Startup T2 Wants to Terminate Twitter
In mid-2021, Gabor Cselle bought a $15 Moleskine notebook to sketch out ideas for new startups. On the first page, he wrote "T2" and began taking notes for a better version of Twitter. Cselle had sold startups to Google and Twitter and worked at both companies. (He was at the time at Google for a second stint, as a director at Area 120, its startup incubator.) But he couldn't figure out how to draw people away from "T1"--the original Twitter--and set the idea aside.
Then came Elon Musk's Twitter takeover, which saw its new owner lay off over half of Twitter's staff, troll the community with alienating tweets, and speculate about adding features like long-form video. "It was basically the worst-case scenario of how Twitter shouldn't be run," says Cselle, who finally did leave Google last summer. (He got out just in time: Last month's layoffs essentially defunded Area 120.) The time had arrived, he felt, to pursue the dream of T2. He finally had his differentiator: His version of Twitter would be more like ... Twitter, in the classic sense. T2 would be less a revamp than a restoration, an attempt to recapture the excitement of early Twitter and build from there.
T2, which won't be the final name for the product, is now live in a very limited test version. Nine people work for the company, including Cselle's cofounder, Sarah Oh, who had been an executive in user safety for Facebook and, most recently, Twitter. Last month T2 received $1.35 million in angel funding from several well-connected Silicon Valley investors.
But T2 is far from alone. Cselle is talking with me at a bustling WeWork with spectacular views located in Salesforce Tower in San Francisco. It could be that maybe half of the bright young techies typing at work tables and sofas are building new social media apps to challenge Twitter or other social apps that have lost their charm in pursuit of mass audiences and ad revenue. T2 faces startup competition from Mastodon, Countersocial, Post, Hive Social, and more. All of them have different twists on a short-form social network. None of them are quite as brazen as Cselle in claiming to duplicate what was once the thrill of the original.
"People can't resist futzing with the format, but it works," Cselle says. "People have a background process in their brains: What is a crispy 280-character thing I can say about this thing that just happened? Why mess with that? And what if you can get that same crispy 280-character thing in front of people who are really relevant to you? I think that'd be pretty cool."
It would also be bucking what in retrospect seems like a gravitational pull away from social networks being social. The pursuit of the viral has diminished the intimacy of the personal, and as the business models of the early networks focused heavily on delivering audiences for advertisers, they increasingly became a new version of broadcasting. Social networks once obsessed on Dunbar's Number, the claim that humans can only meaningfully interact with 150 people they know well. What you saw was determined by who you knew, or who you wanted to know more about. Now Meta, Twitter and the rest algorithmically connect you to "content you may be interested in," which more likely than not involves influencers who spend all their time concocting ways to grab your attention with calorically empty content. Or stuff that enrages you. Cselle wants to roll back the clock as if all that never happened. "It's kind of retro," he says. "Remember what Twitter felt like in 2007 when it was real people sharing things from their life and not airbrushed Tiktoks?"
That's where T2 and other startups see an opportunity. Instead of worrying about Dunbar's Number, Cselle is more concerned with the Allee threshold, a biological effect named after the ecologist who postulated it. Basically, it states that in any ecosystem there's a particular number of inhabitants needed to sustain a healthy population. Cselle thinks that T2 will hit the Allee Threshold this spring when it includes 5,000 people, enough to generate the level of relevant content needed to keep users engaged. The trick needed for T2 to thrive is for it to host a multiplicity of communities, so that the service itself can grow while still maintaining that relevancy for individuals. (That idea is taken to an extreme on Mastodon, which immediately forces newcomers to pick what tribe they will be associated with.)
To make sure that T2's communities are healthy, Cselle and his cofounder, trust and safety specialist Oh, are planning to authenticate users. But they won't be charging $8 for the privilege like Musk's Twitter. Instead they will actually attempt to confirm users' identities, which will cut down on hostile attacks and harmful misinformation. "I think there's a space for building an authenticated network, where you actually are pretty sure that if someone claims to be Steven Levy on T2, it is actually Steven Levy," he says. Right now, that's easy, because only invited people can join. Later, they may check IDs or try to cross-check with other social media accounts. But Cselle says that while not everyone might have airtight credentials, certainly those with the equivalent of blue checks will.
Cselle got me past the waitlist to peek at the just launched, very alpha version of T2. (I'm user number 76, and the total is still in double figures.) The app is a clean, chronological feed of comments from testers drawn from personal connections of the founders. Much of the discussion is about T2 itself, as the community workshops the product. But the feed also weaved in asynchronous discussions on vegetable gardens, Oxford commas, and the most recent episode of The Last of Us. The minimalism is refreshing, the vibe is welcoming, and it's a relief not to have to pose as or promote something. One user comments that the experience is like "a quiet gathering on a large lawn with wine served and a hushed din as guests arrive." But what will happen when giant flat-panel screens are installed and hordes of beer guzzlers arrive? That's the hard part.
For now, Cselle's team is bathing in an idyllic moment of pure potential. T2 is nascent enough that Cselle can ask around for ideas on its interface, taking a poll to see whether people prefer flat or dimensioned (that is, shaded) icons. But it's telling that his poll was not shared exclusively with the double-figured T2 community, but instead was conducted on the noisy scrum of the platform he's supposed to be replacing, owned by Elon Musk. Cselle is even planning for T2 to support cross-posting with the rival. Even he realizes that, lost magic or not, T1 will be hard to topple.
When Twitter was a startup, Jack Dorsey, Noah Glass, Evan Williams, and others had to figure out the rules that would allow the service to reach its own Allee threshold. I wrote about this for WIRED in October 2009.
When Twitter's creators designed the service, they made a series of crucial and deliberate decisions--ones that seem brilliant in retrospect--that created the conditions that allow users to innovate.
The first was a commitment to simplicity. Jack Dorsey suggested a system that mimicked the simplicity of SMS to let users send messages from the web or their cell phones about what they were doing from wherever they were. In March 2006, a small group including Dorsey and Stone built a prototype in two weeks. "The important thing was that we spent a lot of time getting it down to its essence," says Noah Glass, a cofounder of Odeo who managed the so-called Twttr project. The system would receive a message from the sender and then forward it to the right people. Nothing else. Simple.
The second key decision was crucial: creating asymmetry between writers and followers. They didn't need to be "friends" or in any way on equal footing. Anyone could read a writer's updates, and that was powerfully liberating. "One thing I didn't like about social networks was that awkwardness of friend requests," Williams says. He wanted Twitter to be more like blogging, where readers pay attention to whatever they like. "That frees up creators, because they can do anything they want," he says.
The implications were profound--and unexpected. No one thought people would want to follow strangers, or that celebrities would use Twitter to apprise fans of their activities, or that businesses would use Twitter to announce discounts or launch new products. Allowing unrestricted following eventually meant that P. Diddy could share the progress of a tantric sex session with a hundred thousand followers, and the Kennedy family could use Twitter to keep the public informed about developments in Uncle Teddy's funeral. It obliterated the line between confidant and audience.
Ken asks, "Thirty years after you first wrote about the crypto wars, privacy is lost, but more to big new corporations than to the government. Does the outcome surprise you?"
Thanks, Ken. You asked this question in response to a tweet--yes, people are still getting value from T1!--noting that three decades ago, my story "Crypto Rebels" appeared on the second ever cover of WIRED. And in 2001 I wrote a book called Crypto with the subtitle "How the crypto rebels beat the government--saving privacy in the digital age." Oops.
Yes it's true that Little Brother--meaning private industry--has run rampant over our privacy by collecting vast amounts of our personal information. That's not the fault of cryptography, but our own failure to force those companies to use that technology to protect our privacy. But we can take heart that while it has taken a long time for privacy tools to protect consumers, it can happen.
It's taken several decades, but several mass market companies, like Apple and Meta, now routinely encrypt consumer texts and messages so that no outsider can view them. Millions of others use tools like Signal or Telegram. The next step should be to popularize the use of cryptographic authentication techniques to limit the information you share about yourself to what's necessary to complete a given transaction. If you're renting a car, for instance, you could verify that you have a driver's license without sharing the ID number. Or you might share your fandom of a sports team with a social network without tying that information to a more detailed profile, or the ID number of your phone.
The data-profiling industrial complex doesn't want this to happen, so consumers must demand it from tech companies, information brokers, and legislatures that have failed them so far. It may take a while, but the algorithms aren't going anywhere.
You can submit questions to mail@WIRED.com. Write ASK LEVY in the subject line.
The IRS has a plan to collect taxes after the apocalypse. (But the Four Horsemen will finally take their ride when the tax agency collects a fair share from billionaires.)
Startups competing with TikTok might not have to worry about competition if the government follows this roadmap to banning it.
Not all crypto products are equally secure, as Telegram users in Russia have found out.
Who would guess that peaceful villagers in remote rural Scotland would quibble with a billionaire building a fire-belching spaceport in their peat-bog paradise?
The ultimate way to duck responsibility: Set things up so you can say, "The AI did it!"