Ask HN: Why don't we have widespread, decentralized social networking? | Hacker News Reader

Ask HN: Why don't we have widespread, decentralized social networking?

4 pointsnobody999915 comments

With decentralized Social Networking (SN) platforms such as Diaspora, Mastodon, Manyverse, etc. it seems like a no-brainer for many people to implement both personal and small group nodes that can either run standalone or in a federated environment.

It seems that there are a number of blockers to widespread adoption. Some of those are:

1. Inertia. Most folks are already on centralized SN platforms, as are their contacts, friends and family;

2. Complex installations[0] keep many folks from even attempting to install such platforms;

3. Lack of symmetric, high-speed broadband access limits the ability to self-host such platforms;

4. Limited/expensive access to domain names, DNS hosting[1] and static addressing (in the absence of widespread IPv6 adoption by consumer ISPs).

Given that most consumer-grade computing systems are mostly idle and storage costs continue to plummet, the argument for decentralized platforms just becomes better and better.

As such, why aren't more folks promoting them and/or developing mechanisms for broader access/acceptance?

[0] Yes, there are Docker images for many of these platforms, but implementation/adoption of Docker is limited and most of these platforms require significant IT and development knowledge/skills to implement.

[1] This could also be addressed via containerization, but there seems to be limited interest in promoting such self-hosted platforms.

Because the average person wants to be where their friends are, and us something that just works. Modernity poses so many challenges that decentralization is meaningless to most "normal" people, whether through ignorance or lack of care.

More importantly, you seem to place a higher value on decentralization than centralization. Why is that? What if centralization allows more people express themselves more freely? Would you place a higher value on centralization?

Decentralization is a design choice, that's it. Unfortunately, we have to come to terms with the fact that outside the sphere of hackerdom and a select few privacy-conscious individuals, decentralization is completely and utterly irrelevant.

>Decentralization is a design choice, that's it. Unfortunately, we have to come to terms with the fact that outside the sphere of hackerdom and a select few privacy-conscious individuals, decentralization is completely and utterly irrelevant.

I'd posit that decentralization is a business model choice that has design implications.

A decentralized model means that one entity can't control access to users, their data or their eyeballs.

As far as privacy is concerned, that's not really an issue WRT centralization vs. decentralization since the whole point of social networking is to share information with others.

Rather, what folks present as "privacy issues" are really more about control, monetization and engagement.

I want control over the information I share. If I want to keep it private, I wouldn't share it.

What's more, I despise advertising and contact from individuals and corporations whose only interest in me is to extract value (whether that be money, attention or access to my social graph) from me. I'd say that most people, if given the choice not to have that in their social interactions[0] would choose control/lack of advertising/marketing overwhelmingly.

So no. It's not irrelevant at all. Rather it's a matter of having a viable alternative.

Unless, of course, your paycheck is contingent on such centralized platforms. Is yours?

[0] Let's say you're at the pub having a pint with your friends and a group of strangers stood nearby listening to your conversation, and every 30 seconds or so, one of them started pitching some product or service to you? How long would you continue to frequent that pub?

> What's more, I despise advertising and contact from individuals and corporations whose only interest in me is to extract value

I completely agree with you here. Despise is the exact word I would use to describe my feelings towards advertising.

> I'd say that most people, if given the choice not to have that in their social interactions[0] would choose control/lack of advertising/marketing overwhelmingly

Maybe it's just me being older and more cynical but I really don't think most people care about control either, at least not in large enough numbers to matter. I'd be happy to be wrong, of course.

> Unless, of course, your paycheck is contingent on such centralized platforms. Is yours?

Not at all. But I am building something in this space, with decentralization/anonymity as a priority. It is entirely possible for me to be in full support of a move towards decentralized services while accepting the reality that it's not important to most people. Pragmatism does not equal defeatism.

>Maybe it's just me being older and more cynical but I really don't think most people care about control either, at least not in large enough numbers to matter. I'd be happy to be wrong, of course.

Given that I'm in my mid-fifties, I'm guessing that cynicism has little to do with your age.

I'm not especially sanguine about the masses being all that concerned about control or centralization vs. decentralization either.

That said, I do believe (perhaps I'm just not cynical enough yet...and I hope I never am) that providing those "most people" with an alternative that allows them to maintain the connections they have (via federation with centralized services) and the added value of controlling their own experience using a tool that's simple to use and ubiquitous (think bundled web browsers), they'd jump (or rather use what you dump in their lap) at it.

>Not at all. But I am building something in this space, with decentralization/anonymity as a priority. It is entirely possible for me to be in full support of a move towards decentralized services while accepting the reality that it's not important to most people. Pragmatism does not equal defeatism.

Fair enough. My primary point in posting this "Ask HN" bit was to spur discussion about both the societal and technical issues surrounding decentralized network services in general, and SN services in particular.

Especially since the centralized models we see today are the result of artificial barriers to direct interaction. Those are partly due to software/platform advances that have outpaced the consumer internet access infrastructure (that's not an accident -- it's not a nefarious plot either, just monopsonies attempting to protect their privileged status) and the co-opting of FOSS for corporate use (again, not a nefarious plot, just folks limiting their economic exposure by using free stuff), combined with the fact that marketers will pay big bucks for eyeballs on their messaging and even bigger bucks for their messaging to reach their "target" market.

All of this, combined with the vertical integration of network and content providers has created a "broadcast" mode for most Internet services.

All of which is bad for creativity, content creation/distribution, freedom of expression and real (rather than highly filtered/curated) human interaction.

Do I have all the answers? No.

But I do have the suspicion that by solving the technical challenges that keep folks in these centralized silos, we can create a better, more open and more equitable environment on the 'net.

If we do so, and help "most people" understand how this will allow them to have a superior experience across a vast swathe of their lives -- not just social networking, but personal control of their devices (including IoT) without centralized groups both limiting their activities and charging them "rent" on their own devices.

It's a complex set of topics and I'm not doing them justice here.

That said, there are many blockers that keep folks from decentralizing their access to their fellow humans.

Most of those barriers are artificial and exist to enrich those who own the centralized platforms -- and only continue to exist by buying, copying or forcing out their competition.

I've strayed pretty far afield here, but the reasons why we don't have decentralized social networking are many of the same reasons why corporations like YouTube and Cloudflare are so successful.

I'd only add that anonymity (or pseudonymity, as I practice here) has its place and I support it wholeheartedly. But decentralization adds so much more, IMHO.

The alpha geeks of decentralization need to prove they are living better lives. And not just by the virtue of doing so, but by creating combining remixing in a way that is more fun, more interesting, in a way that the staid legacy platforms can not.

It's not about decentralization, to me. To me it was always more about extensibility. About freedom to try new things, to bend tech in ever more interesting ways. There's a wide internet out there, and decentralization wins when it's not just it's own platform, but when that decentralized platform is a part of a wider internet surfing experience, in a way that the walled-garden social intra-nets can not be.

Nobody wants to run their own servers. Period. You absolutely, positively will not get widespread adoption of any process that involves Docker, DNS, or anything that involves more than a single click. Nobody wants to handle backups, upgrades, or move hardware.

You might, maybe, be able to get a federated system, whereby people could sign up for major providers that all talk to each other. And that could include being your own provider. For most people, that would be no harder than signing up for Facebook.

Now you just need to give them a reason, and no, nebulous "privacy" isn't going to cut it. Nor is "portability". If anything, the easier it is to move your data around, the more it's going to seem like picking one provider over another is a waste of their time.

The whole point of social networking is networking. People want to connect. "Private social networking" is a contradiction in terms. There's little motivation for people to take control of their social networking, and they're happy to put up with the mediocre privacy controls because the vast majority of the time they want to be more public, not less.

Right now all that ops stuff is a pain in the ass, and most everyone does it differently, and when you do do it, you get something that resembles a walled-garden social intra-net platform, but decentralized.

I strongly believe it can be cool to run systems again. And that we can be good at it. We are only just emerging operational paradigms that span multiple concerns, where-as historically the panolopy of concerns each needed their own daemon, their own configuration, their own tending, and there was no common operational paradigm. That, finally, is changing.

Still more fantasy than real but I do think we are starting to pave some of the cowpaths, & build a realistic, practical, well-trod path for running your own private cloud.

What remains is to make that fun. Pretend the ops box is ticked, what now? NextCloud is like the most advanced all-in-one suite you can do for yourself, and it's pretty cool, but it's still kind of this big monolithic jumble of PHP features. It's it's own point of of presence on the web, containing a lot of things. But I don't think it's enough.

ActivityPub & federation presents some good interesting grounds we can begin to inter-network on. But it's still just a start. Ultimately I think a lot of computing itself needs to be re-imagined to be networked-by-default, with good underlying general paradigms to build specific systems atop. We have a lot of exploration to go before we start to understand what that looks like. But, with good & sensible paths to home clouds (even just a raspberry pi), I think starting to make our personal computing also networked computing, & then internetworked computing, is what will topple the remote, far-off walled gardens that we commute to to compute on today.

I might be looking at this at a wrong angle, but why not just have a blog where you post your stories and share your pictures?

People could still "like" your photos and comment on them. Make URLs shareable and then they can email or what ever to their other friends to see if they feel like it.

I know, too much friction. But if they really cared about the content I would guess your friends would hop into the opportunity.

Because there isn't a business model for decentralized social networking.

Exactly. If data is hosted on personal devices, it’s not available all the time at high-bandwidth. If data is hosted on a dedicated server it costs money and time to maintain. That means users will need to pay a fee, or see ads. Open platforms are a means to circumvent ads, so that leaves paying a fee. But it seems users prefer free with ads, so those businesses will attract more users and capital.

The one thing that could change this is a major change in internet connectivity, such that it’s feasible to host on your own devices. As personal connectivity and bandwidth improves, P2P becomes more viable.

With P2P you still have to have a device that's always on, which most people don't already have, so that costs money. I actually think a VPS is cheaper than a home server. But either way it costs money and people don't want to spend it.

And software development costs are another thing that needs to be paid for but can't.

Don't forget one major thing, think about how many people on social media are not old enough to have a debit card, PayPal account, or similar.

You effectively instantly exclude them.

In significant parts of the world, including the US, many ISPs are turning to CGNAT also. Static IPs and being able to open a port is an addon.

I am working a technology solution for this. The idea is to invert how the web works so that each person/device is a server responding to clients. The result is inherent controls on privacy and superior technology, such that a web browser becomes your OS desktop available from multiple computers across the internet with a security model in place that allows sharing with friends.

We do, they are in our phones. Everyone's contacts list is their decentralized SN.

The concept of the contact list is primitive though. The open source community has neglected to develop tech for easily reusing it, and neither apple nor google have any interest in making it interoperable with other services.

I believe we're close. I believe it's only a year or two away.

Reply on news.ycombinator.com