Stop worrying. According to internet physics facebook will gradually diffuse into distributed services.
The first mass migration away from the silos is likely going to happen as early as within this decade. As soon as some kid in his basement creates a set of viable protocol specs (read: not diaspora) and as soon as a few relentless hackers smell blood in the water.
Once initiated no Zuckerberg or Schmidt has the resources to argue with the second law of thermodynamics. The social graph is destined to gravitate outwards until it's more or less evenly distributed over our always-on devices. Supported by a network of interchangeable service-providers for caching, storage and, of course, "apps".
There you have it. But don't let that stop you from writing the next series of sky-is-falling posts. They're part of the process.
When has this "diffusion into distributed services" happened before? It seems to me the trend is most often in the other direction.
There were once many search engines, now there is Google. Usenet was a distributed system; now there are centralized forums. Email was once something that was largely administered locally, now everyone has a Gmail account. Sharing things with friends and family used to be done through a constellation of different services, like blogs and Flickr and email. Now Facebook does all of those things.
Go back a bit further to the days of the big online services like AOL, Compuserve, Prodigy, etc. I mean before they were ISPs. They competed with the internet and lost, because they couldn't match the wild creativity unleashed by the web, which made everyone a publisher/broadcaster/developer/priest.
Social networks have taken a few of the biggest distributed applications of the web and started mass producing them, in watered down form. But notice how they have to keep absorbing every new fad that comes along? The coup will start with some platform that allows ideas to be realized faster than the networks can clone them.
Search is not easily distributable and will remain silo'ed up for a long time.
GMail is huge. But SMTP is still a distributed protocol, you can switch providers in a blink and not lose anything other than a shiny interface.
NNTP was always run in a hierarchical fashion and more importantly is centered around discussion groups. People a) prefer more control over their publishing/group-comm (blogs, forums) and b) demand the "follow individual" pattern (RSS).
Consequently the Usenet-monolith shattered into said services, all of which you can either run on your own server, or rent from a service-provider.
Sharing always sucked and that is the niche (apart from the social graph) that made Facebook big. However, today we can build mashups that plug into all these services and present them under a unified interface - just as pretty as facebook.
The only missing bit to lace it all together is the social graph. Today our little mashup has to load the graph from either facebook, twitter or google.
In the future some smart kid will come up with a "DNS[1] for social relations" and everyone will start using that, just because it's the natural thing to do and because futzing with proprietary API-keys is so 2011.
But SMTP is still a distributed protocol, you can switch providers in a blink and not lose anything other than a shiny interface.
I think you may underestimate the importance of the shiny interface. Regardless, the use of an outside provider is still a move toward centralization. Ten years ago, the small offices I did software work in all had local email servers. Today, running an email server well is a big job, and those offices all outsource their email service to much larger companies.
People ... demand the "follow individual" pattern (RSS) ... In the future some smart kid will come up with a "DNS for social relations"
It's interesting that you bring up RSS. I was thinking about that too. I think RSS is already very nearly a "DNS for social relations." But it's also failed to catch on outside the circle of techies, early adopters, and web-potatoes.
I'd say the failure of RSS to achieve mainstream popularity is an example of a decentralized system losing out to centralized alternatives. Like Facebook. Facebook is just plain simpler to get started with than a self-administered approach like RSS or a hypothetical next-generation DNSfSR would be. And Facebook and Twitter both killed off their RSS feeds once they'd felt they'd achieved a certain critical mass. They understand that RSS is an alternative for much of what they do.
I don't celebrate any of this. I use and like RSS. I wish I didn't feel the need for a Facebook account (but that's where my real-world friends are). But the pattern I see is that convenience beats out control every time. For complicated devices like computers and for complicated situations like the (real, human) social network, the inevitable complexity that comes along with local, personal control is often a problem, not a feature.
Under such circumstances, I don't see how Facebook (or its successor) will be overtaken by a distributed anything.
I have a gmail account, but I don't use it. I would say the same for any business. No business with any sense is going to rely on Gmail for their internal business communications.
Any companies on that list using Gmail as their primary email service are foolish. But I'm not sure why you think I was somehow equating Facebook with Google simply because I don't agree that everyone has Gmail and is using it.
Ah, now here's a different thing: should a comet fall on Google's data centres, you still keep your data. I infer that you probably thought about privacy risks. I also suppose your company's email address isn't "@gmail.com". That would be foolish.
Despite the other comments, I completely agree with you.
It's just foolish to give up control of one of the most important communication resources your business has to a company with a notoriously poor customer service record.
That's your prerogative, of course, but more and more businesses are outsourcing their internal email to an outside service. Sometimes that's the commercial version of Gmail and sometimes it's a competitor.
Regardless, that seems like a move toward central control rather than toward anything I'd describe as a distributed service.
I'd bring up that "outsourcing their internal email" is not the same thing as "using Gmail". But as to your point about centralization, companies (like Facebook) have tried for years to centralize data and it hasn't worked yet. Microsoft has tried it, IBM has tried it, and Google continues to try it.
I think the issue is that there is difference between allowing people to log into an account with Facebook and companies actually sharing their data with each other. There's also the reality that as companies die, merge or change course, the massive amounts of data they have often simply becomes unused.
Have you ever suffered a Google outage? I don't mean the kind of system-wide failure that gets all Googleplex hands on deck. I mean the kind of random, individual catastrophe that comes with a crash-course in just how astonishingly unreachable the Googleplex really is.
Yes, I have first hand experience with this. Dealing with Google "support" taught me a harsh reality that Google simply isn't interested in support in the way that Fortune 500 companies need and expect.
How is this any different from storing source code on GitHub, or using Campfire to communicate?
A very reasonable argument is that you shouldn't be using those either. GitHub is fine for open source projects but if I were involved in writing proprietary code I'd probably not put it up there.
You can't guarantee anything. It's all about trade-offs and managing risks. Putting your source code onto a private server doesn't guarantee that your source code is safe, and there are far more examples of source code being stolen from private servers than there are of companies appropriating source code from hosting companies like GitHub.
This is true. But would the story get out if it happened? And it's not limited to appropriation. There are stories about, say, Assembla, and the terms of code storage changing dramatically. You could also have a company where the security on transferring the code was inadequate. That's a concern you don't have as much if your servers are local (though you need VPN for remote coders).
It's also hard to talk percentages -- there are a lot more companies storing source code on private servers than using external hosting. So yes, more stories of that, but it's hard to talk percentage chance rather than absolute numbers.
Is the risk large enough to justify using a more expensive and less functional in-house system? To me, the risk of using a third-party host seems vastly outweighed by the benefits.
Well written articles never get boring to me, even if I don't agree with the author or if it's not really new information or particularly deep in some sense. I found this article a very good read.
Could you give an argument, why there should be something like internet physics, which follows the laws of thermodynamics or the laws of gravitation? To me this sounds rather metaphoric and like an abuse of terminology.
>Once initiated no Zuckerberg or Schmidt has the resources to argue with the second law of thermodynamics.
Life processes move in the direction opposite of the 2nd law. Though I generally agree, if we view the market and technology as evolving we should see oscillation between diffusion and convergence.
I get your point, but I'm still uncertain about whether to feel happy or disappointed at the fact that everything goes in circles.
We (the internets :D) spend all this time getting out of the AOL walled garden, only to go back to the Facebook walled garden. Seems like an utter waste of effort.
But maybe the glass is half full, and we will be out of here soon.
In a sense, isn't your entire social life a waste of effort? Even if you think the ultimate goal is to construct the ultimate hivemind, massive cultural flip flops can still represent progress.
I think something big has to happen before people start leaving. Diaspora can launch tomorrow, but I don't think people will start flocking to it. Most end users don't care about stuff like this.
I'm inclined to think the same way, that there is some kind of natural / emergent property of networks that acts as a counterbalance. Trouble is, how do we know this isn't just wishful thinking.
It's less an emergent or mystical property of networks, and more a function of human psychology. The decision to keep using Facebook, or to abandon Facebook for alternatives, depends on each user's calculation of Upside vs. Downside. As soon as alternatives emerge that offer equal or greater upside with less downside, users leave for those alternatives.
The one thing keeping Facebook alive and well right now is that no alternatives yet exist whose upside outweighs the hassle of losing all of one's Facebook friends, data, etc. (In economic terms, this is known as a very high switching cost).
Sooner or later, though, it's possible that a) more alternatives will emerge; b) some of these alternatives will be massively more attractive than FB for one reason or another; or c) FB will become increasingly annoying via feature creep, privacy concerns, homogenization of experience due to lowest common denominators among users. I'm banking on "c." There are, and will be, viable competitors to Facebook. But these competitors won't get a lot of attention, relatively speaking, until Facebook becomes more annoying than it is convenient.
Here's my take:
Stop worrying. According to internet physics facebook will gradually diffuse into distributed services.
The first mass migration away from the silos is likely going to happen as early as within this decade. As soon as some kid in his basement creates a set of viable protocol specs (read: not diaspora) and as soon as a few relentless hackers smell blood in the water.
Once initiated no Zuckerberg or Schmidt has the resources to argue with the second law of thermodynamics. The social graph is destined to gravitate outwards until it's more or less evenly distributed over our always-on devices. Supported by a network of interchangeable service-providers for caching, storage and, of course, "apps".
There you have it. But don't let that stop you from writing the next series of sky-is-falling posts. They're part of the process.