>
This leads to a fracturing of consensus.But is it necessarily a bad thing? I think whether the outcomes are good or bad is going to be highly subjective and debated ad infinitum. I, for one, welcome our new Internet overlord.
There are two interesting articles on this topic, I strongly recommend everyone to read them. Gwern's article has interesting observations about politics, subculture, Unicode, and programming languages. David Perell examines the effect of the Internet in a bigger framework of world's politics, education, and commerce.
* The Melancholy of Subculture Society, by gwern
https://www.gwern.net/The-Melancholy-of-Subculture-Society
* What the Hell is Going On?, by David Perell (has a playful foreword, if it annoys you, ignore it and continue reading).
https://www.perell.com/blog/what-the-hell-is-going-on
--
My perspective of the issue is, under the previous centralized system, the entire nation is subject to the identical propaganda. The effect of the system can already be seen from the extensive propaganda during Mexican–American War of the 1840s - nearly 200 years ago.
> You furnish the pictures, and I'll furnish the war.
And during the World War II, and later the Cold War, the power of state propaganda reached its height, and I'd say this system is responsible for the massive thought manipulation and greatest violence in the human history. But there are good sides as well - information authority, good writings, and strong consensus.
> We have had Edward R. Murrow talking straight at us and gripping the whole nation's attention, we have had Thomas Paine standing in the street, telling us common sense that changes our lives, we've had shots heard round the world, revelations shocking the whole nation at once. - said Shii, a early and influential 4chan moderator, also a major contributor of English Wikipedia.
Then came the Internet revolution. Sure, under this system, the absolute notion of truth is deteriorated. There would be Holocaust deniers, anti-vaccine activists, flat-Earthers, moon-landing truthers, foreign propagandists, wild populism, among other groups - heck, even a Facebook poster can start an absurd national hate movement (see my comment at https://news.ycombinator.com/item?id=20012564). But simultaneously, it's a feature as well, the same system also gave voice and self-determination to those who didn't have - although not in a totally egalitarianist manner, but at least a positive contribution, and brought democratization of communication, which ended, or greatly reduced the power of centralized propaganda.
> Getting your views on not just politics, but also physics, biology, economics and who-to-burn-at-the-stake from your local religious official. This was the default for most of human history. I’d much rather have flat earthers than the Spanish inquisition, thank you very much. - An author's response to the concern of Fake News and disinformation, who is working on replicating GPT-2.
It disintegrates physical and national barriers and identities. The horror of nationalistic violence had seen it better days, e.g. "the country I live in now is the best country in the world for people like me; I would be terribly unhappy if I was exiled", and now, it's something like, "‘Why, what’s so special about the USA? It’s not particularly economically or politically free, it’s not the only civilized English-speaking country, it’s not the wealthiest..."
And the subcultures rule. and in a sense, liberates individuals by giving one the option to opt-out. I, for one, welcome our new Internet overlord. As gwern said,
> If I’m a programmer, I don’t need to be competing with 7 billion people, and the few hundred billionaires, for self-esteem. I can just consider the computing community. Better yet, I might only have to consider the functional programming community, or perhaps just the Haskell programming community. Or to take another example: if I decide to commit to the English Wikipedia subculture, as it were, instead of American culture, I am no longer mentally dealing with 300 million competitors and threats; I am dealing with just a few thousand. It is a more manageable tribe. It’s closer to the Dunbar number, which still applies online. Even if I’m on the bottom of the Wikipedia heap, that’s fine. As long as I know where I am! I don’t have to be a rich elite to be happy; a master craftsman is content, and “a cat may look at a king”.
> Leaving a culture, and joining a subculture, is a way for the monkey mind to cope with the modern world.
But from another perspective, it's also harmful in some ways. It was once possible to leave or escape from one's tribe by physically moving, but now it's an iron cage that nearly impossible to escape. It may intensity the world's geopolitical conflict, as everything has been balkanized. On the other hand, perhaps the society can be better off by learning to operate without a strong consensus.
--
To start fresh and frame the topic differently, we can use decentralized systems as an analogy, the traditional mass media is like the Certificate Authority model. It guarantees absolute truth on whether a public key is real by providing a central consensus, but on the other hand, when things go wrong, the entire public key infrastructure is vulnerable to rogue actors, especially state actors. In comparison, we have the PGP web-of-trust model - although the actual implementation turned out to be a total failure due to legacy code and design issues, but the ideas remain valid - that iut derives the trust not from an authority, but from the collective opinion of a group of people in a community you know. The good thing is that anyone is free to make one's own judgements, and the system is resistant from a central rogue actor, the bad thing is that there is only localized consensus, not centralized consensus, you cannot surf the web using this model.
Recently, there was a chaotic argument online about on how to scale a Bitcoin-like decentralized, P2P protocol. Two authors purposed the DCS theorem an an analogy of the CAP theorem. It's not really a theorem or a research paper, I'd say it's just a personal opinion, but somewhat interesting [0]. Basically it says, a decentralized system cannot simultaneously satisfies the three properties: (1) Decentralization, (2) Consensus, and (3) Scale. A traditional bank is C & S: having a global consensus and no scaling difficulty, but it's centralized. The original Bitcoin achieves D & C: decentralization and global consensus. Everyone knows every transaction - by running a blockchain, but it must come with an extremely high cost. On the other hand, Layer-2 solutions bypasses the blockchain, thus it achieves D & S: decentralization and scale, by avoiding to broadcast the transaction to the blockchain and use P2P communication instead, thus abolish global consensus. Using the same line of thinking, I think this conclusion can be applied to a lot of other systems, not only blockchains, for example, the Certificate Authority is C & S, and the web-of-trust is D & S. The mass media is C & S, the Internet is D & S. Doing D & C requires everyone knows everything, which is simply not the Internet.
Anyway, what I'm trying to say is that the pros and cons are inherent in each paradigm, you cannot both have your cake and eat it, and whether the outcomes are good or bad is going to be highly subjective.
And back to the issue of biases of mass media, one possible idea is to impose some limitations on the freedom of the press to the mass media as a counterweight to the freedom of speech: In the age when everyone can say everything online, perhaps the mass media establishment should be forced to engage in a journalism with a higher standard, serving as a reliable reference source.
[0] https://arxiv.org/abs/1801.04335