Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Decentralized Social Networks Won't Work

wired.com

Post reply on HN
Print view View original

41–50 of 143 posts

Re: Decentralized Social Networks Won't Work

#41

Post by marian0_ »

panic wrote: post #35
lxrbst wrote: post #30

Earlier quoted context omitted.

Mastodon is not successful, at all. Apparently content that was not accepted on twitter, such as lolicon, has found a new home on Mastodon. The average user would never bother to migrate to decentralized services for any other reason then necessity, lack of choice, or to escape rules/laws.

Sure, but centralized services have the same user acquisition problem. Is there something that makes decentralized services uniquely bad at getting users to switch?

Decentralised services always come out years later centralised services are launched and their only raison d'etre is not having the rules of those services.

Re: Decentralized Social Networks Won't Work

#42

Post by jswizzy »

Just about everyone I know on Facebook would jump ship if they could. Not a day goes by that I don't notice someone complaining or leaving over anger at Google and Facebook supporting invading their privacy, promoting extremist politics, and censorship of conservative and liberal ideas.

Re: Decentralized Social Networks Won't Work

#43

Post by infinity0 »

I've never seen an actual well-thought out technical argument on why decentralised social networks can't work. That's because there isn't one. People who are familiar with the technical issues surrounding this are humble and don't make bogus over-generalised statements like "X can't work [ever, everywhere, in all contexts]".

This article is not a well-thought out argument either. It's a bunch of random irrelevant anecdotes with some hand-waving tacked onto the end that the writers though was good to write down here because the same words impressed their friends.

The main barriers here actually are:

- over-advertising and self-promotion of non-advanced stuff like Diaspora etc that doesn't actually solve any inherently hard problems in this area, wasting people's time especially that of interested people that want to enter the field

- lack of funding on solving the hard problems in this field

- lack of structured educational material to get good new engineers onto these topics

When people say this stuff "can't work", they really have no realistic concept of the amount of actual resources it takes to bring a great idea into reality, and how little resources are being (and have been) fed into this topic that they so over-confidently dismiss "can't work".

Re: Decentralized Social Networks Won't Work

#44

Post by cetra3 »

Kiro wrote: post #18
cetra3 wrote: post #14

Earlier quoted context omitted.

It's decentralized but needs: - My Email Address - My Phone Number Not really the sort of decentralized I was hoping for..

You can spin up a node or use a wallet to create an account directly on the Steem blockchain instead of doing it through the Steemit webpage.

If that's an option I haven't found it from a quick glance on the website. I just followed what was there without looking too deeply into it. Maybe if that's something you can do, there should be a clear process written up somewhere

Re: Decentralized Social Networks Won't Work

#45

Post by leoc »

edhelas wrote: post #31

Decentralized social networks aren't working for now because they reinvent their own protocol each time which creates fragmentation and incompatibilities between the platforms. I think that decentralized social network projects should relies on existing standards and help with their development. That's why I'm working on Movim ( https://movim.eu/ ) for several years now, which is fully based on XMPP and is de facto c…

> Decentralized social networks aren't working for now because they reinvent their own protocol each time which creates fragmentation and incompatibilities between the platforms.

Recall that all computing problems are solved by adding more abstraction ( ;) ) and not all problems yield to a direct attack with the current tools. Maybe the point of maximum leverage here is actually the problem that interoperation between different protocols is too difficult?

Re: Decentralized Social Networks Won't Work

#46

Post by zAy0LfpBZLC8mAC »

marian0_ wrote: post #41
panic wrote: post #35

Earlier quoted context omitted.

Sure, but centralized services have the same user acquisition problem. Is there something that makes decentralized services uniquely bad at getting users to switch?

Decentralised services always come out years later centralised services are launched and their only raison d'etre is not having the rules of those services.

Yeah, I still remember when the web was only a product of Web Inc., and how long it took for independent webserver implementations to take off.

Or, before then, do you still remember when email was this walled garden where everyone had an account with Electronic Messaging Solutions Inc.? It took decades until sendmail, exim, qmail, postfix, and so on came along, largely thanks to Google's help in breaking open the former monopoly, and people could finally host their own email servers!

Re: Decentralized Social Networks Won't Work

#47

Post by zemnl »

albertgoeswoof wrote: post #23
panic wrote: post #13

Earlier quoted context omitted.

Where would you find these metrics? Looking at the federated timeline, I see posts appearing at a pretty fast rate (more than one per second), though most of them are in Japanese.

go and ask everyone you meet in person today if they've heard of mastodon

I bet that few of them will reply that it is an extinct animal, some others that it is a cool band, and most of them that they have never heard of such a thing.

Re: Decentralized Social Networks Won't Work

#48

Post by marian0_ »

zAy0LfpBZLC8mAC wrote: post #46
marian0_ wrote: post #41

Earlier quoted context omitted.

Decentralised services always come out years later centralised services are launched and their only raison d'etre is not having the rules of those services.

Yeah, I still remember when the web was only a product of Web Inc., and how long it took for independent webserver implementations to take off. Or, before then, do you still remember when email was this walled garden where everyone had an account with Electronic Messaging Solutions Inc.? It took decades until sendmail, exim, qmail, postfix, and so on came along, largely thanks to Google's help in breaking open the fo…

I mean the decentralised services of today, no need to write a snark response. THink XMPP, or Mastodon.

Re: Decentralized Social Networks Won't Work

#49

Post by edhelas »

leoc wrote: post #45
edhelas wrote: post #31

Decentralized social networks aren't working for now because they reinvent their own protocol each time which creates fragmentation and incompatibilities between the platforms. I think that decentralized social network projects should relies on existing standards and help with their development. That's why I'm working on Movim ( https://movim.eu/ ) for several years now, which is fully based on XMPP and is de facto c…

> Decentralized social networks aren't working for now because they reinvent their own protocol each time which creates fragmentation and incompatibilities between the platforms. Recall that all computing problems are solved by adding more abstraction ( ;) ) and not all problems yield to a direct attack with the current tools. Maybe the point of maximum leverage here is actually the problem that interoperation betwee…

The interoperability between the protocols is indeed difficult. But pushing a standard protocol is, I think, the best way to promote this compatibility.

Imagine if we didn't have a standard protocol for emails (SMTP/IMAP/POP) but instead several protocols built using JSON on top of HTTP. We would have to take most of our time to work on broken transport libraries and fix incompatibility issues between all those solutions. Hopefully we have a standard (even if it's not the best one) and we can "just" build on top using the many existing libraries and projects out there.

I think that such "social standard" should be build on Internet protocols (TCP/TLS/DNS/IP…) and not Web only (HTTP), even if it's "easier". Also instead of re-starting from scratch I prefer to improve something existing like XMPP that have flaws but is already an industry standard with a strong community, many libraries/clients and servers deployed all across the world.

Re: Decentralized Social Networks Won't Work

#50

Post by madnerd »

lxrbst wrote: post #30
panic wrote: post #7

None of the reasons they give -- difficulty of user acquisition, conflict between security and convenience, curation, or economies of scale -- are fundamental obstacles to making a decentralized social network. Mastodon has found some success despite facing the same problems.

Mastodon is not successful, at all. Apparently content that was not accepted on twitter, such as lolicon, has found a new home on Mastodon. The average user would never bother to migrate to decentralized services for any other reason then necessity, lack of choice, or to escape rules/laws.

Personally I like using Mastodon. There is an open source community (mostly french) that post interesting things.

A good social network shouldn't be about who has the most users, but who has the users you want to interact with.

Mastodon will probably never be the average user social network, but it works pretty well for people with specific interest, or marginalized people that doesn't want to interact with hateful people.

As for the lolicon content, it just doesn't exists except on specific instance, which are blocked on mine.

Post reply on HN