Hacker News new | past | comments | ask | show | jobs | submit login
Why there is no Facebook killer: the death of the P2P dream (2014) (disconnectionist.com)
123 points by wheresvic1 on Sept 12, 2016 | hide | past | favorite | 136 comments

This scarily accurate extraction of our personality and needs is advertising gold, and advertisers will pay a lot more for it.

I once liked a dance company on Facebook because I know the person who made their website. Now I see adverts for dance productions. That's not advertising gold; that's a waste of advertising money.

I reckon most people's profiles are like this - a few likes or posts about something that you'd never actually buy and all that ad money is just thrown away. 90% of the adverts could be getting your profile wrong but simple confirmation bias makes us notice that 10% far more than the useless, poorly targeted adverts, so when advertisers survey people they get the impression their profiling is working. Or worse, the ad networks know that the profiling doesn't work but they carry on the myth in order to charge advertisers more.

I might well be wrong on this, but while I see adverts for things that have no appeal to me I'll keep suspecting profiling doesn't actually work.

>I once liked a dance company on Facebook because I know the person who made their website. Now I see adverts for dance productions. That's not advertising gold; that's a waste of advertising money.

I think it may be more subtle than that.

- not all clicks are created equal: clicks advertised to people like you might be cheap, so why not pad out the "gold" budget with a few chances elsewhere?

- the advertiser might be targeting incompetently and hoping for the best (possibly even succeeding!)

- you might coincide with 90% of "gold" clicks --> still worth it

- your demographic might be an experiment

- etc. etc. etc.

Marketing seems - to me - to have an awful lot of "let's make assumptions and test them" about it. With anything like that, you're going to reach people who aren't interested.

Advertisers don't really care about the wasted clicks; they care about the successful ones. If Facebook can give more efficiency (even 10% more) than other forms of advertising, which it tries to do through targeting, then it will attract more advertisers.

As long as this is the case this is a recipe for encouraging users to block ads. I'd welcome relevant ads. I'm always about to buy something. If you could predict what I was about to buy and give me truly relevant ads, I'd be delighted.

For instance, right now I'm really looking for the best way to deep clean the upholstery on my couch. Let's see what's in my Facebook feed right now...

1. A mortgage ad talking about HARP... nope, don't qualify for that 2. Home Depot wants me to buy the Arctic Cove Fan... In the midwest, in mid-September... 3. Vizio... I'm not in the market for a TV, but I do like TVs I guess... but who doesn't... 4. HTC Vive... already own it

It just goes on and on.

The reality is that not all progress can be brought by for profit organizations. There simply are systemic improvements that can't be monetized.

These should have been the job of the governments, but unfortunatelly, those too are the puppets of big corporations and private interests..

Some people should just break the cat and drive this type of change as an open source movement. I for one am dedicated to try..

I don't think it's a coincidence that Linux, the SYSTEM of OPERATION that actually runs the world is supported by an open source movement..

But there's a pretty specific analysis tool for this distinction. Goods like this are called "public goods" and they are both inherently non-rival and non-excludable.

What you say is fine, so long as this tool is in used. But people don't tend to do that.

Also, don't assume that the manner in which open source came to dominate software was inevitable. Had AT&T and SCO been able to come to an accommodation there's no telling what the world may have looked like. Plus, Linux is exceptional because Linus was exceptional. He's a stubborn guy :)

Actually on a second thought, I don't think that supporting data privacy is the job or the interests of governments. In fact, they are usually the first ones ready to sacrifice privacy in order to provider you with "security" services.

I wouldn't even define data privacy as "public goods" because you now, it's private, it's really a whole different thing than what governments are willing to provide.

Profit making is again in conflict with data privacy, because there are much more money to make by profiling the hell out of your customers then by selling technology to aid real data privacy.

So there's this niche of progress that can't be pushed neither by governments nor by businesses.

So in order for it to happen, there's the need for another force to drive it. And in my oppinion that force is the open source / grass root movement .

And I don't agree that linux is special because Linus is special. We're talking about a mindset here, and Linus is not the only one having it.

I use an adblocker. If your website doesn't like that, I'll stop using your website. So far as I am concerned, how you make money is your problem, not mine.

Data privacy isn't obviously a public good - it's both rival and excludable. There's way to look at it as non-rival, but since much of it comes down to trusted key authority I'd say that makes it rival.

I just recall the world of shareware before open source. It was certainly a way and much was done in that way.

It's interesting you would blame a lack of government progress on businesses. Historically businesses are happy to jump on R&D the government has done and made public.

I suppose for any potential innovation there are those that are opposed, but there are those eager to profit from it too.

The inefficiencies of advertising and marketing is the True Barrier to Entry for modern companies. One needs to have the financial resources to throw away 90% of one's advertising dollars - because that is not really throwing it away, it is padding and paying the Internet infrastructure that will pad and prop your company's reputation because you're a good revenue source for them, as you buy ads from them.

You can go into your Facebook advertising preferences and fix that, so you can still "like" the dance company but not see ads based on that.

Or just block all ads if you prefer.

Stockholm Syndrome.

What is that supposed to mean?

S/he means you think YOU'RE the problem, rather than the victim.

I don't think Facebook will be dethroned by another Facebook. I also do not think the differentiator for such a new king is technological (e.g. P2P). The only kind of FB rival I can imagine succeeding is an open ecosystem one.

A second FB would have to overcome the network effects, and the only chance for that to happen would be for FB's network effects to decrease.

While negative network effects do occur -- it's what killed Orkut, for instance -- Facebook has learned that lesson well. FB's value for participants has decreased, but is now stable and that network is still a huge barrier for entry.

Technology is largely irrelevant. People do not care if it's P2P, runs on functional code or if it is an engineering marvel. They care about the value provided to them. The article posits that P2P adds negative value, and it is probably right. The article infers from that to the impossibility of a FB killer, which I think is wrong, because the scope of the analysis is wrong.

An open ecosystem social network:

1) Has the possibility of creating use cases that Facebook can't compete with;

2) May be extremely vibrant, if the ecosystem is large enough;

3) Can't be bought by Facebook, as it runs contrary to the main strategy of a walled garden. (It may be bought in order to be killed, though);

4) Can live perfectly well, for an indeterminate amount of time, in second place.

I can't see any other avenue for Facebook's demise.

Side note: Twitter nearly had these characteristics for a while. They squandered the opportunity when they kicked it's developer ecosystem in the nuts. Sad.

It seems that the constraint "Can't be bought", or "Can't be bought to be shut down", is an important one and fundamentally incompatible with the goals of a VC-funded startup. That's also why Twitter failed in that respect.

Potentially workable governance models to prevent this kind of takeover: - Community-driven open source - Bootstrapped company with limited profit ambition (think GitHub or Basecamp) - Non-profit/charity organization

If you're a startup millionaire and care about the real-world viability of P2P, decentralization, etc., consider donating to / investing in one of these vehicles instead of looking for the next big business model. Any startup ever will sell out to centralization and control eventually, unless they specifically put themselves in a position where they don't control their own ecosystem anymore (having open-sourced the software and gained a large enough audience that can decide to fork if necessary).

In Asia, I see LINE, Viber, and WeChat displacing a huge chunk of facebook usage for a lot of people. The network effects are intensely strong, too.

I think we've reached the stage where second order effects influence the discussion.

So its not just a product which has to be killer, it has to be able to defeat facebooks increasing mass of capital.

For that matter - lots of Facebook competitors HAVE been created. They've just been bought out by facebook or someone else.

I suspect that eventually people will just die out of facebook, and at some point it will gain the moninker of gravebook. Once facebook stops being cool, it will stop gaining younger users and the network effect will end.

Of course, people at FB have obviously considered this problem and have a plan for it. Perhaps they will live forever.

Maybe we'll discover some first order effects of the network effect which limit its growth or even reverse its growth.

Other than these 2 outcomes though, I don't think we have several degrees of freedom.

Dear <whomever wants to kill facebook> - the answer is simple: Leverage email.

Send marked up messages that look good in both HTML and can be transformed into threaded comment strings in your client. Embed XMPP chat, and set up enabled-by-default encryption around the whole thing. Leverage mailing lists for your special interest groups and your public pages. Sell server time for managing the mailing lists and image hosting, and for a web client.

The infrastructure is there, just forgotten.

This has been banging around in my head for a few years now, but I'll never have the time to do it myself. If you can poke holes in the core concept - please do. It needs to be fleshed out.

> Send marked up messages that look good in both HTML and can be transformed into threaded comment strings in your client.

Sadly, that is impossible. The only mail client that does real HTML is Thunderbird (because under the hood it uses the fully featured Firefox engine).

Outlook even in modern versions still uses the horrifically broken Word HTML engine. Lotus Notes uses, depending on client configuration, either a OS-provided web view or its own stripped-down, worse-than-Word engine.

The webmailer options are even worse. Lotus Notes webmail passes most, though by far not all, HTML tags over to the client, which is pretty cool but you can't do anything with mediaqueries etc. as <style> tags get removed. GMail strips next to everything except images. GMX / Web.de are similar.

And mail support on mobile... don't get me started on this one. Android default mail does stuff differently than the Google Mail app; on iOS Mail at least responsiveness sorta works. Android Lotus Notes Mail is a mixture of "works" and "works not", probably decided by the current position of the moon or whatever.

And Javascript: doesn't work everywhere I tested except on Thunderbird, but there it invokes a prompt...

> The only mail client that does real HTML is Thunderbird (because under the hood it uses the fully featured Firefox engine).

And this is widely considered to be a bug and a source of insecurities, not a feature. The few remaining people I know who use Thunderbird use it in text-only mode in order to control its HTML-rendering tendencies.

Most people don't want a full-fledged browser in their email; it opens up too much attack surface, opportunities for "read receipts" and tracking cookies, etc.

That said, I think email is a good and underutilized tool, and it's unfortunate that more MUAs don't support features like digest handling properly. I think there's a pendulum that's going to swing away from public services like Facebook and back to direct-communication services (Snapchat might already be arguably a part of that) and email is already well-positioned there.

But forklifting the web and HTML into email clients is probably never going to happen.

Thunderbird by default doesn't load external web content, you have to explicitly allow it.

Very true.

I didn't appear to flesh out the details enough - the actual contents of the email is simply a transport mechanism - not the implementation of the client. All that needs to happen inside the email is that it looks like a good, well formatted email message when viewed with a regular email client (and interact properly when treated like and responded to as an email), with predictable content in predictable locations.

The app can then take the contents of that email and format it to appear like your standard social networking platform.

A social networking platform is simply a combination of 1-1, 1-many, and 1-all broadcasting of messages; a format which email is currently capable of.

Maybe you have a great concept, but as a general principle that sounds bad. If someone doesn't have your special client, it will be disjointed. If someone does have your special client, why didn't you just use a saner protocol.

Also, email is not really very decentralized. You need to be one of the major providers to avoid the spam filters.

I realized I didn't address the more important of your two points, figured I'd do that now.

> why didn't you just use a saner protocol.

Mostly because your email network with your friends and colleagues already exists. Even more - we already rely on email as the ultimate identity authority for just about everything based on the web. My idea simply adds crypto-based authentication, UI improvements, and chat over the top of that.

Emails are so simple and ubiquitous that people have already implemented email <-> help-desk bridges, this idea just takes that bridge concept and applies it to interactions between user.

Take a moment and picture what it would be like to have every thread in Facebook delivered to your inbox. It would be a mess with the current mailbox UI, but there's not much impedance mismatch. Now picture the reverse - every email you get shows up as a "post" on Facebook. Both consist of the asynchronous distribution of messages from one user to many others, it's just that one relies on Facebook's network of servers, and the other relies on mail servers.

It would not be any more disjointed than any other form of email exchange - just not as well organized as a social network.

> You need to be one of the major providers to avoid the spam filters.

Yes, but you could take advantage of the existing email providers' IMAP and POP implementations, and companies have long worked to ensure their emails are not sent to spam folders; it's not entirely a black art or impossible.

Is this not what Google Buzz was supposed to be?

Google Wave to be more precise :)

Even better - leverage NNTP. Choice of transport is nearly arbitrary.

Gmail / Inbox does some really cool stuff with markup like that:


This sounds like something you could have proposed 15 years ago. It assumes local, custom email clients for "enabled-by-default encryption" and XMPP support. People have moved onto web and mobile email clients, and you can only use whatever is compatible across all platforms.

I also think it overlooks mobile adoption. Many people using the internet today have only accessed it from a phone. They don't necessarily have email addresses. They may not have ever used a full sized keyboard.

This idea does not preclude the use of a mobile app. The mobile app will simply act as a IMAP and XMPP client on the network side, and look like a Facebook/Google+/MySpace app on the user side.

P2P Facebook needs a killer function to overcome the network effect. If Facebook can compete by throwing money and engineers at it, you loose.

Here's a terrible idea.

Build a desktop app that stores all it's data locally (and is easy to configure with Dropbox et al). The extremely popular social networking activity of sharing photos is indistinguishable from sharing all other kinds of files. Now everyone has a more convenient way to loan movies and books that's difficult for the RIAA to snoop on.

Like I said, terrible idea. What's something legal that Facebook can't just outspend a P2P app on? Other than privacy.

What is a "desktop"?

Is that where I place my phone to charge?

Ha, way to make me feel old. A mobile app with one of the cloud storage providers can also work without killing your battery and data plan.

I think you've just poked the point the article author misses.

I'm not sure I see a technical reason why my nexus of identity has to be physically colocated with my person, and there are a lot of good reasons for it not to be. The important point is that, be it wherever it is, my nexus of identity is not Facebook's or anyone else's but mine.

Also, why should a nexus be a 1:1 relationship with your person?

Should a person be allowed multiple nexi?

That's actually pretty close to what Google+ tried with "Circles" when it launched: the idea that people exist in several distinct yet fluid social universes and that there should be a way to apply fine-grained yet intuitive rules to your sharing behavior.

Turns out most people exist in one social universe, namely Facebook.

I'd say it more turned out that Circles were decidedly unintuitive, and far more of a pain to use than Facebook's single list (which can still be heavily customized, most people just don't bother.)

I think the circles were fine, the problem g+ had was that it lacked plenty reasonable features (events, anyone? never understood this when google calendar integration would have been a killer feature).

That's facebooks main use case for a lot of people and it's absence made g+ pretty useless.

Doesn't Facebook also have a nasty habit of surprising people with breaking changes to the way those customizations work? I seem to recall hearing of such things on a fairly regular basis.

Most people over the age of 20...

Absolutely! This is something my prior comment overlooked, so thanks for catching that. It's another good argument for remote hosting, too; correlating identities is easy when they all originate their communications from the same network address, especially if it belongs to a phone or a residential broadband account. When the phone's just a client, such correlation becomes much harder.

I would go a bit more general than what you said, though - instead of a given person being allowed multiple nexi, I'd say that ideally there should be no one in a position to grant or revoke such permission. If the fashion in which I manage my identity is not wholly within my control, to what extent can the identity so managed be said to be mine?

I am actually advocating a new form of authorization that is based on AOL, Prodigy, and CompuServe handles.

That'd be funny if only it were funny, you know? The names have changed, and the scale, but that's about all - Facebook is a better AOL than AOL ever dreamt of being.

AOL had better non-user generated content, same with Prodigy, and better sense of communities (which I see in Reddit sometimes).

Note that I didn't say Facebook was better for the people who use it.

I always wanted to be known as two integers separated by a comma, but I was too late.

Nice, did they finally get a new UI?

I've toyed with the idea of an open source P2P social network for a while now. You're idea isn't horrible, it's essential.

We can keep trying to support start-ups which will compete with Facebook, but they will all either get bought out, or succumb to the same profit motives which make Facebook corrupted by advertising and unchecked government surveillance.

The only way a social network can grow to the size where it can displace Facebook without getting shut down for political reasons is by providing real privacy over a distributed open source infrastructure.

While it may be abysmal from a PR standpoint, It needs to be the kind of thing where you can post a video of yourself running down mailboxes while DUI, share copyrighted files, or even illegal activity such as black market distribution of drugs, weapons, and CP. That doesn't mean everyone on it needs to be a criminal, but it needs to be that level of bulletproof in order to gain widespread adoption.

With the right selection of cryptographic protocols and algorithms, it should be possible to build a distributed system where each user hosts and advertise dozens or hundreds of profiles without being able to access them. The only way to access the data is by request.

The only major issue I see is an immense data and computational overhead if your profile must be encrypted with dozens of public keys. From a usability standpoint, accessing it should be no different than installing an IM client.

Edit: Essentially, the killer feature is being able to ask your circle of friends is anyone's got some extra weed they can sell you, and actually getting responses because there's no surveillance chilling effect.

Terrible? Rather interesting start point, I'd say.

It should be trivial to install, maybe come preconfigured on a stick computer or even run directly in a mobile phone?

The Dropbox functionality should be encrypted and download stuff only when asked for. With very carefully designed access rights. Ideally, people should feel safe putting nude sunbathing pictures there together with their girlfriend/boyfriend. And their sport tournament meant for the whole world. (Make it two small computers, one for (semi) public data?)

Then add blogs with allowed comments. An 80% solution of FB, if the UI is simple.

Add some P2P or global directories of people, so they can connect their servers.

Some more features are needed to follow e.g. your favourite music bands and local music avenues if those only use FB.

(My personal fear with privacy is that I'll end up on some terrorist list, then be sued for big damage by a police officer for emotional damage, because of my dullness and sense of humour. :-) )

Edit: If you and your friends backup your data, it should probably be an encrypted backup. But then, an encrypted cloud service function would probably be worth the money instead.

This is a bad excuse.

Today you can do a Facebook that lives in your own server. Your server living in your house and being a simple box that you can buy like a TV or car and is super easy to configure.

I already have something like this, but it is complex to configure as it is not mass manufactured but custom made for me and a small environment of people.

Everything else, an excuse for not working on finding solutions for the problems that you have to face, like with any engineering problem.

Of course you can't replicate the work of more than 500 programmers(facebook) with just one person, but you can go far reusing open source components and coordinating the work of other people.

Instead of trying to replicate facebook(a clone), focus on making something useful for the people getting advantage of distributed computers. Facebook did not try to replicate myspace.

For example, focus on networks that PAY real money for having a social media they can control(companies).

> For example, focus on networks that PAY real money for having a social media they can control(companies)

What are some examples of companies or vertical markets that would pay for control that is missing at FB? Publishers seem to be looking for new revenue sources, not control. Are you thinking of advertisers?

There aren't many proprietary content distribution networks. Bloomberg is one example. Note that P2P may conflict with a "control" objective.

Synology NAS has valiantly built a standalone networked home server, complete with NAT hole punching (they run a central broker) and usable "sharing" and "consumption" mobile apps that connect to your home NAS. They earn revenue on hardware sales, similar to iX Systems and FreeNAS. Synology has an OSS pseudo-clone called XPnology, which is apparently compatible with Synology's free mobile apps.

Tanium has a billion-dollar valuation for endpoint management and surveillance, allegedly with P2P techniques: http://www.forbes.com/sites/briansolomon/2015/04/15/meet-tan... & https://kb.tanium.com/Unmanaged_Assets

Various people have tried to charge for "social media within the company", and I think the only one that's really been successful is Slack. Maybe Bloomberg chat and Skype count (although comms is not the same as a social network).

> Facebook did not try to replicate myspace.

No, Facebook tried to replicate Friendster (fairly slavishly), as Myspace was getting popular. The only reason Facebook caught on was 1) by starting with college kids, and 2) building it as a platform that other business could build on.

I don't think decentralised P2P is really the answer anyway: the answer is separating the content from the convenience.

Facebook killed blogs because when I want to find my friend I just searched for their name, and if I was their friend their content made it to my wall.

Because of Facebook's convenience my friends use it, and because it's a walled garden I am compelled to use it to.

My dream for this kind of thing would be something akin to DNS but for identity (pseudo or not), and an open API for content.

So, Adam knows nothing about computers and creates an account on Facebook, which creates an IdentityDNS entry for him on his behalf.

Belinda is also not that techy, but likes G+'s layout more, and so has an account on there instead. She also has Twitter.

Zer0cool is a massive nerd, and has their own domain and hand-written blog software. They went to some website somewhere and created their IdentityDNS entry, and has hooked up and validated their blog on their domain as their content.

Adam can search for Belinda and Zer0cool in Facebook, which in turn searches IdentityDNS. He can add them to his Facebook wall, and their posts will show up in some fashion (maybe Facebook shows its own posts normally and only shows titles of external blog posts, FB can do what it wants). Belinda can do much the same with G+.

Zer0cool has also added Adam and Belinda, and reads their posts through muttdentity, a mutt clone but for identity stuff.

OR something. I don't know. Effectively something that means that I can choose FB because I like how it displays photos, or because it has the best mobile app, or because it has lots of storage. Or I can run my own blog because I want that control. ETC.

I shouldn't have to use Facebook just because that's all my Dad can work out how to use.

http://indieweb.org allows independent blogs to aggregate discussion from centralized social networks. RSS used to enable centralized social media to aggregate indie blogs, but is less used these days.

See recent discussion on these topics at http://www.decentralizedweb.net

> My dream for this kind of thing would be something akin to DNS but for identity (pseudo or not), and an open API for content.

SDSI/SPKI had an interesting mechanism for getting around Zooko's Triangle (before, I think, the latter was formalised): identities are keys (and therefore globally unique), with each user assigning his own names to people (e.g. 'Bill' or 'Ted'); users can then refer to other people as 'my wife's Uncle Bill' and 'Microsoft's Bill' and not confuse them.

This is decentralised and federalised to the max, which is awesome.

It'd be great for something like a Facebook killer, in which when I search for 'Adam Smith' I probably want the Adam Smith I went to gradeschool with, as opposed to the 18th century Scottish economist.

Hum, For my point of view Facebook didn't killed blog. Blog is used not only to show pictures of yourself, but writing. Nobody write on Facebook, they only share things. It's totally different, but hey anyway you can make your blog what you when that they are. Multiple account kill the service. You have at less 5 social media on every newspaper website. I don't use anymore twitter and I hope people will do it too. It's really a waste of time using 10 social media. I even don't use Facebook too.

You can write or blog on Facebook but few people choose too. Techies seems to flock to medium nowadays... So there is some kind of theming around hosts that decides where you publish. I can see why, in many cases Facebook can be too private to share in the professional world.

This sounds really awesome. The thing you are missing, I think, is the way Facebook lets you post things for just your friends (people you have approved to follow you) to see. That's a negotiation between the origin of the content and the consumer of the content that a simple DNS-like system doesn't solve.

The problem is, each 'viewer' application (Facebook, G+, Twitter, mutt, etc.) has an incentive to add proprietary features which only work on their system. Not only does this differentiate them from the competition, but it's also an easier problem to solve than if they tried to make it distributed (e.g. Facebook can add features to Facebook much more easily than they can to email, since they store all of the Facebook data and broker all of the Facebook messages; their only problems are how to make it appealing to users and cost-effective to run).

Over time, the walled gardens return.

In fact, we've basically seen this happen before: many people ran their own standalone Web sites on their own domain. The "viewer" was the user agent, which could be your choice of browser, aggregator, etc.

I think the killer feature of Facebook over standalone websites and blogs is that with Facebook you can choose who is allowed to view what you post. Maybe I'm wrong, but I remember people being a little uncomfortable posting pictures of their kids on a blog for the whole world to see. With Facebook, only your friends are going to see those.

> with Facebook you can choose who is allowed to view what you post

Well that's their kool aid, which many people have been drinking.

I still follow the rule that you only post stuff online that you're comfortable sharing with the world: it's not a matter of if the databases will be leaked, but when.

> Maybe I'm wrong, but I remember people being a little uncomfortable posting pictures of their kids on a blog for the whole world to see. With Facebook, only your friends are going to see those.

Use a blog software that allows to enable viewing of some content only to users that authenticate (for example via OpenID) and are of course in the list of allowed users (that you carefully maintain).

> Because of Facebook's convenience my friends use it, and because it's a walled garden I am compelled to use it[,] to[o].

Rather: Because it't a walled garden I am comelled not to use it (because otherwise I would violate the principles that I stand for).

The P2P network doesn't exist, because it doesn't solve a problem. Facebook solves the problem of "I want to keep in touch with people in my life digitally" and has built a sustainable (to say the least) business model around it.

Being P2P by itself is not a reason to create a social network, unless it brings benefits for users that simply don't exist in FB. What are those? Privacy? Security? Only a small number of people care about those.

Products are not created because they are technically possible - they are created to solve problems. This reasoning is flawed because it starts from the solution, not the problem.

Following your line of reasoning it's worth asking what would have to change to make Facebook's business model vulnerable.

When Facebook IPO'ed the biggest technology threat on the horizon was mobile apps [1]. However, the FB folks negotiated that one pretty successfully. Since then they have been very paranoid about new types of chat applications, in short anything that could move communications onto a platform they don't control.

I'm just speculating but the biggest threat to Facebook would be if people started to look for something that Facebook has a hard time delivering. For instance, what if users decided that getting real news from reliable sources was high value? (Something like a return to properly curated journalism rather than entertainment--this is of course just wild speculation.) In this case, it's something that puts a premium on quality of the content, not the size of the network.

Another obvious threat would be new ad delivery technology, which attacks the "buy" side of the network. No doubt FB watches that area very carefully.

[1] http://gadgets.ndtv.com/mobiles/news/facebooks-mobile-transi...

Fundamental problem with article -- says that Skype changed from p2p to centralized because it was a better experience for mobile. It was political, not technological.

I believe it had more to do with 'Project Chess' which was LE/IC access to the Skype calls: https://www.theguardian.com/technology/2013/jun/20/skype-nsa...

Facetime, which is a similar thing to Skype for the Apple platform used to be p2p in ios4. Patent troll VirnetX sued and won for 600 million dollars to get them to stop being p2p. Facetime has never worked as well since, and VirnetX are still trying to get iMessage and Facetime blocked.

No, p2p works fine. Better than centralized. There are forces that don't wish this.

Well yeah, that's one technical reason. But I suspect we would have found a work around if these systems had caught on to begin with. If a site or service is popular, then that can influence what technology is adopted and how we use it.

No, I think the real major reasons are marketing related. For example, Facebook's network effect is very strong, even compared to sites like Twitter and Instagram. The service is only replaceable by one that grows extremely quickly, to the point things don't settle down enough for the site to feel like a ghost town.

And it ties into the real reason said effect is so strong.

These P2P and decentralised networks are only appealing to a certain minority of users.

They're basically marketed as 'censorship resistant' or 'free speech focuses'. And while these are nice things to be, they're not a selling point for 90% of the social media userbase. The vast majority of people either don't feel they're being censored or don't really care about anything controversial.

As a result, the only people using these services are niche groups using subject specific networks and activists banned or unwelcome from the major ones. It's like why Voat failed to unseat Reddit; because the only people who joined were those interested in discussing controversial issues and hence the site wasn't useful for people wanting to discuss more mundane things.

It's the social and marketing factors that really sunk any P2P attempts at a Facebook clone (or any other decentralised social networks). They just don't have value for people wanting to discuss non controversial things.

"Censorship resistant" also means, in practice, that a large amount of your content is awful. Whereas the average person wants something with the sharp edges removed. Facebook's moderators, inconsistent as they are, are genuinely adding value.

I don't see why do you think such system will force you to experience content you don't want.

It's just when your tastes diverge from what's forced on us as acceptable, you shouldn't let in the dark.

There's no way a priori to determine whether I want content without looking at it. At best I can delegate this up to a moderator to look at it for me.

There are entire subcultures of people who make a game of trying to get people to see images they don't want to see. This has been happening since slashdot and goatse.

> There's no way a priori to determine whether I want content without looking at it

This is plain false. HN does it, for example.

> At best I can delegate this up to a moderator to look at it for me

It's an algorithm, not a moderator. Moderators are currently employed for two things: a) censorship, b) influencing opinions. Everything else algorithms do just fine.

HN is also a moderated system, there's this guy called dang you may have heard of.

But he's not going around curating your reading lists, does he? That's the point.

In your own P2P Facebook, you'll be your own dang.

HN's algorithm for deciding what gets on the first page is not totally trivial. Is it published anywhere?

I think it's tweaked periodically, and unpublished to make manipulation harder.

Yep. Another way to put it is that by ditching Facebook in favor of something open-source and decentralized, you automatically become "that guy": https://xkcd.com/1105/

Isn't it also the case that web usage is much more wide-spread now than it was when Facebook became popular? It just feels off to cram all of your contacts into one platform (the dreaded will my granny see this post effect).

I think end-to-end encrypted social platforms could have a future if they are oriented towards a niche that sees advantages from the encryption and offers federated access for mobile devices in addition to full p2p connectivity.

If you want to build a Facebook killer, then answer this question first without coding a single line:

How do you plan to sign up non techie females? Do you have anything to offer to them?

More importantly, if you want women to use your service, how are you going to handle the inevitable abuse problems?

I think if users paid to use social networks then those networks wouldn't have to sell ad data to make revenue.

It makes sense that when Facebook, LinkedIn et. al were set up social networks were something new so people wouldn't understand the benefits enough to pay.

Right now, we've seen what social looks like when you don't pay for it. You become the product.

I would have happily paid for WhatsApp instead of letting them give all my numbers to fb.

> I think if users paid to use social networks then those networks wouldn't have to sell ad data to make revenue.

But the networks would still do it, even if they didn't have to. Look at Microsoft - you paid hard cash for you copy of Windows, and they still vacuum all your data and behave as if they were scummy cloud company.

Even if you pay, there's no reason for companies not to do this. Neither users nor regulators don't punish companies that spy on their customers and sell their data, so for the companies, it's free money. Why wouldn't take free money?

Sure, they could not be evil, but not being evil doesn't take you far in the business world - see every big company ever.

I would love to be able to name a company that's done well without the being evil component.

He omits a huge source of emerging P2P potential: cloud instances run by power users and businesses. These are currently powering the various blockchain communities and represent the next stage of P2P. Due to the security and accountability advantages of blockchains, these networks could lead to a large-scale "redecentralizion".

What blockchain communities represent the next stage of P2P? Twister is the only thing I can think of that vaguely meets that description, and as far as I know it's not particularly thriving.

From my understanding of what he said, he thinks that distributed social media would mean that the hosting would happen from our mobile devices as a P2P network. So yeah, he did completely miss something, that would be a terrible architecture that would not only be completely unreliable, it would offer no scalability, except for really awesome cat videos. Distributed social, such as diaspora allows anyone to host their own social media platform, but also provided a federation interface to other nodes. The problem with diaspora is that it offers a small subset of features compared to Facebook, and the problems they needed to solve, especially with federation and content control, are really difficult. Further, the branding will confuse almost everyone as most people can't be bothered to understand the difference between a site, application, and protocol.

One model that could work I thought of works like this:

A packet based P2P network, where peers can also act as buffers to relay the packets. So basically the same as servers in the traditional sever client model. However your account isn't bound to a server like with email or diaspora. So the networks of computers of your friends can act as servers serving your smartphone. Or the developer of the app you use can provide a server that can also send you push notifications.

The real problem for decentral messaging/social networks is discovering your friends. Friends of friends being suggested to you certainly is a good step forward but it's not enough. It doesn't work when creating a new account and when adding people outside of your friend's circle.

I think what always gets ignored in these "Facebook sucks, we should use/build X instead" arguments is the grandparents factor.

My gran cannot be expected to learn server administration and security just to have an inferior* version of what she has right now with Facebook. For that matter, that's not a tradeoff you can expect almost any non-tech (or most tech - we have other things to be doing) people to make.

(Inferior, meaning the supposed benefits to her are completely philosophical and theoretical, rather than practical. She doesn't give a tuppeny crap what advertisers are doing with her likes.)

Unless she can right now click on an icon that brings her to something that looks like Facebook and has her family on it and is easy to use, all the alternatives in the world don't matter. And I get the sneaking suspicion she's not in the minority here!

Everyone always starts looking at this from the tech side, as if having the system be P2P should be its primary selling point. No! In the real world, that isn't a selling point any more than something being Free Software is a selling point to anyone who doesn't read Stallman. The "threat" from advertisers, corporations, and governments accessing your Facebook data is going to have to get a lot more practical and a lot less theoretical (the NSA might do X Y and Z!) before that even winds up in the same universe where it can appear on a normal person's radar.

The costs of running even a p2p system need to come from somewhere and you need to attract the right people (hubs in the social graph lingo). The trick is incentivizing them. There are going to be adverts somewhere but the people and content producers could get a cut. Synereo has a better chance of succeeding than a wishful thinking p2p open source project because there is a source or funds that can power the network. It still has to come from ads though, unless people pay for it.

You can complain about Facebook's appetite for data all you like, but reason Facebook is successful is that most of the users of the site are willing to trade their data for utility. To beat Facebook you need to create a new product based on something that people actually care about, by providing either a better social network, a better news distribution system, or a better way to share photos and memories (to name a few possibilities).

> It's very difficult, inefficient, and unreliable for your phone to serve out data to all-comers: it's a one-way street.

But your phone can talk with your server and anybody who wants to access what you allow them to, can simply talk with your server. It is perfectly possible to host your own server nowadays.

> Imagine if the messages and photos you shared with your friends only lived on their phones or computers, and not in some server being scrutinised by security agencies, used for targeting ads, or at risk of being hacked.

Yes, it is possbile by implementing the idea described above where everyone hosts its own data.

Actually I'm working on something like this. Currently I have developed some minimal applications that allow you to share your status and chat. It still uses a single server for now to relay events between the clients, but I am working on making it P2P.

You can check them if you want :

https://fruit-fly.herokuapp.com/client/statusapp.html https://fruit-fly.herokuapp.com/client/chat.html

"Your 3G connection isn't like your home ADSL or cable internet connection - it is designed for browsing web pages and downloading data off big servers, not serving up content to the whole world, like a mini version of a Facebook or Instagram server."

If you had a container running somewhere in the cloud talking to containers owned by other users then the mobile device doesn't have to be a server?

You have to sign up for a new paid service for that. Significant barrier vs doing something off your existing wired home connection for free and having the easy to understand mental model of "stuff running on my PC there in the corner".

Heroku/Lambda/App engine help with that a little but they aren't newbie friendly and there aren't many existing server apps for those that you can just install. Sandstorm has the right idea.

And now you have to trust the service provider that the container runs on etc etc etc.

Technology advances occur because something is provided what people want. Advertisement is not what people want. At least not too much of it. If something better comes up Google and Facebook will die.

As with each technology advance it's impossible to know beforehand what will be a hit. Nobody knew that smartphones will be a hit before the first iPhone came out.

However we all know that advertisement is something companies force down the throat of people. This won't be sustainable in the long term. We will experience the Facebook and Google killer.

Perhaps it's neural networks people can train themselves and providing a layer above all communication technology. Perhaps with these neural networks we will recognize that email is truly all we need for asynchronously communicating with each other.

We will see.

Well, tech firms and tech creators want money....so?

unless the incentives for technology creators change this problem is going to remain as old as time.

And we are doing a pretty terrible job of setting incentives which are not monetarily oriented.

Yes that's why the people will perhaps run their neural networks themselves. Because nobody will do that for them for free.

Issue isn't tech, it's that the average person doesn't see any value in what P2P offers.

Not to mention the average developer. I just spent the past few days wrangling NAT-punching issues for a small p2p project of mine. To say "the first rule of p2p is don't do p2p" is an understatement.

The (very unfortunate) truth is that p2p is just more difficult than client/server.

That problem is partially solved by using a supernode overlay structure. It isn't particular elegant, and it relies on certain nodes being willing to work more for the network - but it can enable nodes behind a NAT to parcitipate relatively easily.

Sure, sure (and the tip is much appreciated), but this just goes to show that P2P is more difficult than client/server.

The point is that one has to have a very good reason to use a P2P architecture, as it tends to create problems.

Maybe it's me, but I'd assume the many-to-many is more complex that one-to-many, especially when laid on top of systems designed for one-to-many.

Good point. I was thinking of issues like peer discovery and NAT punching, but it's true that traffic topology changes as well.

So it's the same issue as with many failed startups - solving a problem, that does not look like a problem for the end user. At all.

TLDR: someone has to pay the bill. If users don't, advertisers will.

Not a problem for email, which is P2P.

Most peoples personal email these days is paid for by advertising.

Sure, but it's not something that makes the P2P not work.

The "problem" is that users have to either administer a server, pay for a service or accept ads, and they mostly choose the latter. I'm not crazy about the fact they make that choice, but the fact the choice exists is not really a problem with the protocol.

At the moment the best plan for that seems to be, make sure that if Facebook somehow kills itself the best alternative is a P2P alternative.

Excessive advertising or conspicuous, harmful info sharing seem like the most likely ways Facebook might drive people away. Obviously the network effect means they're going to have to fuck up quite badly to drive people to a competitor.

The reason why there is no Facebook killer yet is that the telecommunications act, [1], does not extend yet to mass communication in the form of social networks.

[1] https://en.wikipedia.org/wiki/Telecommunications_Act_of_1996

Care to expand? There was a brief period of telco (CLEC) competition, followed by a series of mergers which returned the landscape to almost the same as the one preceding the Act. If such an Act applied to social networks, how would you envision the effects?

Well, take this quote:

> Since communications services exhibit network effects and positive externalities, new entrants would face barriers to entry if they could not interconnect their networks with those of the incumbent carriers.

Since social media are communications services, and also strongly exhibit network effects, the act could not only lower the barrier to entry, but actually create the possibility of entry (which is now virtually nonexistent for any newcomer, even if you have the money of Google).

Good point. In practice, ILECs did everything possible to delay the provisioning of interconnects, until most CLECs failed. But it was a good objective, albeit difficult to enforce. Without a physical layer monopoly (social networks), there would be less plausibility for broken interconnections.

Their has been no Facebook killer because Zuckerburg has bought anything that he felt will supplant Facebook or at least take away users from them. He bought Instagram and WhatsApp was after Snapchat as well but Snapchat founder didn't want to sell.

Zuckerberg is a smart competitor. If only he'd sell the site to Yahoo or Murdoch it would be much easier to compete.

Maybe we need something in the middle of p2p and centralized. Main big server by company works like current facebook and the ability to create your page on your p2p node if needed. And you will can choose between a general server or host p2p node...

Very few people care about the issues raised by the author. From a practical pov, I think the best way to create something like this would be from some prebuild community that might care like Burning Man or something.

Some Retroshare developers are working on an mobile app for Android and like Bittorrent it won't need a server farm. Peers can carry their own weight using home servers that can talk to phones etc.

I think a P2P social network that positioned itself as a counter-cultural movement could do pretty well. People are willing to put up with a lot to be part of something cool.

> People are willing to put up with a lot to be part of something cool.

Funny enough, this is exactly how Facebook first marketed itself: only for students of Harvard, and then only for college students.

Once it had enough on-board members, it pivoted and started allowing the general public, until today where even your grandmother uses it.

A few people are.

If P2P is dead, does that mean that 99% of cryptocurrency traffic go through server farms?

You are solving the wrong problem. The reason there's no FB killer is the network effect. You can create the best alternative in the world, but to actually populate it with users is the task that nobody knows how to solve.

> You can create the best alternative in the world, but to actually populate it with users is the task that nobody knows how to solve

Nobody except Instagram, Snapchat, Pinterest, et cetera

S/he means that the core problem in creating a "Facebook killer" is not creating a social network, but figuring out how to get users through it. Instagram, Snapchat, Pinterest etc did not use any new protocols or tech, but they were able to attract users and grow through their distribution methods (on mobile) and features (not their underlying tech). Using a new P2P protocol will not attract enough users to beat Facebook.

Instagram = photo sharing app

Snapchat = chat app

Pinterest = photo sharing app

They are not social networks. They have social elements, but that's not their core design. And none of them are FB killers.

Google+ and Diaspora are examples of social networks, but they couldn't solve the network effect problem.

Umm Facebook lets me chat and share photos with family..?

Yes, so? These are secondary functions.

There is no Facebook killer because of the lack of interest of EU industralists for the Internet. They dont care because they dont see it as an infrastructure. They think it's a gadget. De minimis non curat praetor.

The article is very ignorant. The internet remains primarily a P2P technology, and will probably continue to be that way forever. The amount of valuable resources on the internet which are utilizing P2P methods are 10000000 to 1 compared to the number of Facebooks.

Also consider that Facebook's #1 product, social blogging, is not profitable, and can never ever be profitable. 100% of their profits are made in other businesses which they've connected to their brand name, which is the only reason for their success. For example, microtransaction-scam "video games".

This is a typical article by someone who does not really understand the internet and has not been using it for very long.


>users can now own their own identity.

Facebook's main feature is precisely the opposite of this. How can this be facebook killer.

Because if users own their identity and there is a new social network built upon the basis of owning your own identity and take it with you were ever you go, which is something most people want, then FB would be screwed. They would have to come up with a new revenue model to compete.

Ad blockers are becoming more and more popular. A social app built on Blockstack would prevent users from being tracked.

There is no Facebook killer because the average user doesn't care about privacy. It's the same misunderstanding about Apple. Sure, techies of the world might rather have a rooted droid, but the vast majority of the world just wants something that works. They don't really care about ideals.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact