Skip to main content

A decentralized Twitter would bring the company back to its past

A decentralized Twitter would bring the company back to its past


Before Twitter was a business, it was a design playground — and with Blue Sky, it might be once again

Share this story

The Twitter bird logo in white against a dark background with outlined logos around it and red circles rippling out from it.
Illustration by Alex Castro / The Verge

Today we’re going to talk about Jack Dorsey’s surprise tweetstorm about potentially decentralizing the service — but first, some history.

Death came to the Twitter developer community on August 12th, 2012. In an infamous memo, the company’s head of product divided potential uses of the Twitter API into four quadrants. In the past, developers had been able to make any sort of Twitter app they wanted to — including a full-featured, ad-free Twitter client that they could customize any way they wanted it to.

It was a policy that led Twitter to become, briefly, a design playground for some of the world’s most talented user interface designers. But the policy also ran counter to the vision of Dick Costolo, who had become Twitter’s CEO two years previously. Costolo had come from Google, where he learned how to build advertising businesses. And so not long after he became Twitter’s CEO, he set about turning Twitter into an ad business.

Among other things, that meant aggregating as many eyeballs as possible into one place. A third-party Twitter client might be prettier and more functional than Twitter’s own client — shout out to Tweetbot! — but it certainly would not be more profitable.

A more ruthless company would have shut down API access to Tweetbot and its brethren. Instead, Twitter opted to let the third party clients bleed out over time — denying them access to new features, such as polls and group messages, and capping the number of users they could have.

That was good news for people like me, who prefer third-party clients for various reasons that I would happily share with you in person until you run screaming from the room in boredom. Every once in a while, Twitter would ask shellshocked survivors of its developer community for suggestions on a path forward. “Re-establish robust APIs that once again allow third-party developers to build full-featured Twitter alternatives!” I would weakly shout back.

And then one Wednesday morning in December 2019, Jack Dorsey ... said he would think about it?

Twitter is funding a small independent team of up to five open source architects, engineers, and designers to develop an open and decentralized standard for social media. The goal is for Twitter to ultimately be a client of this standard.

This five-person team, to be known as Blue Sky, will be charged with the project — effectively turning Twitter the platform into Twitter the protocol. In such a world, Twitter would be to tweets as Outlook is to email: one client for reading and writing messages among many.

Why does Twitter want to do this? Dorsey seems to feel less comfortable with the idea of a single, centralized network with one global set of rules. He notes that it places terrible challenges on content moderators. He argues that Twitter’s value lies in directing your attention toward valuable tweets — not hosting all the content. (As critics have noted, this direction could allow Twitter to avoid responsibility for some of the platform’s unintended consequences. If it doesn’t “own” every tweet, it’s no longer accountable for moderating them.)

In any case, this move toward protocols was one that many developers hoped Twitter would take back before the four-quadrant memo of death. And it was a direction that others would try to take in the wake of that memo: most notably, a Twitter clone that set to build an open standard and somehow build a business around it. It was built by a guy named Dalton Caldwell, who began work on it after writing a popular essay called “What Twitter Could Have Been.”

To understand both the promise and the perils ahead of Twitter as it pursues decentralization, you ought to read Caldwell’s follow-up to that essay. He writes of the reaction he received to his Twitter critique:

The responses to my post largely fell into two camps. One group is of the belief that a non-commercial, open source, open standards federation of real-time protocols is the solution. The opposing group has pointed out that these decentralized efforts never work out, and the API-focused service I wish existed is the fevered dream of navel-gazing geeks.

You could see those takes repeated all over Twitter today as geeks imagined what Twitter might look like as a protocol. Many people pointed to the quick collapse of, the service that Caldwell founded to embody the ideals in his manifesto; and the relatively slow-growth of Mastodon, a decentralized Twitter alternative that I profiled for The Verge in 2017.

Mastodon’s challenges give you some idea of what Twitter is up against. Decentralizing a network makes it harder to find people, and half the appeal of Twitter is the sense that everyone is there. Organizing people back into tribes can do wonders for a social network — it’s why, for example, Reddit is my personal social network of the year. But it can also mean that you’re enabling the formation of hate networks. Does anyone doubt, for example, that a decentralized Twitter would have a fork that closely mirrors the right-wing service Gab? Mastodon sure does!

But let’s not get ahead of ourselves. At the end of the day, all we have to go on is a tweetstorm. (Well, two tweetstorms.) Twitter’s historic pace of development has been glacial, and the act of converting a public company into a decentralized protocol sounds extraordinarily difficult. Former employees I spoke with today were intrigued, if not exactly optimistic.

“Well it’s Twitter,” said one, “so nothing will happen for 20 years.”

YouTube takes a fresh look at harassment

This summer, YouTube found itself facing serious criticism after it declined to remove videos posted by right-wing commentator Steven Crowder in which he repeatedly called video host Carlos Maza a “lispy queer,” among other things. And so, after many fumbles, YouTube said it would reconsider all of its harassment policies.

It was confusing, because YouTube’s policies already banned “content that makes hurtful and negative personal comments/videos about another person.” The issue, YouTube said at the time, is that Crowder made his comments in the context of much longer videos that amounted to fair-game media criticism.

At the time, I laid out what I wanted to see from YouTube: to hold big creators to a higher standard of behavior generally; to hold creators accountable when they incite harassment campaigns; and for YouTube to begin discussing its decisions in public, on the record.

Anyway, today YouTube published its revised harassment policies. And I got ... one of the three things I wanted? One and a half? Here I am at The Verge:

One, the policy expands the types of threats that are now banned. Historically, YouTube has banned direct threats like “I’m going to kill you.” Now, more veiled and implied threats will be banned as well. That means no brandishing a weapon while discussing someone, or altering a violent video game to put someone else’s face on a murder victim.

Two, the policy now bans targeted harassment campaigns. In an interview, the company told me that harassment on YouTube often doesn’t come down to a single insult. Instead, it’s a sustained effort over many videos. Under the new policy, YouTube will now take a more holistic view of what a creator is saying on their channel. Even if individual videos don’t necessarily cross the line, if they still contribute to the persecution of another person or creator, they’re eligible for removal.

This expansion of the policy directly addresses an omission that contributed to Crowder’s harassment campaign, which Maza illustrated with a viral supercut of the times Crowder had targeted him. At the time, YouTube said that because Crowder’s insults came within the context of longer videos about many other subjects, it would be unfair to remove them. The new policy should make it harder for other bad actors to use YouTube the way Crowder did.

Three, the policy now bans insults on the basis of a protected class, such as race, gender expression, or sexual orientation. So: no more “lispy queer” slurs. The policy applies to all individuals, whether they are creators or not, and even if they are public figures, where social networks have historically tolerated much more offensive speech.

Taken together, the changes would make the sort of harassing that Crowder became famous for against YouTube’s policies. But arguably, it already was. As Maza said today: “‘Malicious insults’ were already prohibited under YouTube’s anti-hate and anti-harassment policies. YouTube rolls out policies like this to distract reporters from the real story: YouTube’s non-enforcement.”

Or as I am fond of saying: your policy is what you enforce.

The Ratio

Today in news that could affect public perception of the big tech platforms.

🔽 Trending down: In the UK, thousands of misleading political ads have managed to avoid scrutiny thanks to Facebook policy’s not to fact-check ads from politicians. Almost all of the Conservative Party’s recent Facebook ads promote misleading claims, according to this report.

🔽 Trending down: Facebook and Google are no longer among the Top 10 “best places to work” in the US, according to Glassdoor’s annual list. The site uses employee reviews to rank office environments, and while high salaries and enviable perks used to be enough for Facebook and Google, employees’ perception of their workplaces is clearly changing.


Tens of thousands of political ads went missing from Facebook’s archive this week, just days before voters go to the polls in the most important UK election for decades. The bug meant people lost sight of who was spending what in the run-up to the election. Hadas Gold at CNN has the story:

A spokesperson for Facebook (FB) confirmed its library went down but could not say how many political ads disappeared. The problem affected several countries, and Facebook prioritized fixing the UK database because of the imminent election.

”We have fixed the bug and all of the impacted ads in the UK are now back in the Ads Library,” the spokesperson said.

Vietnam is the most important market in Southeast Asia for both Google and Facebook, but neither company will open an office there. They worry employees would be vulnerable to government pressure to turn over sensitive user data, due to the country’s oppressive censorship laws. (Wayne Ma / The Information)

A Facebook contractor was fired after being paid thousands of dollars in bribes by a shady affiliate marketer to reactivate ad accounts that had been banned due to policy violations. “This behavior is absolutely prohibited under our policies and the individual is no longer working with Facebook,” the company said. (Craig Silverman / BuzzFeed)

Former advertising executive Dina Srinivasan has helped popularize the antitrust case against Facebook. She argues that rather than raising prices like an old-school monopolist, Facebook harms consumers by charging them ever-increasing amounts of personal data to use the platform. (Jeff Horwitz / The Wall Street Journal)

France’s proposed copyright law weakens the already flimsy user protections from abusive infringement claims in the EU’s Copyright Directive. The rule suggests France thinks copyright laws are a tool to enrich the entertainment industry, rather than a way to benefit the public, this writer argues. (Mike Masnick / Techdirt)

China nudged past Turkey as the leading jailer of journalists this year, a new survey from the Committee to Protect Journalists shows. The change came in part because of severe repression in China’s Xinjiang region and Turkey’s eradication of “virtually all independent reporting,” which has left many reporters unemployed, driven into exile or intimidated into self-censorship. (Rick Gladstone / The New York Times)


Smart devices like Amazon Alexa, Google Home, and Apple HomePod depend on thousands of low-paid workers to analyze sound snippets. These people, who often work as contractors, are privy to peoples’ most intimate moments, report Austin Carr, Matt Day, Sarah Frier and Mark Gurman:

Former contractors describe the system as something out of the Tower of Babel or George Orwell’s 1984. At a GlobeTech office near an airport in Cork, Ireland, some say, they sat in silence at MacBooks wearing headphones, tasked with transcribing 1,300 clips a day, each of which could be a single sentence or an entire conversation. (This quota was reduced from as many as 2,500 clips, others say, to improve accuracy rates.) When a contractor clicked play on a voice recording, the computer filled a text box with the words it thought Siri “heard,” then prompted the worker to approve or correct the translation and move on. GlobeTech didn’t respond to requests for comment.

A program the workers used, called CrowdCollect, included buttons to skip recordings for a variety of reasons—accidental trigger, missing audio, wrong language—but contractors say there was no specific mechanism to report or delete offensive or inappropriate audio, such as drunk-sounding users slurring demands into the mics or people dictating sexts. Contractors who asked managers whether they could skip overly private clips were told no clips were too private. They were expected to transcribe anything that came in. Contractors often lasted only a couple of months, and training on privacy issues was minimal. One former contractor who had no qualms about the work says listening in on real-world users was “absolutely hilarious.”

Even as Big Tech is under attack, Apple, Google, Amazon, Facebook and Microsoft are doing remarkably well in the stock market. The trend suggests investors aren’t too worried about low public opinion or even mounting government investigations. (Matt Phillips / The New York Times)

TikTok’s parent company ByteDance is testing a new music app called Resso in emerging markets. The app, which displays real-time lyrics and lets users post comments under individual songs, is now available in India and Indonesia. (Zheping Huang and Lucas Shaw / Bloomberg)

Twitch just signed exclusive deals with streamers DrLupo, TimTheTatman, and Lirik, who have a combined 10.36 million followers on the platform. The news comes on the heels of a rash of big names leaving Twitch to stream exclusively on sites like Mixer, Microsoft’s streaming service, Facebook Gaming, and YouTube. (Bijan Stephen / The Verge)

A look at female leadership and modern communication tools like Slack, in the wake of the Away investigation and CEO Steph Korey’s resignation. (Carolina Milanesi / Tech.pinions)

The swipe defined dating this decade, taking people from a world of meticulously curated online profiles to split-second decisions about whether or not a person looks hot. Now, more apps are leaning into real life experiences, trying to facilitate interactions that actually end up in a date. (Ashley Carman / The Verge)

Instacart, an online grocery delivery company, has been experimenting with how it pays workers by using tips to supplement their wages. Now, some of the companies’ independent contractors are banding together to demand change.

Inside the insanely prestigious world of falcon influencers, people who drop millions of dollars on high-quality birds and race them in Dubai and the United Arab Emirates. (Isabelle Kohn / Mel)

And finally...

Talk to us

Send us tips, comments, questions, and decentralized tweets: and