Decentralizing Media

Redistribute power from centralized legacy media and social media corporations to a million hubs and a billion spokes.

• 34 min read
Decentralizing Media

As the saying goes, freedom of the press belongs to those who have one. An old saying, but newly relevant, due to real news that was dismissed as irrelevant.

Li Wenliang was a doctor who tried to warn people of the mysterious new pneumonia that had arisen in Wuhan, China. He wasn't a prestigious journalist or a tech entrepreneur. He was just a citizen, posting in a social messaging app, trying to inform his fellow citizens of the story of the year that was absent from the news.

His warnings fell on deaf ears. He was denounced for rumor-mongering and forced to publicly apologize while institutional media downplayed the outbreak’s severity by equating it to common pneumonia. And then, tragically, he died from the illness he tried to warn his country and the world about.

The unholy fusion of state regulation of social media, untrustworthy institutions, and social media de-platforming that so many have been warning about for so long may literally have put the whole world at risk by silencing a single voice.

And this is not a one-off situation. For example, Caijing is one of the most reputable outlets in China. It published an article that strongly indicated that the Chinese government was downplaying the scale of the crisis. As people were viewing the article, it suddenly 404'd in real-time! Those around the world trying to get a sense of the scale of a crisis that might end up going pandemic were suddenly thrown into the dark by the actions of some nameless functionary pressing a key somewhere.

The act of deleting information was itself informative: the official news was actually fake, the actual news was officially censored. In a moment like that, one feels viscerally how important it is to decentralize the power of large tech and media institutions.

To paraphrase Andy Warhol, in our future, anyone can be crucial for 15 minutes. The people quoted in that Caijing article, the Chinese doctor whose voice was silenced online…had these people been able to get the word out to more folks, we might not be in the middle of what the WHO has declared a global public health emergency.

Trust because you can verify

If freedom of the press belongs to those who have one, we need to redouble our efforts towards turning people from mere users of social media and readers of institutional media into first-party operators of social media platforms, first-class publishers of their own media content, and first-class fact-checkers of everyone’s content.

To not just hope that Twitter shows your post to your audience, but to reach them through uncensorable, encrypted communications channels in the event you need to guarantee receipt. To not just post occasionally on social media, but to have the production values and monetization capability of an institutional publisher when you have something important to say. And to not just commit acts of journalism like a traditional reporter, but to use decentralized verification to establish institutional-grade credibility such that you are taken as seriously when publishing something true.

That last sentence deserves elaboration. How could a citizen possibly figure out when a random blogger should be trusted over the official organs of the Chinese government, or an American outlet like 60 Minutes for that matter? When could we practically substitute decentralized verification for institutional reputation? Don’t we need to just trust centralized institutions and hope they don’t abuse that trust?

This is one of the most important questions of the coming decade, but there are a growing number of technological tools that allow the crowd to confirm or refute the claims of an institutional authority in at least some cases. For example:

Shadow statistics

Returning to the example of the virus, a Chinese investor named Dovey Wan posted a calculation that showed that the actual percentage of infected people in Wuhan on Feb 3 may be around 1%, rather than the official 0.1% reported at that time.

The Billion Prices Project at MIT has done something similar with inflation numbers by scraping the whole internet for prices to compare them with government statistics. Sometimes they confirm the official numbers, and sometimes they refute them.

Digital signatures

News articles over the years from outlets like Newsweek and Wired have purported to identify Satoshi Nakamoto, Bitcoin’s creator. All of them have been shot down by the lack of a confirmed digital signature that proves that the purported Satoshi actually has control over the earliest mined Bitcoin.

Device logs

When Elon Musk had a dispute with the New York Times in 2013 over whether a Tesla actually had car troubles, he was able to post the device log on the internet to show that at least some aspects of the story were false.

Conversely, Apple Watch and Fitbit logs are in some cases helping to confirm the claims of other institutional authorities, by helping police identify and prosecute criminals.

Timestamps and digital attribution

President Macron of France shared a photo of a fire in Brazil that was retweeted 40,000 times but turned out to be fake. The reason is that it could be attributed to a photojournalist who died in 2003, meaning that it was at least 16 years old. The combination of reverse image search to trace the photograph and biographical timestamping to date it invalidated the widely-circulated-but-false institutional claim (which is still up!).

As a different example, the internet court of Hangzhou was able to use content hashed to the blockchain to establish priority in a patent dispute. The defendant could show they had not infringed the patent by pointing to an earlier timestamp. Here, institutional authority itself used decentralized verification as a complement to its pre-existing process.

Censorship resistance

Finally, again returning to the virus crisis, we only know about the Caijing article because it was archived, translated, and shared on Western social media even though Chinese censors took it down. A site like is one way of accomplishing censorship resistance but even more powerful decentralized web technologies are coming that will let us create truly perma-links.

These techniques are ad-hoc right now, but we start to get a picture of how individuals can confirm or refute the claims of institutional authorities with decentralized technologies.

Would this have helped Li Wenliang, though? It certainly wouldn’t have been sufficient. But it might have helped on the margins. For example, we know that there are popular wearables that give everything from ECG to heart rate to body temperature, the last of which is an input for COVID-19 diagnosis. And we know that there are public dashboards like the Strava global heatmap where people share and aggregate their wearable data.

Had there been some kind of censorship-resistant public dashboard where everyone could see the rising number of sick people in a district as revealed by privacy-preserving aggregation of their health data, we’d be using individual device logs to generate collective shadow statistics. Perhaps that indisputable data would have helped more people believe Li Wenliang.

But he would still have needed the distribution to reach people in the first place, and the reputation for those people to judge him worth listening to.

Both reputation and distribution are centralized chokepoints

Li Wenliang had the facts, but he was hobbled by a lack of both reputation and distribution. He was not taken as seriously as the institutional journalists at state-controlled Xinhua News who downplayed the crisis as “controllable” till it was too late. And he was not able to continue speaking to his audience of medical school colleagues on the centralized WeChat. A switch was flipped, and he was identified and then silenced.

So in the case of China we see the alternative timeline where centralized power has consolidated control, where decentralization does not succeed.

The free world hasn't yet gone down that dark path. But we're starting to, due to worrying levels of technological and cultural centralization.

The technological centralization of China's WeChat and Weibo is obviously similar to the centralized servers of Facebook and Twitter. But less well appreciated is that the cultural centralization of China's institutional media is operationally similar to the cultural centralization of America's institutional media.

Yes, China has a state-controlled media and the US does not. But on a daily basis, the Chinese people have no way to know when their state-controlled institutional media outlets are less credible than a doctor with local knowledge posting on WeChat, in much the same way (say) the American people have no way to know when the New York Times, Wall Street Journal, and Washington Post – or government pronouncements for that matter – are less credible than an engineer posting on a mailing list.

Organizationally, the countries are different. The US actually does have state control of public education, government statistics, and much academic research – but the press still stands above the state in the West rather than vice versa as it is in China.

Operationally, however, they are increasingly similar. In both China and the US, there is a small group of technologically centralized institutions (WeChat, Weibo, Google, Facebook, Twitter) that can control what is seen and culturally centralized institutions (Xinhua, People's Daily, NYT, WSJ, WaPo) that determine what is accepted as true.

This centralization may be acceptable if we trust the centralized institutions. But it does not work when these institutions violate our trust, as they did in China with Li Wenliang and have done many times before in the West.

So we need to figure out some way to complement trust in institutions with universally accessible technology, such that “trusted” institutions can be corrected when wrong and the average citizen's voice can be elevated when right.

Because distribution alone is only one piece. Both distribution and reputation are centralized chokepoints. Technology can help us with both, but we’ll need cultural changes as well.

Decentralizing these things is a tall order, both technologically and culturally, but as we will see there are now some powerful ideas that may help us do it. And we must at least try. Because the public health of the entire world was endangered by the fact that Li Wenliang was just a user of social media and reader of institutional media, by the fact that he had no means to establish the institutional credibility of a publisher or the robust distribution of a platform.

A standing media is as undesirable as a standing military

Stepping back for a second, there is an important parallel between a standing military and a standing media.

The founders of the United States were against a standing military, a military comprised of full-time rather than citizen soldiers. Why? Because they feared a praetorian class that was isolated from the populace at large, that considered itself separate and better, that had the power to attack the population they were charged with defending, that could draw the country into unnecessary conflicts.

The historical (and possible future) alternative was to draw the military from the population at large, to have part-time citizen soldiers rather than a full-time standing military.

Now consider this quote from New York Mag's 2016 survey of journalists:

Journalism’s biggest blind spot in its coverage is:
Groupthink. We draw from a limited pool of people who generally have a similar background and class. They simply are unable to see the perspective of people who are not like them, and tend to drive out those who don't "fit in."

Similar criticisms have been made of technologist demographics.

If neither journalists nor technologists look like America at large, that means we have a standing media, just like a standing military, comprised of a group of people with collective control over our public narrative that are not demographically representative of the American people, let alone the world.

Here, the term standing media refers to the fact that these journalists and technologists are, for the most part, full-time employees of institutional media corporations and social media companies respectively.

And if we think about the journalist's focus on holding people accountable and the technologist's power to moderate social media, they really do both wield significant power. Indeed, both journalists and social moderation teams often see themselves as a type of police.

The analogy to a standing military is obvious. And then the question arises: if the alternative to a standing military was a citizen’s army, can we continue further along the path towards replacing (or at least augmenting) our standing media with citizen journalism?

The two kinds of centralized media corporations

Another term we can use besides standing media is centralized media. Let’s define two kinds of centralized media corporations:

  1. Technologically centralized social media and internet companies (Facebook, Twitter, Google, etc)
  2. Culturally centralized institutional media outlets (NYT, WSJ, WaPo, etc)

Despite all the digital ink spilled over "tech vs media", these two institutions are more similar than different in at least three respects: demographics, wealth, and centralized control.

First, on demographics. Much ink has been spilled about tech's diversity or lack thereof, and indeed much remains to be done. But media companies like the New York Times appear to be significantly whiter than companies like Facebook.

Second, on wealth. It is absolutely true that Larry Page, Mark Zuckerberg, Jack Dorsey, and company are very wealthy people. But Rupert Murdoch of the WSJ is also a billionaire, Arthur G Sulzberger of the NYT may be worth close to $100M, Kara Swisher of Recode has inherited wealth, and a large number of journalists have private school educations, trust funds, and/or well-to-do spouses:

Jeff Bezos of course is both a tech and media mogul, and the wealthiest man in the world, which illustrates the point well.

Now, we’d be remiss if we didn't mention that there's huge variation in both camps. The CEO of a multinational tech company is hardly the same as the founder of a tiny startup. And a famous journalist with an NYT column and a trust fund is not the same as a freelancer at a struggling local paper. This is true not just economically but politically as well: just like GDPR is advantaging giant tech companies over small tech companies, AB5 is advantaging full-time journalists over freelancers. Still, at the elite level when we think of “tech” vs “media”  it really is billyuhnaires vs multi-millyuhnaires, as Bernie would put it.

Third, in terms of centralized control. There have been many articles written about the dual-class stock of folks like Mark Zuckerberg, and how they would mean that people couldn’t fire his kid’s kids. But James Murdoch inherited his position from his father, as AG Sulzberger did from his. And like Zuckerberg,  both the Murdochs and Sulzbergers have dual-class stock. They have as much control over News Corp and the NYTCO as Zuckerberg does over Facebook.

So that's the dirty little secret: tech and media are actually very similar on key dimensions. Journalists are like VCs in that both are always scouting for new ideas and writing up theses on the future. Engineers are like writers who want to change the world by typing into a keyboard. And employees of media and tech companies alike really do work for something more than a paycheck.

This makes sense because tech is a cultural fork of the East Coast, in much the same way the Puritans were a cultural fork of the British Isles.

So in many ways it's not useful to call it "tech" vs the "media". They are more similar than different in key respects. And they're demographically and economically different from the public at large.

Address disalignment with decentralization

So we've established that the two types of centralized media corporations (and their leaders and employees) are both very similar to each other and very different from the population at large.

One can argue that this is problematic for many reasons, but perhaps the most compelling is the lack of alignment. If what's good for GM isn't necessarily good for America, then what's good for Facebook or the New York Times isn't necessarily good for the world.

This is not an insult towards either Facebook or the New York Times. It's simply an observation of economic reality: there are many win/lose scenarios where Facebook or the Times wins, and the user or reader loses.

For example, Facebook may shut off a feature that your app depended on. Similarly, the Times may decide not to have a reporter fluent in your mother tongue. This makes economic sense for Facebook or the Times, but does not make economic sense for you.

And that's just one relatively uncontroversial example for each type of centralized media company. You can come up with many more. The point is that it's perfectly reasonable that they make a decision like that, one that is adverse to your interests. FB may not be able to support that feature, NYT may not be able to fund that reporter.

But then, it's also perfectly reasonable for the 99% of the world that is not an employee of a centralized social media or institutional media corporation to seek alternatives.

What might that alternative look like?

One way to address this is by trying to change the composition of tech and media companies alike. Great organizations (Black Girls Code, SAJA, etc). are working on this, but progress is relatively slow.

But there is another alternative: make everyone equal to both a technologist and a journalist.

After all, there are only a few thousand tech and media jobs. But there are billions of tech users and media readers. Even if tech and media were fully representative of the population, we know that two people of the exact same socioeconomic background can be economically disaligned. That is, one of them can win while the other loses. Representation may well help, but if centralized media corporations continue to dominate then a user or reader is still trusting someone within a giant social media or institutional media corporation to do the right thing.

The alternative solution is to remove the obligate need for trust entirely, to fully decentralize power outside these organizations, to obviate the issue with truly universal equality of opportunity.

After all, over the last decade, billions of people have become users of social media platforms, and millions of people have become readers of digital media. Now we need to take the next step: to allow anyone to become full-fledged operators of social media hubs and individually-remunerated publishers of digital content.

Technological and cultural decentralization is not trivial

Note that this is not an easy task. The centralized media corporations of both kinds aggregate many specialized functions and operate on trust.

Facebook, for example, is a technological marvel which requires millions of servers and the best minds in computer science to serve billions of users (if you doubt it, see here, here, here, and here).

And the New York Times has more than a century of institutional credibility, with reporters in dozens of countries and the ability to publish >1000 articles per week under the glare of the 24-hour news cycle.

Decentralizing these functions is technologically and culturally nontrivial.

But Li Wenliang's experience during the virus crisis illustrates that we are in a race between technology and politics, between freedom and centralized control. And that relying on centralized social media companies or centralized media outlets may actually be dangerous to public health.

So, again, I think we really need to decentralize media, and do that right now. Here are some ideas on how.

Ten ways to decentralize centralized media corporations

Learn to write, report, publish, direct

Learn to code, yes. But we also need to learn how to write, report, publish, and direct.

Put another way, the tech industry already understands the importance of turning everyone into a technologist. I believe deeply in this myself, and taught a MOOC in 2013 on the topic that attracted more than 250,000 students.

But we also need to turn everyone into a journalist, and a publisher, and a media company too.

Some of this is instruction. We want to put the Columbia School of Journalism online, or the equivalent. Teach everyone how to write. And not just how to write, but how to create slides, images, video, and audio content. At a higher level of abstraction, far more people should understand how to do investigative journalism, edit audio and video, write screenplays, and craft stories.

But some of this is technology too. Like Instagram turned everyone into a photographer, and Canva is turning people into designers, there's so much we can do in terms of templates, tutorials, and the like. We can put all the repeated steps into code, and we can make high-end content production accessible to all.

Now, the primary reason to do this is because we need to give a polished, powerful, uncensorable microphone to billions of people worldwide. To give a stronger voice to the next Li Wenliang, and to the next citizen who witnessed corruption in local government or the next victim of an unpunished assault.

But turning everyone into a technologist and a journalist also may have benefits for the ongoing East Coast journalist / West Coast VC conversation.

If you read recent pieces by Hunter Walk, Alex Wilhelm, or Chris Anderson, a common theme is the lack of understanding of the other side. Ben Horowitz talked about this as well, in terms of his Freaky Friday management technique. Being on both sides of the table builds empathy.

As a small example, if you tweet enough to see something go viral, you understand something of the incentives of a journalist and something of the power law of VC returns.

But we need bigger examples. Substack is wonderful because it turns people into both journalists and media companies, and allows people to understand the challenges of reporting and running a small business. As a writer, you'll find yourself resisting the urge to write a provocative headline on some business; as an entrepreneur, you'll find yourself realizing how hard it is to build any kind of business at all.

We need that for every other kind of content too. But we'll only get there if we make it easy for people to learn to write, report, publish, and direct.

Every company a media company

The decentralization of media does mean moving power away from centralized social media and institutional media companies to billions of individuals, many of whom also become one-person companies themselves (more on this below).

But it also means decentralizing it to millions of medium-scale companies.

Specifically, we are in the middle of a phase transition where every company is becoming a media company. Social media and content marketing was version one of this. Every major corporation has a Twitter, a Facebook, a blog, a YouTube, and so forth. And at first they used them mainly for corporate announcements and customer service.

But now you have brand accounts doing risque jokes on Twitter. And you have tech companies are pursuing original content. Netflix and Amazon in particular showed the way, and are some of the biggest studios in Hollywood.

The next step is go from blog posts and content marketing to continuous arcs, to entertaining and educational narrative storylines. There's a real story behind any company, there's something that motivated people to quit their job and work full time there, and making that purpose legible and tangible to people brings them in as partners and friends rather than just customers. Wistia's one, ten, one hundred is a great example of this.

One of the benefits of building an audience this way is that when your company needs to get an important message out, you can. You don't need to trust a centralized media corporation in your time of need. Instead, if you build an audience by giving them educational and entertaining content, they will amplify you in return. Which leads us to our next point…

Build your own distribution to avoid distortion

After you learn to write and report, and start creating original content for your new one or N person media company, you need to build your own distribution. This can be a following on social media, but ideally it's an email list or a set of phone numbers along with the explicit permissions to contact the people on that list for the unspecified purposes they've opted into.

The reason is that an email list or set of phone numbers is exportable in a way that a social media following isn't. Put another way, it's easy to export all your tweets from Twitter. But you can't contact your Twitter followers without permission from Twitter. You can, however, email the people on your email list without permission from Substack.

That's what I mean by building your own distribution.

One reason to build our own distribution is that in the event you are cut off by any given centralized internet company, you can still recontact your users.

But another reason is to avoid distortion.

To understand this, we can make an analogy to the concept of a noisy channel from electrical engineering. In a noisy channel, you might send in a binary 1 but the receiver might hear a 0. Getting your message across in the presence of noise is a fundamental problem in electrical engineering, and inspired both Shannon's information theory and Hamming's error-correcting codes.

But while the undergraduate model of noise is usually just random, there's a whole new level of difficulty when dealing not just with noisy channels but adversarially noisy channels, channels which are economically or otherwise disaligned with the sender and which have an incentive to distort the message. They aren't injecting random noise, they are injecting noise which is designed to corrupt the message you're sending in some way.

Both kinds of centralized media corporations do this.

Centralized social media and internet companies like Google, Facebook, and Twitter are frequently called out by people inside and outside tech for deplatforming companies, cutting off their API access, downweighting their posts, and otherwise interceding themselves between sender and receiver. It's understandable that they do this – they are economically incentivized to do so – but their actions are not always in the best interest of sender (the posting entity) or receiver (the reader).

Centralized media outlets like the New York Times are also frequently accused by people inside and outside journalism (like their former editor-in-chief, Jill Abramson) of things like clickbait headlines, quoting people out of context, editorializing the facts, and otherwise interceding themselves between sender and receiver. Again, it's understandable that they do this – they are economically incentivized to do so – but their actions are not always in the best interest of sender (the subject and sources) or receiver (the reader).

Now, one way to deal with a noisy channel is to encode your message in such a way that it's harder to be corrupted. By analogy, this is what Paul Graham advises doing.

But another way is to simply construct a different channel with better noise characteristics.

Once you realize that centralized media corporations of both the social media and institutional media type can be noisy, adversarial channels, that's when you realize you need to build your own distribution to avoid distortion. And that's true whether you are a tech or media entrepreneur.

A million hubs and a billion nodes

We need to technologically decentralize social media platforms.

Just to level set, I like Jack and Zuck and respect what they've built. After all, many people only complain about Twitter because they use it so much and gain so much knowledge out of it, and wish it could be better (for some definition of better!). And whatever you think of Facebook, it has given free long distance phone calls, texts, video chat, group messaging, photo albums, event management, white pages, yellow pages, and dozens of other services to literally billions of people. (Yes, there are ads. No, you don't have to click them. Yes, we should build more private systems over time. No, that doesn't mean that giving billions of people these services at no charge was not a step forward.)

Despite my respect for what they've built, I also know they run large centralized social media corporations, and their interests are sometimes adverse to mine. And so do many other people. Some of the many criticisms people have include:

  • Too much moderation, or not enough
  • Arbitrary demonetization and deplatforming, both at individual and company level
  • Too much of a walled garden, or too freely accessible APIs
  • Unreliable APIs
  • No explicit governance or social contract

I'm sympathetic to many of the critiques of social media by traditional journalists. These platforms are in nontrivial part responsible for not just clickbait, but culture. Whatever ideology people may nominally have often degenerates into Twitterism.

I'm also sympathetic to the problems Jack and Zuck have. Jack is like the leader of a 300M+ person country that seems to be at civil war with itself. Zuck has many of the same problems at even larger scale. They are like the presidents of gigantic nations wracked by civil war. They actually don't seem to have much power to stop this war.

But to their credit, they have both realized they need to give that power away. Jack is doing this with @bluesky and Bitcoin. Zuck is doing this with the move towards private messaging and decentralization.

We'll see how much progress is made on these initiatives, but decentralization of power away from the platforms is the right long term direction.

Li Wenliang's experience with WeChat shows us what happens if this doesn't happen in time. It's extraordinarily technologically impressive but already lost on the dimension of decentralization. It has become a tool of social control.

So we need to decentralize the platforms. That likely means a federated model with a million hubs and a billion nodes, each with their own ranking, governance, community norms, and the like.

We know from GitHub that there are millions of developers, so there are enough people to deploy and maintain technical hub software. The end result might be sort of like Substack, except rather than each journalist maintaining an email list of readers for a newsletter, each developer maintains an email list of users for a small social network.

We have version one of this already with Mastodon's Fediverse and Secure Scuttlebutt, but there will be more. Once we have large numbers of hubs with different characteristics, we start to reintroduce choice into social media for billions of users. And thereby decentralize the social media platforms.

Redistribute your influence to help the up-and-coming

Last year I retweeted @skweird's new publication on tech innovation in Africa, This was a tiny effort for me but brought exposure to a wonderful project by an up-and-coming young founder. I think tech and media people alike need to do way more of this.

That is, we should redistribute what influence we have to rising stars, both in the US and around the world. Just like we want to technologically decentralize away from centralized social media platforms, we want to culturally decentralize away from institutional media outlets.

The concept of cultural centralization

To explain what I mean by this, it's obvious that centralized social media corporations like Facebook and Twitter are technologically centralized. To send a message to someone else you are going through their servers. They are a single point of control. They can downrank your post, suspend your account, or even deplatform you entirely. They can hobble the API, raise prices, or change the terms of service on a whim. They are a single point of control.

But for institutional media outlets like the New York Times (or even Vox or Buzzfeed), this is less obvious. Aren't there many more outlets than social media companies? How are they centralized?

The answer is that they are all culturally centralized. They're all singing from the Brooklyn Hymnal. You can see this from LinkedIn prosopography. Go through three bios as an example. It's all the same. Exeter to X to Y to Z. W to A to B to C. A to B to C to D. It's a closed loop, a culturally centralized system, a hermetic closure.

That's why Brian Stelter and Jake Tapper refers to 'the media' as a singular entity. Because while the boundaries of the set are fuzzy, there is something there.

The reason cultural centralization is a problem is the same reason technological centralization is a problem. The problem is not the existence of Facebook or the New York Times, or not the existence of Twitter or any other publication. The problem is the lack of real choice, the presence of technological and cultural centralization.

Now, that choice is starting to return with the advent of dozens of decentralized social media platforms and millions of influencers, freelancers, blogs, and part-time journalists. But we need to consciously support them when they are small.

Like retweets, interviews and quotes can be endorsements

Retweets really are often endorsements. So are links. And, less obviously, so are interviews and quotes.

When you give someone a quote for attribution, you are giving them power. The power to amplify, but also to distort. This is the same as the alignment problem discussed above. A centralized media company is an economically misaligneddisaligned actor that may not act in your best interests, and that may distort your message for its own ends.

For example, a social media company like Twitter may put your tweet in front of people that hate you to try to gin up a mob, as its algorithm says that it results in more clicks and "engagement". And an institutional media outlet will sometimes seek to do the same thing with a quote, as some of the journalists in NY Mag's 2016 survey wrote.

The convention of giving quotes to large centralized media outlets arose historically due to distribution limitations. Not everyone could afford a printing press. But now that everyone has a smartphone, you need not always give quotes to people who still buy ink by the barrel.

Put another way, when you post something on a centralized social media website like Facebook or Twitter, you are giving them power in the form of pageviews, ad inventory, and the like. You may be getting something out of it, but you are giving something as well.

Similarly, when you give a quote to a centralized media outlet like the New York Times or Wall Street Journal, you are giving them power in the form of pageviews, ad inventory, and the like. You may be getting something out of it, but you are giving something as well.

But just as there is a way to decentralize centralized social media platforms technologically, you can decentralize media outlets culturally. For example:

  • Turn down the Recode podcast to go on the podcast of a woman starting her first fund
  • Ignore the New York Times reporter in favor of giving a scoop to that rising influencer from Nigeria
  • Stop reflexively giving quotes, interviews, fundraising announcements, or even responses to institutional media outlets and start thinking about freelancers, specialist publications, startup blogs, and local outlets

Basically, if you're a tech founder or VC, you should think twice about giving the Nth scoop to an established outlet, and instead redistribute your influence to help up-and-coming media entrepreneurs, freelancers, and influencers.

There is a global community of budding media entrepreneurs and technologists both in the US and around the world who would be delighted to hear from you, to interview you, to have you on their podcast, to teach you about what they know.

Cultural decentralization is both ethical and rational

This works for both ethical and rational reasons.

Ethically, every time you give an interview to a rising start you are redistributing influence away from multinational media corporations and towards freelancers, influencers, individual experts, startups.

Rationally, by doing this, you start leveling the playing field between a one-person Substack and a 1000-person newspaper posting on Twitter. It is in everyone's interests to do this to culturally decentralize the institutional media. The only people who might quibble are the employees of large media companies, perhaps, but they're doing fine these days.

Otherwise we have a boring world where the Times can use its brand to get a top podcast, just like Microsoft could use its Windows brand to promote Internet Explorer. And where a few large institutional media companies gobble up all the subscriptions, leading to subscription fatigue.

You might think that giving an interview to a smaller outlet is less exposure than institutional media, and it can be, but go and take a look at the NYT's twitter feed. The engagement numbers really aren't that high for such a huge account. Nowadays the content often stands on its own. So long as content is compelling and the person you're talking to is reputable and fair (more on that in a second), you can simply share the link on your own social media and there will be no shortage of views.

By the way, Jack Dorsey seems to agree. Last year he tweeted this:

While Jack did end up doing a Twitter interview with Kara, he went in person to visit entrepreneurs in Africa:

That's what cultural decentralization means. Fewer interviews for well-established employees of institutional American media corporations. And more content for freelancers, substack writers, media entrepreneurs, and influencers around the world.

Move towards decentralized PR

This section is for founders, execs, VCs, entrepreneurs, and their PR people.

Let's be clear that the previous recommendation is a major adaptation. A move from centralized institutional media to decentralized influencers and freelancers is a complete rewiring of the reflexes of traditional PR folks, who are often ex-employees of said institutional media companies.

Operationally, shifting from cultivating a few relationships with employees of centralized media corporations towards cultivating hundreds or even thousands of relationships with decentralized influencers, freelancers, experts, and related startups around the world is a huge transition.

To be sure, it can be useful to have a PR person who knows someone at a centralized media corporation like the New York Times, just like it's useful to know someone at a social media company like Facebook. But you want a PR person who thinks "influencer first", who thinks about decentralized journalists first.

And yes, there is such a thing as a decentralized journalist. The left (Splinter), right (Federalist), and center (Shafer, WaPo) all agree: if they commit acts of journalism they are a journalist. Perhaps more importantly, so do many shield laws.

Some privilege schemes are narrow and apply only to full-time employees of professional news outlets, while others are broad and extend to bloggers, filmmakers, freelancers, book authors, and student journalists

This means that legally and formally, in many places a blogger, filmmaker, freelancer, book author, student journalist IS on the same plane as an employee of a professional news outlet. Let's call the first decentralized journalism and the second centralized journalism.

With respect to PR, there are two ways to handle the transition to decentralized journalism.

The first way is to look for those PR people who can handle the transition from centralized institutions to decentralized influencers. It's like the desktop to internet transition, or the finance to crypto transition. Some folks from the traditional world get it instantly and can make the jump.

The second way is to hire new PR people who come natively from that world, just like institutional media did over the last 5-10 years. Many bloggers and tweeters are now working at centralized institutional media outlets, or even running them (like Ezra Klein).

This a refreshing change for PR people. Rather than spamming jaded institutional journalists with unwanted press releases, you can give a big break to a decentralized journalist. Rather than some high-stakes nail-biter of an interview with a legacy outlet where you hope they get it, you can have 1000 casual conversations with people who have context and expertise in the field.

Collectively, those decentralized actors have more reach and genuine engagement. They're community members, not tourists, and you're playing an iterated game. And morally, as a PR person you are doing the right thing by raising up new voices and decentralizing power away from centralized media corporations towards decentralized journalists, freelancers, bloggers, experts, influencers.

A great way to start is with something like Buzzsumo. I have zero involvement or stake in the company, but it gives a useful way to search for influencers in your topic area.

Go ahead and try it for a week or a month. You'll find that it's possible to completely ignore traditional institutional media channels in favor of a decentralized approach.

Turn every user and reader into a paid publisher

Today, 27% of kids want to be YouTube influencers. Substack and Patreon are growing. And, interestingly, centralized journalists are starting to get paywalled from below.

Why is this happening? It's actually often apolitical. These Gen Z kids think of themselves as peers of the journalists who are asking for their content and don't want to give away free content without reimbursement.

I'm sure that some of these users dislike institutional media for political reasons (whether on the left or right). But for many of them, theirtheir their mental model of institutional media is neither rightful guardians of democracy, nor ruthless oppressors and fake news, but simply a rival gang.

The Gen Z kids understand the value of attention and would rather build their own brand rather than that of the institutional journalist asking them for free content. They know in their bones that they are equal to the institutional journalist.

Now come back to that number: 27%. This is a growing wave that is just getting started.

Millions of American kids these days are writing stories for TikTok videos, editing videos on YouTube, creating content, growing up as digitally native journalist equivalents. The production values of user-generated content are consequently about to soar.

And for those production values, these kids want to get value for the time they put in. Due to the YouTube adpocalypse, people are looking into new decentralized video platforms like dlive. These are still in their infancy, but they are coming along.

To recap, we are still in the early phases of the total decentralization of media. The future may look more like substack with 100 people at $1000 per year rather than 1 at $100k.

That's the same dollars in, or perhaps more, but with vastly more knowledge, experience, expertise, etc. It's a huge part of moving from a centralized standing media to a decentralized citizen media.

There are many tools to build: copy editing, grammar checking, AI video, style transfer, etc. It's already happening with stock photos and photography. We could probably improve the experience of one-click incorporation with Stripe Atlas and Carta, and crypto over the long term.

Set up community review and rating systems

If we decentralize media we will need decentralized quality control.

We already know that people seek out reviews of corporations, products, and apps. And we know that books and movies have been reviewed for decades with explicit 5-star ratings without any ill effect. Extending the same concept to more forms of media is an obvious next step.

Indeed good authors want a review system because it lets them stand out. This is why people brag about being on the NYT bestseller list, winning a Booker Prize, being #1 on Amazon, or getting a 5-star review.

There's another aspect which is not just review from everyone as a whole but community-based reviews. I want to know what the Muslim community thinks of the NYT's Iraq War coverage, what Chinese people think of Western coronavirus coverage, what Uber and Lyft drivers think of AB5 coverage. Is it accurate? Did it correctly predict events? Does it reflect the values of their community?

If it doesn't, that reviewer should be able to jump off the page and set things right from their community's perspectivepersspective. This will be valuable for everyone. Amazon book reviews weren't the end of books, Rotten TomatoesTomatos wasn't the end of movies, Glassdoor wasn't the end of startups, and community reviews of news won't be the end of media either. We need third party critics who don't work for media corporations to score the correctness of old stories.

Automatically find each topic’s Gell-Mann

The concept of Gell-Mann amnesia is critical:

“Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well. In Murray's case, physics. In mine, show business. You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. I call these the "wet streets cause rain" stories. Paper's full of them.

In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know.”

– Michael Crichton

Every article has a Murray Gell-Mann. How do we find them? One answer is to combine something like with and textual search. To generate a power ranking of experts within a field, plus a search widget that displays those experts next to the text along with their commentary.

Change Incentives to reward positive behavior

Right now social media is a pure status game. What is rewarded is popularity, not truth or building valuable things. However, that doesn't necessarily mean that you only focus on truth and value.

Think about residential, industrial, commercial zoning. It's unnatural to separate them out. People want to work, shop, and live near each other. Perhaps the first generation of internet companies built out models that work on their own (like social media, Saas, and so on). And the next generation will start to combine them into the equivalent of walkable neighborhoods online.

One way to do this is to allow people to make money online. Another is to let them post pseudonymous. A third is to reduce more issues to bets, with resolvers, rather than making them all arguments.

Everyone verified and everyone pseudonymous

Decentralizing the media also means scalable authentication of sources, one of the five pieces of identity.

It's one thing to trust a few centralized media outlets, but it's quite another thing to verify facts coming from a billion decentralized influencers, freelancers, journalists, and bloggers. A big part of this is verifying who they are, especially in this dawning age of AI-generated photos and bot armies.

But we also need to protect sources, and protect people who are speaking up from being retaliated against. So pseudonymity is important. And yet on the third hand, if there is a quote from someone anonymous or pseudonymous, we need some way to verify beyond just simply trusting the writer.

Fortunately, there may be a technical solution to this which I sketch here.

The basic idea is to have a social network where people can verify their credentials under their real name, but then use zero knowledge proofs to port a subset of those credentials to a pseudonym. It's the same technology underlying privacy coins like Zcash, except applied to credentials rather than cash.

A high profile example is the problem of verifying who Satoshi Nakamoto is. The real Satoshi would be able to establish themselves through a digital signature that showed they owned the first mined Bitcoin. This would not give up their real name, but would verify the professional 'credential' that they were the inventor of Bitcoin.

A zero knowledge proof would provide even more privacy protection than this sketch, as it would prove only that the user was Satoshi, and not reveal other information like which coins they held. However, by rough analogy, many other credentials could be established in the same way as proof-of-Satoshi pseudonymous digital signature.

This would allow us to combine source protection (via pseudonymity) and source verification (via zero knowledge) in such a way that even a skeptical audience could be shown the relevant credentials of the source, but nothing else.

Replace trust with technology by decoupling fact and narrative

Finally, let's talk about the big one: how do you do decentralized fact checking in the absence of institutional trust?

Perhaps the single biggest problem to solve is how to increase the credibility of a decentralized freelancer or citizen journalist when they post something true, and to reduce it when they post something false.

As mentioned at the beginning, is it possible for some unknown to not just commit acts of journalism like a New York Times reporter, but to be taken as seriously when publishing something true? Or, conversely, for a New York Times reporter like Judith Miller, Jayson Blair, or Walter Duranty to not be taken seriously when publishing something obviously false?

This is a huge ask and not possible for every kind of fact. It is possible for some kinds of facts, but it will not be an overnight transition. One way of doing this is to start gradually replacing trust with technology, by leaning on decentralized verification rather than institutional reputation.

We're not going to solve this in one subsection of a blog post. But let me sketch a 1/2/3 strategy to do this, essentially a way to develop unit testing for text.

  1. First, parse text with natural language processing to computationally identify factual assertions as opposed to narrative/sentiment. Make graphs. Simultaneously, seek to put facts online with crypto oracles, as a generalization of twitter feeds.
  2. Second, set up a configurable truegle which hits knowledge graphs, oracles, wolfram alpha, etc to verify different kinds of facts algorithmically and explain how truegle knows what is true.
  3. Third, once we know what is fact vs narrative, plug these facts into explicit epistemology algorithms to see if the conclusion follows from the facts. If there's a statement that truegle can't verify, indicate that.

The overall concept is to develop something that is to natural language roughly what a compiler is to code. Hit enter and you get out a visualization of whether your argument appears to match known facts, whether it contradicts them, or whether it introduces new ones.

In terms of fact-vs-narrative, sites like Axios are already doing v1 of the facts with their bullet pointed stories for mobile and sites like Vox have invested in narrative with their explainers. But there's a lot of technology that can be brought to bear to streamline this.

In short, even in a decentralized media ecosystem, facts can be checked and processed by open source algorithms and databases, while narratives can be written (or detected, eg the use of Russell Conjugation) by an algorithm.


The bullet points above could be expanded into articles of their own. And there will be many more ideas about how we can decentralize media.

But consider this a partial roadmap for how and why we should decentralize power away from centralized media corporations (both social media and institutional media) and towards decentralized citizens, influencers, freelancers, and experts.

The how is both a technical and a cultural matter. It may mean zero knowledge proofs and explicit epistemology algorithms, but it also means remembering to give that scoop to an up-and-coming freelancer rather than reflexively giving it to an employee of a giant media corporation.

As for the why, as Li Wenliang showed us, the fate of the entire world may depend on giving someone who is neither technologist nor journalist the reach of social media and the credibility of institutional media.

Right now, Chinese social media companies like WeChat cooperated with the Chinese government to silence a voice, while American institutional media companies are publishing articles on how "free speech is killing us".

It's quite the contrary. It's the lack of free speech that's killing us, it's the centralized control over speech that's killing us. Had we listened to a decentralized citizen journalist talking about the people who'd fallen sick at a seafood market on the other side of the world, we might all be in a different place.

← 100% Democracy vs 51% Democracy
Getting started with 1729 →

Subscribe to 1729

Subscribe to the newsletter and unlock access to member-only content.

You've successfully subscribed to 1729
Welcome! You are now a 1729 subscriber.
Welcome back! You've successfully signed in.
Success! You are now a paying member and have access to all content.
Success! Your billing info is updated.
Billing info update failed.