeed. I’ve just started recently using FriendFeed. I am impressed with it.
It seems to me that when you have a social network, every social network has a nucleus. In other words, you got one or two platforms occupying 80% of your online usage. For me, it’s my blog what puts my cartoons and my longer thoughts. Socially, just day to day kind of random interactions with people I use Twitter.
Sadly, Twitter has been a bit not working very well recently, which saddens me because I’m getting quite dependent on it, business wise. I’m looking at FriendFeed, saying, OK this works really well. These guys from Google and they have done some really good stuff.
And Paul, when you say you’re just using FriendFeed just exists as an aggregator, I have 6400 followers in my Twitter and they are important to me. I want to be able to talk to them and have them talk to me on FriendFeed, whether they are signed up with FriendFeed or not.
I want to be able to talk to my friends on De.li.cious, whether they’re on FriendFeed or not. Dare I say it, I want people to be able to talk to me on my blog comments and talk to them back, whether they’re on FriendFeed or not.
This whole thing where you have to become a member Now, I can understand and it would be great if everyone would become a member, but what if they don’t, then all of a sudden it’s less useful to me.
I’m not saying your sky is the wrong one, I’m just saying, wouldn’t it be great if I had this one place I could go and talk to everybody that could talk to me and I could talk to them, get rid of all the assholes I don’t like and just talk to people I care about, whether or not, regardless of what particular cornucopia of Web 2.0 tools they are using at the moment. That to me would sound more powerful.
Paul: Yeah, that seems very reasonable and I think, it’s really in line with the kinds of things we are trying to build, but we’re just taking it one step at a time. But, that sounds like a great vision.
Hugh: Yeah, and it’s… I’m not saying that you are doing anything wrong, I’m just saying every Web 2.0 user has a nucleus and I suppose the question is: are you a nucleus page or are you just like a random out there aggregator page, because being a nucleus to me is the money shot. If you look at Facebook, that’s a nucleus shot for millions and millions of people.
Hugh: Especially, the people who don’t have time to learn about all the wonderful world of Web 2.0. They just want to go to one place where they can do it all.
Steve: Right. Right.
Hugh: And I’m just saying, is there a nucleus, you know, to me the money play is you figure out how to make FriendFeed and there is essential nucleus for people, like Facebook. Or like Twitter. And I’m hoping you do that, but I’m…
Bret: This is Bret again. Certainly part of the FriendFeed has been sort of a nucleus for content discovery and discussion for a lot of our users. That’s where a lot of our growth is. But, at the same time, one of the reasons that I created FriendFeed was that what I defined as my nucleus of online activity was different from people that I knew.
I worked at Google for a lot of years, so I drank the Google Coolaid and I used Picasso web albums instead of Flickr and I used Gmail instead of Yahoo email, and whatever. But, that doesn’t mean that all of my friends do that.
What happened in the past five years was the proliferation of syndication format and APIs is that that data isn’t necessarily siloed in all those services. Likewise, I don’t think that everyone should have to use the same nuclei to participate and syndicate with each other.
I don’t think FriendFeed has really accomplished the vision that you said, which is a really good vision. At the same time, I really would like the world where if I want to use Flickr for my photos, and I want to use Pandora for my music, that shouldn’t necessarily depend on everyone that I know using those services. The thing about FriendFeed is that we hope that you can have that diverse array of services and they can all sort of interact with each socially.
Hugh: That’s good. I suppose there are like two kinds of Web 2.0 users. You got the Gillmor model and the Scoble model. Scoble views hundreds and hundreds and hundreds of different things all the time. And then me and Steve Gillmor, where we just like one or two or three things.
Bret: And Steve Gillmor, probably the common case. But, what’s interesting about that is that I think, that the union of all of your friends’ one or two services is a really diverse set of information and a really diverse array of services. I think, that’s a natural tendency. I think, in general, people, a small group of developers focusing on one thing and doing that one thing really well almost always results in a better product than someone trying to do everything.
And I really think that that’s the natural tendency of the Internet, and I’m hoping that FriendFeed really like plays off that trend as opposed to playing off the, you know, we already have the distribution channel, so we’re going to make copycat products of everyone of them.
Steve: So, who the hell was that?
Bret: That’s the Union Pacific Railroad.
Steve: Who’s that? Robert?
Feldman: Feldman, Feldman, Feldman.
Steve: Hey Loring, how are you?
Feldman: I’m well. I’m well.
Steve: OK. So, what I was going to ask. I was going to point out, Bret, that basically that’s the second time that you guys have suggested that what we are talking about here is a copycat service. I think, you are missing the point. So, I’m going to bookmark that and then let Feldman and Scoble ask some questions and then I’m going to get back to that. So, go ahead.
Feldman: So, I’m jumping in a little late, but I’m totally into FriendFeed. I agree with what I just heard. I really like it. I like it a lot. I think, with all the Twitter comparisons and all that, not that I have… My one opinion would be like if you just separate the conversations from the feed somehow in a view. That’s it. And then, Twitter at this point, is just laughable to me. I’m sorry.
I mean, yeah, it’s cool, but the fact that they can’t keep that up. I mean, it’s laughable at this point. And truthfully I’ve always been at the FriendFeed. Period. That’s what I have to say.
It would be cool to just separate the conversations. With all that screaming about the comments and all that, I don’t care at all where the comments are, truthfully That whole post that Fred did about; comments are payment for bloggers. I mean, I just thought that was idiotic. You comment, you put it out there. You don’t have to track every single piece of data about yourself. That’s what I have to say.
Bret: Shall I respond?
Hugh: I have a question. I am using FriendFeed. I find it very useful, but I also find it difficult to navigate. There’s a lot of data in there and what are you doing on the user interface side to make it more usable?
Bret: Yeah, that’s a really good question. I would say probably our biggest complaint from users is just that sense of being overloaded by the amount of information on the site. So, we’re working on a number of things. We’re just sort of in the testing phase, so I’m not sure which will pan out.
Two of the trends are summarizing the best stuff, you know, a period of time so that you don’t need to sift through every single piece of information to find the ten or so best items, and then also, combining things that are duplicates or related. And I think, those two things will probably do the most to at least address most of our more vocal, the more obvious problems.
Hugh: You know what would be really cool is you could just search aggregate like pictures, video, conversations.
Bret: You could do that with search I think. But, what’s interesting to me that the aggregation idea, which is combining things, which are duplicate or related, because you could start to see clustering of means that are going on at the same time and related.
Hugh: Yeah, that would be huge if you could join that all have the same title, because a lot of people are using Google Reader to share items. If something is really hot, like the Puppet Show, then all of a sudden, Loren’s title repeated 50 times and I see the same thing 50 different times. Some of them have comments and some of them have likes and some of them, a lot of them.
Bret: I think, that actually your mention of the comments and likes is actually the user interface contrast, it certainly been something that a lot of people have talked about. But, the fragmentation of conversations on FriendFeed is actually one of its most liked features by a lot of users, primarily because if someone writes say, for example, if you’re a big Obama supporter and you read an article that’s pro McCain, on FriendFeed there will be a little conversation among people who know each other discussing it. They don’t necessarily want to participate in the public forum because they might be attacked or they might not have many people on that public forum who agree with them.
But, there might be like 20 different conversations about that blog post among 20 different social groups, all with that sort of their particular perspective and their particular inside jokes, or particular social dynamic.
And so, for a given article I might have one social group might be my family and one might be sort of my more techy friends. And keeping those conversations separate is actually a feature to me. So, the challenge for us in the user interface perspective is how we can eliminate the noise and duplication, but at the same time not merge my coworkers set of comments with my mom’s set of comments, because those social groups are a fairly distinct.
Steve: The problem is what you have just defined what many people see as a problem, as a feature.
Bret: That’s correct. I..
Steve: Let’s drill down on that a second. One of the strategies that is employed in managing information is to allow the user to be [inaudible] you don’t allow the user to [inaudible]. There are no tools as far as I can discover in the limited amount of time that I have been in FriendFeed, because I’ve been embargoing it as attempts to interest me in it.
But, one of the biggest problems I see is that there are no tools to be able to aggregate conversations across different identities. So, you have the exact opposite of the problem that Twitter has, which is that with Twitter it’s all about the identity. And the conversation, literally, most people spend all their time trying to figure out who somebody is talking to, because they can’t understand because of the @ reply syntax. They can’t understand why they’re talking to or why they’re not talking to one person in particular.
Bret: I don’t think that… I’m not sure if I completely understand your comments. And you can let me know if I misunderstand then.
Bret: There is right now, if you have a blog and you have a FriendFeed account, there’s a WordPlus plugin that will pull in all of the comments and likes on your FriendFeed item corresponding to your blog’s post. And pull them into your blog.
Steve: Right. You’ve got that right.
Bret: So if you bring up the O’Reilly’s Inquisitor blog, it will be at the bottom of his post, for example.
Steve: I understand that. But, then what you have is the duplication problem. So, you have to employ resources to try and filter the stuff that shouldn’t be there in the first place, because you don’t have the capability of allowing the user to be able to create the conversations across domains.
I mean, I understand that you think it’s a feature. I’m not sure it is.
Bret: I acknowledge the fact that there are positive and negative aspects to it. I do think, though, that in practice… As we’re growing, and we’re small enough now I don’t think we’ve encountered most of these social problems we’ll encounter on our site.
But, if you look at typical public forums, whether it’s Newsnet, or Flash.Comments, or Digg Comments, or even like the probably lowest common denominator, which is YouTube Comments. When you have a public forum where anyone can participate and identities are typically sort of virtual in those environments, they tend to devolve to the lowest common denominator.
And you can look at any YouTube video to see an example of this. I mean, most of them are like derogatory or antiSemitic or something similar or a combination of those. And then, there’s some public forums that sort of evolve to kind of work over top. But, they sort of are like radio talk shows that take in callers.
So, people call who call into those talk shows tend to be the people who sort of like to hear themselves talk. But, it doesn’t necessarily encourage people, so exhibitionists tend to participate in the conversation.
Whereas, I think, that more traditional social networks the conversations tend to be extremely personal, like a wall post on FaceBook for example. It’s two people who know each other and they’re communicating directly with each other.
You know, FriendFeed is kind of a hybrid of those models. And we are really sensitive to the fact that unifying comments is something that is a really great feature for publishers. But, for participants in the conversation, you don’t necessarily want, say, like the author of a book to hear all of your comments on it.
Your comments might be more directed toward people that you know, your friends and family. And so I think, it’s a difficult challenge to balance the feeds of public users versus the private…
Steve: That’s a fair response. But, what it doesn’t seem to incorporate… I mean, the idea that you’re creating an environment where people can feel comfortable enough to be able to engage in less than broadcast conversations, that’s a good thing.
The problem is that when you scribe swarming characteristics to a publishing model, you’re not thinking about the user. The user wants some sort of aggregation technology that is going to help them find things that are interesting to them.
Bret: Yes. That’s absolutely true.
Steve: You’re starting to do this in many of your extensions to… I mean, rooms is a… I don’t think is a particularly good implementation. But, it’s certainly toward this issue.
Lorne: Can I cut in, Steve? Going back to the original point of privacy, you know personal and business kind of comments. At some point the user has to take responsibility. You’re not going to comment, as the brought up, direct persontoperson FaceBook kind of message isn’t going necessarily to be the same kind of messages that you leave on a blog.
Steve: Lorne, I disagree. This is the same mismatch problem that Twitter has gotten into, where they’re suddenly outraged and a lot of users are outraged by the notion that people, like myself, will conduct conversations in the open on an instantaneous basis.
What we are talking about here is what I was originally talking about, which is the XMPP real time capability. If you introduce that to the product, and I certainly hope you do, then you’re going to have a different class of issues that relate to the notion of what you call comments.
Bret: Yeah, that’s true.
Steve: It’s going to be a conversation. A conversation in this media age and I’m not talking about publishers here, I’m talking about people who need to survive in this information madness that we’re going through that they’re going to find that direct pointtopoint and onetomany conversations, of the kind that you and Twitter have so far characterized either as a broadcast issue or a personal spam issue, because the verbosity of the backandforth can really piss people off.
Those things are going to become the coin of the realm. That’s what’s happened with Twitter. And in my opinion, that’s why they haven’t been able to sustain the architecture that was designed for something else.
You guys are in a position if you start to implement these things, if you provide a service that can come alongside and draft off of what Twitter’s doing, you have an opportunity to be able to do your architecture on the fly as opposed to doing it in panic mode, which is where Twitter is right now.
Lorne: Steve, I have a couple of things that I would love to see that goes straight to this. One is I need to be able to pull stuff out of the FriendFeed database, using some sort of simple query language.
And here’s the query language I’d like to use. I want to see everyone’s post that has two or more likes and four or more comments on it. And replace your numbers with my numbers.
So, maybe, some people might want to see one comment, and some people might want to see 10 comments, because they’re only really looking for the really noisy posts. Because those are usually the ones that get on Techmean are the ones with 100 comments on them.
But, I need to be able to pull things out of the database in that way, and then create a URL that I post around.
And say: “Hey, check this out. Here’s what…”
Steve: They’re building that already. That’s a service that they’re already constructing. Before you got on the call, they were talking about this as a response to my request for a track capability.
That, however, is not about real time conversation. It doesn’t mean that it’s not important. Obviously, it’s important. They’re already doing it.
Lorne: That’s one of the four. The second thing is I need XMPP capability. I need to be able to not just put stuff into the FriendFeed engine from Google Talk or from some IM client like we used to use with Twitter. But, I need to be able to watch what’s going through the FriendFeed engine through on an IM client. And that would give you real time capability.
Steve: And they’re talking about that. OK, what else?
Lorne: I need SMS compatibility, so that I can use… First of all, that would limit the post size to 140 characters. And have some way to read these things on SMS. And also put things into the FreindFeed database from SMS.
Steve: I just want to bring up something in the comments by Keshep. It’s says: “Mr. Scoble seems to like to take services and completely warp them to fit what he wants to do.” Absolutely.
So, do we all. Everybody on this call wants this service to work as best as possible for their particular agenda. What’s wrong with that?
Bret: Exactly. And it’s our responsibility as the authors of the application to make it work for all of you. And I keep telling you that we’ll try.
Hugh: Well, hold on. That’s not entirely true. I mean you don’t want to do what anybody wants to do with it, right? It’s not built for everything. You can’t make everyone happy. And if you do that, your product will go into the toilet.
Steve: If the users decide that they want to do something, and it doesn’t work for them effectively or efficiently, it’s the service’s job to either figure out how to do it, or tell them to shove off.
Mark: Can I say something?
Mark: All right. So, I want to get back to when Steve insinuated about when we were talking about cloning Twitter, versus the FriendFeed guys going about it a different approach, a different set of issues. And I think, over the past 10, 20 minutes we have elucidated a lot of different kind of issues.
Steve seems to be focused on conversations, Scoble is in this kind of mass aggregator mode. And Hugh has a very simple notion of community.
And I suggest that to create the proper architecture and I’m saying this to the FriendFeed guys that I believe that all these issues have to be taken into account. Steve brought up, for instance, the filtering of information and also a persistent ID that works across lots of different services.
And so, let me just take a wild guess that the FriendFeed guys already have their money. They’re comfortable. They don’t have to freak out every morning figuring out if they can pay the bills. So, they have the luxury of being able to architect something that could benefit everybody.
It’s not just about making money, but it’s an architecture that can enable an open mesh environment. And so, the discussion that we have been talking about, each of the people on this call have their own agenda. We all want to try to get the features we want. And FriendFeed does have the ability to take the perspective on what is the right kind of open mesh architecture that can interconnect together and to create the redundancy.
What that means is that it won’t just be FriendFeed. No matter what you come up with. I don’t care how much money you got and how important you are, the right kind of architecture has to have redundancy in it so that if one node goes down, the rest stay up.
Steve: Right. Now. We got the question, Mark. I’d like to hear the answer. And then, I want to attempt to remove the pigeonholing that you put what I’m talking about into. So, go ahead.
Bret: I’m not sure, maybe that was sort of an… I understood the concerns about scalability but I’m not sure I completely understand what the question is. I’d be happy to answer that but maybe you could just restate it.
Mark: OK. Instead of it just being a FriendFeed only solution, that what you come up with is an open architecture that others could support and participate in.
Bret: Yeah, so that conceptually sounds great. I think, that architecturally that the devil is always in the details. And so, I think, for the foreseeable future, we’re going to be focused on making sure our service is scalable and the data is open enough and the publishing mechanisms are open enough that there is implicit user interface replication and there is lots of ability to publish to FriendFeed in a really distributed way, as well.
That said, I’m not sure… That’s a very broad statement. I’m not sure if I completely understand specifically what you have in mind, but we’re very interested in being as open as possible. Our service is built on these open syndication formats that really power the data going into FriendFeed, so we’re very committed in concept to that, at the very least.
Mark: So, let me just summarize quickly by saying, if there were open conversation formats, and open ID formats, and open media formats, and open commenting formats, then that would achieve the goal.
Bret: Right. And then we would have to boil the ocean about 12 times in a row.
Mark: The Google. The Google. It’s one step removed from Google, so they can do that.
Bret: Well, Google is not the only solution that’s open. So, what you are talking about is some sort of aggregation of open technologies. My thesis would be that the marketplace will decide that and probably very quickly.
Mark: OK. Then, get out of your pigeonhole then there, Steve. I’m sorry for your pigeonhole.
Steve: The pigeonhole is: what I was talking about in terms of real time conversations is not… I mean, first of all it’s the only intellectual property that Twitter holds outright at the moment.
And therefore, given the fact that they have a strangle hold on the user base at the moment, the only issue is whether or not there is going to be some pragmatic resolution of the one throat choke problem that we are under right now, which is, not only do they have all the users, but it is broken.
So, how do we get from here to there is the sole thing that I am interested in. The vehicle for that is to provide incentive for a service to be able to come along side of Twitter and provide an identical service base.
Now, just to put some context around this, as Mark has talked about quite a lot when he talks about the notion of Open Mesh, there’s a player that could definitely come along side of Twitter in the reasonably near future and that’s Microsoft.
Mark: I think, Facebook will beat them.
Steve: I think, that’s possible. There are several players. I don’t see any difference between Microsoft and Facebook. I think, they are the same company. But, the other company that I see being able to come along side at some point in the reasonably near future is Google.
What we have seen with Friend Connect and many of the other technologies that they showed at Google IO over the past two days, you can begin to see the speed with which they have developed Android for example, is stunning. And so there is every reason to believe that for getting the JiKou purchase or any of that kind of stuff….
It will be interesting to talk to these folks here from FriendFeed who are involved in the original build out of Gmail, which as we all recall, didn’t work nine nines on day one. Not on day 100. It was something that was built slowly and the problem that Twitter…
Hugh: For the record, Google Map, which is my product works nine nines.
Bret: I’ve been a faithful Gmail user since the first day that it was available and I’ve been happy about it ever since. So, I don’t consider that to be any way a critique of the product.
I’m just talking about how do you get a solution online that provides, what I consider to be, the most viral and disruptive aspects of this technology?
And that’s what Twitter is. So, when we talk about a Twitter clone, I’m simply talking about, as one part of the feature set that FriendFeed is developing. If they were to either by themselves or in concert with other companies, including Twitter, for that matter, to be able to provide services, which have a fail over capability, then the marketplace would be created.
In that marketplace, the kind of standards that you are talking about, Mark, would be quickly manifested.
Mark: Just so we are clear. When you say fail over, that’s what I mean by redundancy, having more than one.
Bret: I know that’s what you mean. That’s my point. That when you talk…
[audio breaks] [crosstalk]
Bret: It sounds like you are describing like a federated system.
Steve: That may be what Mark saying… I’m talking about a real world business economics framework. One that exists because there are two companies that provide the same kinds of services and people then make the choice based on the investment.
Bret: I don’t think that technology is typically the reason like social services succeed. It’s a necessary part of the service, but certainly not sufficient from… I think, there are a lot of…
Steve: You’re misunderstanding me. Right now the business of Twitter is nonexistent. I mean, everybody talks about the business model. Well, in my opinion, their business model is extremely successful, right now.
They have got a group of addicted users who, in the face of almost complete failure, are sitting here waiting like rabbits, like rodents punching a button to get another pellet waiting for our track to come back. They are completely successful on a business level.
But, the issue is can any other service provide the same functionality? Because Robert Scoble has made an investment in FriendFeed. Now, he has very many, a lot of reasons to do this. He’s stress tested Twitter to the point of failure. He’s stress tested Facebook to the point of 5000 users at which point they stopped letting any more in.
He is making an investment with his gestures into a system, because he feels that they are going to be left standing at the end. And I think, that’s an intelligent investment on his part. I’m also making an investment in Twitter, because I consider that they will be left standing at the end of this cycle.
I think, that both of you have an opportunity if you both come up to the same level of functionality for those people that want it. You essentially isolate the rest of the marketplace and you bulletproof yourself from attack by everybody, except those two large clouds that we mentioned before, and they have their own problems in terms of social contracts with users, which is they have to retain credibility and they have to be… You know, it’s not even do no evil anymore, they have to do good. They have to convince users that they’re not going to misuse the social contracts that users supply to that.
You guys have got a tremendous social contract with your users. Twitter obviously has a tremendous contract with their users. You guys are in the position to be able to afford a framework that people can use. And what’s emerged here and we can all look the other way but in fact what’s emerged here is something that’s going to unify IM and email and broadcast publishing technologies in one client. That’s what you guys are doing here.
So, are we going to get some standards? Probably when there’s what he calls redundancy. So, I think, that if you guys ignore this investment, it’s going to be seriously nonstrategic for you, because there are going to be people like me who will look at you and say, “I’m not investing in you. If I have to invest in one place, I’m going to invest somewhere else.”
If I can invest simultaneously in you and somewhere else, that’s even better. That’s what you need to be able to provide. And right now, you’ve got a lot of the capabilities but not the one that is the most difficult to clone.
Mark: So are you describing essentially just better interoperation with Twitter? Is that, am I understanding that correctly?
Steve: No, I’m talking about this real time conversational capability based on the discovery of the entire network of users. That is a disruptive technology which right now Twitter owns. Now, you can come alongside, if you encourage investment in your cloud. I mean, this is what Dave Winer has been talking about since day one with Pounce and all these other services including yours which was go over there, invest. Start flowing your data through those clouds and then if something else fails then you just drop away and you’re OK and you’re in this new system.
You have a certain critical mass, but there are these problems with the siloing of conversations on your service that I think, tend to make people not trust the investments in the service, because it’s not equally distributed across new players.
You’ve got a lot of new players who are coming in who are able to be very visible in a smaller pond. But, what you don’t have is the critical mass that Twitter has already invested in and desperately kept up and running.
So, if you come alongside them, in other words you open up an XMPP gateway as Robert’s suggesting and as I am suggesting, that allows people to do the same things that are being done with Twitter. Namely, pump it in through the Jabber portal, pump it into Gchat and Gtalk, which means that it’s automatically incorporated into the Google search in Gmail. Then you’ve got an archive.
Then, if you’re away from the machine, all the sudden all those details that you’ve missed are easily searchable and you can find them using somebody else’s server. That’s redundancy in a powerful way.
Steve: All right. So, that’s a check off as far as I’m concerned.
Mark: OK. Well, we’ll have to figure that out.
Steve: Once you get the track, then you get to track filtering. And track filtering opens up verticals, it opens up personalization, it opens up the whole ball game. So, you’ve got to be in it to win it.
Mark: Steve, is it OK if I have a different opinion?
Steve: Yeah, of course.
Mark: Because I think, you’re making the assumption that FriendFeed and all of these sites have to be about interactive conversations.
Steve: No. Again, you’re pigeon holing me about something. I said, do whatever else you want to do, but if you don’t do this, I’m not going to invest in you.
Mark: OK, but that’s not the world..
Steve: But, there are going to be people who are going to be outside the service that doesn’t respect something which is absolutely transcendent, which is what that service is. I mean, you don’t use it that way, I get it.
Everybody who I’ve ever talked to about this on the shows or in writing who’s figured it out, gone through the pain of trying to make it work. And then, they get there, they’re all sitting here like I am now, completely addicted to something that we can even get. It’s like all the heroin has been taken off the streets. [laughter]
Mark: It’s true.
Steve: Right. So, I’m not saying that what FriendFeed is doing in so many other ways isn’t valuable. Obviously it has value or we wouldn’t even be talking to them. But, what they’re not…
Mark: You think they’re the shortest path to get there to create the redundancy? They’re the fastest way to get there.
Steve: That’s exactly right. That’s exactly..
Mark: Well, that I do agree with. I do agree with that. But I think, you’re putting a lot on these guys. That’s not necessarily why they wake up in the morning time, to fill in the redundancy and solve the Twitter problem.
Steve: Maybe, they will start waking up in the morning and thinking about that a little bit longer. I kind of hear that in the tone from Paul, I think, it was.
Mark: Yeah, OK. We’ll have to think about these features. But, I guess I am optimistic that, I think, Twitter is going to resolve their issues and I know it’s a very sort of difficult time for the Twitter addicts right now. But, I do think that they’re going to manage to get that figured out.
And that’s just my opinion. But, I think, that once they kind of get that resolved, some of that urge to have essentially the second supplier is maybe going to…
Steve: No, I don’t; you’re misunderstanding me. It’s not about a second supplier. One of the great powers that Twitter has enabled for themselves is their third party application support.
Steve: As have you. I mean, you’ve got the alert thing for example.
Bret: That’s a big deal. You’ve got a small engineering staff and yet you’ve got somebody else out there working for you for free. That’s a good deal.
Bret: You know, API environment, the network is producing. It’s accelerating the velocity of these kinds of applications. This is in your best interest just as much as it’s in Twitter’s best interest. Twitter does not have anything that it can go to the bank with except its relationship with users.
And so do you. You’re in the same position. If people say, “Oh Twitter’s bullshit, I can’t stand it, it’s still going down, yada yada.” Then, they’re going to come to you and start complaining about the same thing.
Mark: OK, so can I clarify so I don’t get pigeon holed?
Mark: All right. So, now, yes interactive conversations are important to you. But, the key essence of the business model you’re talking about and the infrastructure that Twitter provides is what I’m talking about, right?
So, what I’m saying to the FriendFeed guys is, you’re smart enough to understand the technical challenge of solving everyone’s problems. OK? And just like a redundant Twitter, you would allow people, route around it and not be caught up in that particular infrastructure.
So too, do we do the same thing for ID and for leading with sharing, and for aggregating activities. And if you stand back and listen to what Steve is saying, the kind of redundancy and what I’m asking for, would not only solve the interactive conversation problem, but would also aggregate ID profiles, would allow us to share media, would allow us to aggregate.
Steve: Absolutely right. We’re in violent agreement here. So…
Chris: Could I just chime in here?
Steve: Go ahead.
Chris: Business courtesy, huh? I just wanted to say I feel really sorry for the FriendFeed guys because you and Mark, sort of blessing them the ability to solve a whole lot of problems for all scenarios. And they are just really trying to a really great aggregate.
Steve: Well, you know, Chris… Well, let them answer that. I mean, is that all you are trying to build was a really great aggregator?
Bret: Well, I think, aggregation is more of a mechanism than a product, so we think of it more in terms of just sharing interesting content and discoveries with your friends. So, the aggregation enables and improves that. For example, if you’re… if you posted a video on YouTube, we bring it in automatically, so that you don’t have to do it manually.
But, the core activity really is about sharing and discussing content for the most part, which I think, is really one of the distinctions from Twitter, which, I think, is more of the pure conversation that I think, is obviously very important to use.
Steve: I think, that’s bullshit. I think, you’re characterizing what you want to do with your service, as opposed to what the users want to do with your service. And what your service provides is excellent. But, it doesn’t mean that there aren’t users out there that take a look at what you have built and say to themselves, ‘I’m going to use this for something that they may not think that is important to me, but I think, it’s important to me.’
And the history of the Internet shows that people do this all the time. Now, I’ve been using… I was using Gmail to write documents because it had a Spell Check function, the second the Spell Check function occurred, long before Google Docs became obvious.
Bret: No, you’re absolutely right. People will use these services for whatever they desire. I think, obviously if we’re smart, we need to pay attention to that and to continue to refine our product to meet the needs of our users. But, of course, there is always a balancing act there, as well.
Just that the many different users all have their own needs and opinions. So, one of the hardest things in product design is just balancing all of those different needs and make sure you haven’t harmed the experience for one group of users in order to meet the needs of another.
Steve: Right, but you’re presenting a false choice here. We all understand that you have limited resources. Who doesn’t? You have less limited resources than many of the other players. There may be people listening to this call who are building FriendFeed clones that are going to be much more aggressive about incorporating the kinds of features I’m talking about.
I wouldn’t want to bet against those people. They’re your worst nightmare and you don’t want that to happen.
Steve: You need to balance what you think you are trying to accomplish, with what you see the users wanting to accomplish.
Bret: You are absolutely right.
Steve: I’m just being the squeaky wheel here.
Chris: You’re right, though. It is true. It’s a big mistake that people make is to ignore what people actually want, when they are building a product. That’s something we try to avoid.
Mark: I think, yeah, I think, part of why you are getting this kind of feedback is because what you built it so good so far, and we start putting: oh, if it could do this it would be even more magical and if we could do this, it could be even more magical.
You built something of value already, so congratulations. I’m addicted to it. Right? But, I want to use it in another few ways and I can see a lot more potential for getting normal people into FriendFeed that are frustrated with it because of the noise problem.
Where I’m coming, with the ability to search, will cure the noise problem. At least cure it for a large number of people, because I can say here’s how you can get rid of the noise. Here’s how you can get rid of Scoble for instance. One of the common questions is how do I get rid of Scoble? I can’t get rid of them.
Steve: OK. So, Barbara K. D. says: “Need to ignore these early adopters. Need to go to the average user.” Marc now swears that. You know, Barbara, you do the biggest mistake that could be made is to ignore the people. We maybe early adopters, we’re also enthusiasts. We’re passionate about this.
We’ve been, in many cases, like mine, I’ve been living in this technology for 15 years. I was trying to extort these kinds of features out of Lotus back in the Lotus Notes 4.5 release. I’m back in 1743. God, so what that we’re early adopters? Who else is… we’re all early adopters. You’re on a chat room on Uscreen TV. What are you talking about?
Barbara: Not only that but go back and listen to what I just asked for. That was not a feature for me. I can deal with the noise. I don’t give a shit about the noise. I can go through thousands of posts every two hours and deal with the noise the way it is.
The reason I’m asking for these features is I keep showing them to people and they keep rejecting it, because it doesn’t serve their needs. And I’m communicating that the market is telling me that there are other needs that are unmet that I want to see happen, because I want to see more people on FriendFeed.
Bret: Yeah, no, I think, you’re right. It’s controlling the velocity of the feed or something like that is actually one of the things that we are looking at right now. So, that if your Scoble and you just can’t get enough, you can get it at top speed.
But, if you are someone who only wants to look in at FriendFeed once every couple of days, we should accommodate those people just as much as we do the people who are every two seconds.
Barbara: And even me, I switch roles that way. Sometimes, I’m extremely noisy on the these systems, because I’m really addicted to it and I’m focusing on it. Other times, I have work to do. Here the reason I was late on this call, I’m doing a show at Revision Three.
I’ve been offline for three hours, four hours. And now, I’m trying to catch up and it’s really frustrating to try to catch up, because I can’t see all the posts that have four comments on them.
Bret: Sure. Sure.
Steve: Well, again, I promised that we would talk about the comments. Do you have any… what I think, Scoble is saying is that the comments are out of control. The comments are spread across rooms, identities, everything, but some sort of social graph model that is understandable.
Barbara: That’s part of it, I’m going…
Bret: I didn’t ask you, Barbara. I was [inaudible] what you said.
[breaks in audio]
Steve: Sorry. Go ahead.
Mark: I’m not sure I understand what you mean by spread across identities.
Steve: All right, so, Scoble keeps telling me to come over to FriendFeed because they’re discussing, you know, something.
Mark: Right. FU, FU. [laughs]
Steve: Right. But, I don’t click there, because I have no idea what he means and because it’s a TinyURL, I have no way of looking at the URL to see what it might be going to.
But, let’s say it was going to go to one of three possibilities. It’s going to go to a FriendFeed location, which is based on the identity of the user, right? So, FriendFeed.com/Scoblizer.
Steve: That’s one possibility. The other one is FriendFeed.com/and room name.
Steve: And is there a third alternative in terms of finding comments?
Bret: Yeah, a comment cluster. I can link you to let’s say Steve Gillmor’s Twitter post that caused 40 people to get up in arms and tell Steve to basically…
Mark: Which is basically a single entry within Robert’s feed.
Steve: OK. So, if I’m interested in let’s say one person’s comments, the only way I can do that is to go to their site, right?
Mark: Yes, so there’s a… If you visited Robert’s FriendFeed page, there’s a link on the right where you can see all the entries that he’s commented on with his comments highlighted in them.
Steve: Right. The problem there is that 95 percent of them are him screaming at people who are coming in off of his Twitter posts with @ signs. What I want to see is the thing that I was interested in. There’s no coherence across a specific topic and a specific idea, except for these comment clusters you’re talking about. And what they fail to do is to roll up the rooms or any of the other stuff that’s going on.
Mark: Yeah, and I think, that’s pretty accurate, like if you wanted to find across your friend network, all the comments on a particular topic, search is really the only reasonable way. And I don’t think it’s completely the thing you had in mind, but it’s probably the most effective way right now.
Steve: No, but as Robert points out and as you pointed out, search is going to be an infrastructure that’s going to be in place for trying to, after the fact, boil down these conversations.
Mark: I mean, like right now I searched for Gillmor on my FriendFeed. It shows me all the people in my network who have been talking about FriendFeed having talked to Gillmor Gang. And so it’s certainly not perfect, but I think, it’s…
Steve: No, but it’s certainly a good thing. And it will expand over time to include “ands” and “ors” and various other kinds of things where you use the social graph to filter out people that you’re not particularly interested in what they have to say about anything.
Steve: Right. And I don’t mean that in an insulting way. It’s just that…
Mark: No, no, no. I think, we’re definitely… I would say that from a user interface perspective, we’re 100 percent focused on synthesis of information and filtering information. And we’re pretty much all working on variations of that right now.
Steve: Right. So, the question then is whether or not you can do that in real time. In other words, if somebody’s talking about something, and they come in, and you don’t know who they are. But, they say something that is apropos of a comments cluster or whatever the architecture that you want to describe this as, that it immediately comes into a track filter, and is available to be absorbed, and then filtered as well.
Then, you’ve got the makings of a conversation engine, which is the highest value of this architecture. It doesn’t mean that all these other things that we’re talking about aren’t extremely valuable, it just means that the one thing that you can’t do, except on Twitter, if it works, is this.
Chris: I’ve been a critic of the FriendFeed commenting model, personally.
Bret: OK, why?
Chris: Because I did see FriendFeed as an aggregator of sorts. And I felt that it defuses the conversation that’s happening on the target item. And it sort of bifurcated all the conversations to the point where some comments were happening on FriendFeed, some were happening on another person’s item in FriendFeed, and some was happening at the source. And some was happening on Twitter.
If it was, to me, a conversation tracker where you could track all your friends’ activities across multiple social networks, instead of achieving that, it’s turned FriendFeed into a destination site, which is maybe the goal. But, it seems to be just more noise, and it’s hard for me to follow anything.
Steve: The other way of saying that is that there’s one site that does do that. And it’s called Twitter, when it works.
Chris: Well, yeah, Twitter doesn’t allow you to… It doesn’t have an explicit way of commenting on an item. You have to jump back to that item and comment.
Steve: That’s not true. It’s got this TinyURL. It works perfectly well.
Chris: No, that’s what I mean. You TinyURL. You go to the target site. So, it’s a blog post, for example. And you comment on the blog post.
Steve: You could create a TinyURL that had a mind type flag of some type that would open it up in the same container. You can do exactly what you’re doing with FriendFeed, their comment model.
Chris: But, that’s a hack around. I’m talking about FriendFeed encouraging it.
Bret: That’s not a hack. This is not a hack. I mean, HTML is a hack. I mean, everything’s a hack if you want to say that.
Mark: Twitter does not have comments, Steve. Twitter has conversations.
Steve: I just gave you a disruptive idea. And you didn’t hear it.
Mark: The comments… I mean, we can go down the comment rat hole for a long time if we want to. [laughs]
Bret: Well, that’s true. At the end of the day, one thing that is… At the corral, we respond to what our users what to do. And I think, that the people who use FriendFeed a lot really enjoy the commenting functionality. And that is first and foremost in our list of priorities, making a product that people love to use.
And so, certainly, I think, it raises a lot of… We’re constantly interested in talking about conversation fragmentation, and unintended fragmentation, and whatnot. So, at the same time, our priority goal would be to our users and making a product that they love to use.
And so, in this case, this is certainly the thing probably that drives most of our growth and makes our product something that people like to use. So, that’s why it’s there. And that’s why we’re continuing to put a lot of work into it.
Steve: Well, let me say that your willingness to come on the show, and answer all questions, and put up with all of this bullshit is fantastic.
Bret: Oh, absolutely.
Bret: We’d be happy to come on anytime.
Steve: The people in the chat room are just saying this over and over again. And in fact, the next time that we have you on, if you’re willing to come back, I hope you’ll take a look at the chat room so that you can answer their questions as well, and not just ours.
Bret: You know, I didn’t really know that you streamed TV into the chat room’s habit. But, next time, absolutely, we can answer more questions directly next time.
Steve: All right, so I want to around the table. Is there anybody else on the call? Mark, are you still there?
Mark: Yeah. I really like FriendFeed, though I don’t use it. It’s definitely a feature. I think, it’s key and core to our future. And I want to encourage you guys to ignore Steve, and do what you want to do.
But, don’t forget about all the rest of us who don’t have the infrastructure and the budget and the funding you have, because I think, your infrastructure could be just as important as Twitter’s.
Bret: Well, we really appreciate all the comments, and feedback, and everything. And we’d we happy to come back anytime to answer more questions.
Steve: Fantastic. And let us know as soon as the XMPP gateway is up.
Bret: We absolutely will.
Steve: Once you do that, I’ll click on all these Scoble.
Bret: [laughs] OK. It was a pleasure to meet you all.
Steve: Thank you so much, both of you. This is Steve Gillmor. This has been the Gillmor Gang for Friday May 30, 2008. And I want to thank everybody who showed up, and especially those who didn’t show up.
We’ll see you again next time. Byebye.