This just isn't true. Yes, running your own blockchain node is infeasible, but a blockchain-based service with hosted nodes is better than a federated service.<p>Proof-of-work creates an illusion of dependence on blockchain miners. Proof-of-stake makes it clear that blockchains are run by a corporation that wants to maximize transaction fee revenue. Transaction validators have to do the best job of maintaining a unified historical record of all transactions for people to keep giving them business. Otherwise, users will choose different validators and carry on without them.<p>Blockchains are fancy spools of receipt paper. If the printer gets jammed, tear off the stuff you care about and staple it to a new spool. The printer doesn't control the people. The people employ the printer.<p>With that said, it's still unclear whether blockchains are the right choice for publishing. A decentralized Twitter doesn't need consensus on a historical record. An IPFS-based Twitter is probably a better way to start, then anchor the parts that need consensus on a blockchain, like usernames and timestamps.
<i>"The alternative on which we can build a great new Internet era is projects like GNU social, Friendica, Hubzilla, Diaspora, and efforts to design common protocols like ActivityPub, not on the blockchain."</i><p>All the new federated social systems have failed, replaced by commercial, centralized systems. (IRC and email live on. Google thinks it owns Usenet, rebranded as "Google Groups", and a big chunk of email never passes through SMTP.) Why? We had a good discussion of this recently by the creator of WhatsApp(?). Centralized systems can be upgraded and enhanced. Federated systems require massive cooperation to change. And federated systems don't generate much revenue.
i'm not a fan of "blockchain for everything", but stating that a distributed system is a danger for distribution, and favoring a decentralized, federated approach instead sounds like plainly wrong.
Imagine if BitTorrent, or Freenet/IPFS/GNUNet/etc. used a blockchain. They'd be completely infeasible for anyone except Google: you'd need petabytes of storage on your computer just to download a single movie (or a Linux ISO, or whatever).<p>But they don't do that. These systems are actually distributed, and that's why they work.
"The" blockchain = one subset of potential uses of Ethereum ?<p>Did someone take some time to tell author that internet is not www, and that www is not maintaining datacenters ?<p>I'm surprised that some people find this article informative.
Threat is poor word choice. If blockchains don't provide good distributed service, then they simply won't be used, in which case they aren't much a threat. Maybe author should have use the thesis that blockchains aren't the end-all-be-all of the distributed internet.
The TCP/IP Internet is also an engine of centralization. By limiting hop counts, and distributing addresses hierarchically, we end up with just a few giant players in each industry.