Public Internet Services
I think centralizing public social services is good. They should be run like small businesses, bulletin board services, or amateur radio clubs; fostering communities around regionalized aspects.
E-mail worked as a decentralized service because the social graph of E-mail mimicked decentralization, and there was little benefit to grouping power. Centralizing E-mail can reduce spam, as Google Gmail has shown. But conflict resolution has not been an issue for E-mail. I don't expect E-mail to act like a closed service.
A public-facing service has more relationship between the user and the service's public presence, and more conflict, and more safety to users from centralizing. People attach more of themselves to their public profile on a social service than they do to E-mail.
There are many problems with focusing power in centralized services.
- Data loss
- Private data stolen
- Mob dynamics
- Large-scale abuse
- Corrupt governance
I argue that they are largely solved by having smaller, more relational and contextual centralized services.
These problems can absolutely happen on a small scale. But of course the damage is localized, and by building services smaller and through relationships, the likelihood is reduced.
Localized Centrality as Decentralization
Modern decentralization is administered much like localized centrality; services centralized but with closer relation to the users. I have two problems in particular with modern decentralized social media.
First, I think it can still suffer the same problems as gargantuan services like Facebook and small services. It often has the worst of both worlds — no protections and still groupthink and lack of diversity-as-security.
Second, grouping the kind of small, relational community I'm advocating for around a locked-in, ubiquitous framework or protocol negatively affects both the quality of administration and the health of the group.
Ubiquity in administration and protocol can simplify it, make it cheap, and more accessible. But I think this cheapness is a bug, not a feature. Smaller, localized software is becoming cheaper because of improvements all around development, so the technical gains in standardizing social media protocols I think is not worth the losses.
It tokenizes the administration. Needs can't be met. The admins themselves can become irrelevant. In a vacuum, they could be appreciated for themselves, but like E-mail, almost no admin will really be appreciated here. Without a relationship with the admin, I believe the service is more vulnerable to the above centralized issues.
Compare again the admin of any online service to that of E-mail. More than likely, you use Gmail for your E-mail hosting, and you have zero leeway with Google, zero relationship with your admin, no way to contact a person or have anything handled almost at all. You do have good uptime and thousands of articles around the Internet for self-help, and you are safe from Google's bullying of minor hosting providers.
More than likely you use Facebook or Twitter, and the same is true for your experience there.
(I do not exclusively mean a literal conversational relationship with a specific administrator of a service. But simply being on a smaller service, you are closer to the admins. This is a broad meaning.)
A challenge the Internet has brought is that context and relationships are unnatural here. Strengths though include that grouping and supporting minorities is easier than ever, and services cost little to build and run. I think context and relationships should be pursued, while playing to the Internet's strengths.
In principle, I am invoking
the medium is the message (Neil Postman) and
context is king (English 101).
In theory, small centralized services are vulnerable to dying off, like decentralized service nodes. I think this is a benefit in many ways. Unhealthy services will die, opening space for improvement. And people's data can be lost.
Major centralized services would be purchased for their users and their data, and may never die.
This is how honey bee colonies survive and how mom and pop stores would exist in America. Where I liken decentralized service to fast food franchises, beholden to the franchise for what they can serve and where they can exist. And centralized services to Walmart.
The Technology Doesn't Matter
All said, what is the difference between running your own closed ActivityPub instance, and using a custom-built API? Shouldn't the former be similar to running your own BBS or forum? It seems like an online service is going to be a lot more influenced by the users than by the technology stack.
Consensus around protocols means the arena is maturing, and approaching lock-in. So it should also be becoming accessible with multiple implementations and good tooling.
At this point, I am practically arguing against E-mail using a shared protocol. Isn't that ridiculous?
E-mail is not public. Social media is public and inherently faster-moving. It is not the same as E-mail. The power and social dynamics are very different — arguably even between how different social networks function.
This is a good time for me to admit that I don't respect social media. It has organized rebellions in Tunisia and Hong Kong, and caused people to lose their jobs, friends, and lives. It is manipulable from Russia to Wall Street; used to experiment and troll; and can be collected and follow you to the ends of the Earth. I don't want it to be an irreplaceable facet of my identity, and I think it's wise not to use it for anything important. Unsafe, unhealthy, unwise.
If you are willing to give as little significance to your Online life as I do, then you don't have bones in the game and you can afford to play with social media. But I still care about the medium I'm working in.