|
|
|
|
|
|
by leppr
2491 days ago
|
|
|
The article, if seen as criticism of decentralization in general, would be a big straw-man. But the actual thesis of the author is that you can't decentralize "everything" (the HN title has the straw man though). Even in successful decentralized networks, some aspects (usually governance) stay or become centralized. The author easily succeeds in making that point because everyone already agrees with it. Whether that's a problem worth discarding the whole movement towards decentralization is mostly a matter of preconceived opinions. |
|
|
Imagine if someone said, "competitive markets aren't always possible, and over-saturated markets are usually bad for consumers anyway. Therefore, people who advocate for more competition with ISPs or who try to block mergers are just being unrealistic. A competitive market just isn't possible at scale."
Yet plenty of people will argue that Git's goals around decentralization ultimately failed just because Github is convenient and widely used. Or that because social media platforms often naturally trend towards centralization, that efforts like Mastadon are bad for consumers and are a waste of time.
That many real-world decentralized systems have some centralized parts shouldn't be seen as unusual, and it isn't some kind of killing blow to people who advocate for more decentralization. This is the world working as expected.
If the only two states of government possible were anarchy and totalitarianism, then totalitarianism might be the better choice. But those aren't the only two governments possible. Similarly, decentralization is not a binary state. You can use centralization and decentralization in tandem to cover for the flaws and gaps of functionality that both strategies have.