Hacker News Reader

Comment by syllable_studio | Hacker News Reader

I CAN'T WAIT for decentralized infrastructure to finally become robust enough that clones of instagram and facebook can all share a network -- they'll all become portals to the same social graph data. When that happens, there will no longer be any reason to put up with this abuse of privacy garbage.

View on HN

Never going to happen. Ever. Decentralized means it's inevitably more cumbersome to use. If you make it easier you're taking shortcuts somewhere, especially if you want a platform that's resistant to censorship and is supposed to provide anonymity. And if you actually do succeed in this you'll inevitably end up with a platform where highly immoral content is just a wrong click away.

It'll be this cycle of a new platform being fresh and trendy, then they become bigger, care about being taken seriously by traditional media, for getting ad deals with the big players, then a new platform pops up, cycle repeats.

> Decentralized means it's inevitably more cumbersome to use.

Which will lead to a lot of users being hosted on a single instance which provides convenient access, which completes the loop to centralized. See gmail.

> And if you actually do succeed in this you'll inevitably end up with a platform where highly immoral content is just a wrong click away.

I don't think that is necessarily a problem for most users.

> It'll be this cycle of a new platform being fresh and trendy, then they become bigger, care about being taken seriously by traditional media, for getting ad deals with the big players, then a new platform pops up, cycle repeats.

This is, though. See Reddit :(

How many people (outside of the HN circle) actually use Gmail for personal email now? It's mostly been killed by Facebook, WhatsApp or their alternatives.

And most of these alternatives still need an E-Mail to sign up, which loops back to gmail. Statista puts the number of its users at 1.5 billion in 2018[0]. There are some alternatives, yes, but if you ever tried to run your own mailserver you'll quickly find out that gmail users not receiving your mail is very much a problem.

[0] https://www.statista.com/statistics/432390/active-gmail-user...

Fastmail? Protonmail? Others? It's not as though the only options are gmail and running your own mail server. They aren't free, but neither is your time to run a mail server and gmail is only gratis, not libre.

I've never said its the only option. I use a rather unknown hoster myself. But the majority of the users are with very few companies; for the US and EU gmail probably dominates the private user market.

Or, let's say it this way: If someone selfhosting his mail server doesn't receive your mail, it's his problem. If gmail users don't receive your mail, it's your problem. And that's not a sign of decentralisation.

I think that, given the change we expect to see in human-computer interaction over the next few decades, you can't say never. The majority of the obstacles you outlined are not fundamental but rather implementation problems that can be solved by motivated parties with time and effort. Of course, saying this doesn't mean someone will solve them, just that someone can.

As an example, the censorship and immoral content dilemma is easily solved by letting the user "censor" their own feeds of content they themselves find objectionable. This is already trivially done via block lists and word filters, in the future thanks to ML there can be a much more sophisticated level of control exposed including fuzzy sentiment metrics. I wouldn't be surprised if tools like this are already being built. The key idea here being that the control is with the user rather than with some 3rd-party whom may or may not be taking the users own preferences into account.

I would instead say that the problem is political (or marketing) rather than technical: Facebook and Google were set up by an undergrad in his bedroom and three PhD students respectively. Building alternatives is easy, getting people to use them will be extremely difficult.

Building alternatives that people will use is itself a technical problem, just not one that's as crisp as the algorithmic issues related to distributed computing.

Well, the Fediverse made a start and it is only at the beginning of its evolution. Not billions of users, no, but ever more interconnected apps with features similar to the big platform (but void of algorithmic feeds, ads and many of the dark patterns)

>I CAN'T WAIT for decentralized infrastructure to finally become robust enough that clones of instagram and facebook can all share a network

What? We HAVE this TODAY. It's called the INTERNET. How would another internet solve the problems that thrived in this version.

If these will be as popular, we'll still be left with mass disinformation campaigns and culture wars though.

Just a general response to some of the skeptics here (I hear your points, I just think we're still early in the game). I do believe this will happen. I believe it's inevitable, it will just take time. The main advantage over current networks+apps is that it breaks the barrier to entry for new competitors. Mastodon etc are still early newcomers who still have to break that barrier. But once there are many of them playing in the same water, they can gradually build a competing network - slowly at first, then they'll win over the entire new generation. Whatever the next "tik-tok" is could be decentralized, and then the flood gates open. Advertising and privacy issues will still exist, but it will be competitive. There will no longer be a monopoly gate keeper who that sets the rules. Lots of details to be worked out for sure. We need user interfaces to improve and get simplified. I agree that right now federated platforms like mastodon are still too cumbersome for users. But I believe all of this will improve and this will work eventually. I think it's coming this decade.

Mastodon and Pixelfed already exist and are very functional, the issue is that all instances of the above are run with a hobby-grade SLA and mixed moderation policies and users that do not understand federated services. Even domains on emails get defaulted to gmail.com when in doubt now.

all become portals to the same social graph data...privacy

How will we have privacy when the graph is open?

same as now, encryption and permissions. We'll just have more options and less lock-in.

Reply on news.ycombinator.com