OPINION: An Open-Source Twitter Is Possible, Just Not Likely

by Kevin Schofield

Without the Emerald, the true narrative of our community would rarely be told. For too long, and for too often, most media has painted our community in a negative light. When I say community, I include everyone who our mainstream media often ignores, diminishes, and casts aside. The Emerald has been here to remind our community of its worth, and that like all emeralds, karat for karat, the people of our community are worth more than gold. Join me in supporting the Emerald as a recurring donor during their 8th anniversary campaign, Ripples & Sparks at Home, April 20—28. Become a Rainmaker today by choosing the "recurring donor" option!—Phillip "Papa" Green, The Publisher's Dad (and Longtime Community Curmudgeon)

This week's news that Elon Musk is purchasing Twitter, aiming to turn it back into a privately held company with even less public accountability, has some asking whether an "open-source" version of Twitter could be created as an alternative for those who don't trust Musk — or the billionaires running the other major social media networks — to appropriately manage the balance between supporting free speech and facilitating the spread of corrosive misinformation.

The easiest part of building a Twitter clone is, perhaps, the software itself: The basic technology underlying Twitter and other internet services is not terribly complex, and much of it these days consists of "off-the-shelf" components. That's not to say that there aren't technical challenges. Twitter claims to have about 215 million active daily users worldwide, and serving that many customers requires architecting the software in specific ways to distribute the load and replicate information to data centers around the globe. It also requires translating the software into multiple written languages to create the web application and the Twitter mobile apps.

But those are largely "solved" and well-documented problems. There is a bit of magic in how Twitter decides which tweets to include in our individual feeds, and this is further complicated by its advertising business that requires also serving us relevant ads. Recommendation systems like these can be very complicated, especially as there is a parallel industry constantly trying to game those systems to their advantage. If an open-source Twitter did away with ads, that would simplify the service substantially, though it would raise new questions about how to raise the revenue to keep the service running. Last year, Twitter spent $1.24 billion on research and development (R&D), much of which went to writing and maintaining the code for the service and the attached advertising placement system.

Twitter spent even more, however, on operations last year — nearly $1.8 billion. Much of that is infrastructure: the data centers, raw computing power, and networking required to host and deliver the service, and the staff to build, run, and maintain that massive global infrastructure.

The good news for a prospective open-source Twitter clone is that the lower layers of that infrastructure could be outsourced to a "platform" company, such as Amazon Web Services (AWS) or Microsoft Azure. Those companies buy the machines and keep them up and running in their data centers, hosting the clients' software on them. Amazon and Microsoft charge for AWS and Azure, however, and both are profit centers for their respective companies, meaning that they are selling the hosting services at above their own cost. AWS and Azure are cost-effective for small and medium companies that don't want to manage their own data centers. However, at the global scale of a service such as Twitter, it would cost less to own and run the infrastructure itself (as Twitter does today) rather than outsource it, especially to a potential competitor, such as Amazon or Microsoft.

Even if it used AWS or Azure, an open-source Twitter clone would still need to manage its own services running on top of the underlying hosting service — deploying the software, rolling out updates, balancing the load across data centers, and monitoring ongoing performance. It would also need its own R&D staff to build and test the service. But its operations needs go far beyond that — it would need to hire several more teams to cover critical functions. This hypothetical company would require content staff, with multilingual and multicultural skills to monitor for offensive and illegal content, including child pornography, black market commerce, and copyright violations, and with customer support personnel to handle forgotten passwords; locked-out, suspended, and hacked accounts; and payment processing. Security experts would have to be on hand to prevent and respond to hackers, phishing schemes, viruses, and denial-of-service attacks.

A government and public affairs team would be needed to manage the geopolitical landscape and the many pitfalls of trying to run a global service, including how to refer to contested regions of the world, such as Taiwan and Tibet, and how to respond to national governments like China and Russia that require censorship of anti-government speech and block services that fail to comply.

And if an open-source Twitter clone service got that far, it would still find itself with the same problem that Twitter, Facebook, and other social network companies face, namely how to set policy on what is acceptable content when no one will be happy when theirs is the content being censored. At the end of the day, someone still needs to make policy, and an open-source system purportedly being run in the public interest still requires someone (or some group) to make those decisions.

Assuming all of those hurdles can be jumped, the service would still need to jump the biggest obstacle of all: getting a critical mass of people to sign up and use it. Social networks are subject to a principle known within technology circles as Metcalfe's Law, which says that the value of a network is equal to the square of the number of endpoints in that network, because every individual one has a potential direct connection to every other one. A social network with 10 people is worth 100; a network with 100 people is worth 10,000. When you hit 1,000 people, its worth grows to 1 million. Twitter, with its 215 million daily active users, is worth 46 quadrillion.

As a network grows, its value increases explosively, and it becomes almost impossible for a newcomer to catch up. People spend time on a social network, such as Facebook or Twitter, when there are people they want to interact with. Conversely, they don't spend time there when there aren't people there. When a network has many people, it's much easier to get more to join; when there are few people, it's much harder to get them to join (or to stay). The history of social networks is littered with examples of those that failed to reach a critical mass of users beyond which the "network effect" sustained its growth.

Companies try to drive growth in their early days by offering "bright shiny objects" — often celebrities — to attract attention and new users. Sometimes, they resort to deals where the company pays a celebrity, artist, or public figure to post exclusive content on their social network. Musicians are often a big draw, as we are seeing today with Instagram and TikTok, though that often feeds "generation gaps" where different age groups embrace different social networks. At the end of the day, the marketing challenge for a new social network is to create a brand for itself that tells people who will be there and then delivers on that promise.

Now, back to the question of funding. Launching a new social media network — even an open-source one — requires a significant amount of money for upfront costs and ongoing operational expenses. Twitter spent about $5.1 billion in 2021; an open-source venture not reliant on advertising would save a considerable amount on sales and marketing, R&D, and operations, but it would still be a very expensive business to run. It's instructive to look at perhaps the largest open-source organization: the Wikimedia Foundation, a nonprofit that runs Wikipedia as its primary activity.

According to its 2019 IRS filing, it had a staff of 291; revenues of $124 million, mostly from grants and contributions; and expenses of $112 million. More than $55 million of that covered personnel-related expenses. In comparison, Twitter had 7,500 employees in 2021. Wikipedia doesn't have nearly the scale and density of activity as Twitter does, and yet it still requires a substantial organization with a large budget. An open-source Twitter clone would need a much larger organization and budget, likely hundreds of millions of dollars every year.

Wikipedia is also instructive as to the content and policy challenges that a Twitter clone would face. Wikipedia has faced similar challenges in its own realm, including power struggles over editorial control and accusations that it spreads misinformation and is politically biased. Being open-source doesn't inoculate an organization from those issues. And, in the end, it comes back to the money: The people and interests who are providing the hundreds of millions of dollars in revenues to keep such a service running will inevitably have thoughts on the policy issues. Even if their influence is somehow curtailed, the general public will still believe that their money corrupts the policy and content decisions being made.

Where does this leave us? In theory, there is no absolute barrier to the creation of a new, open-source Twitter-like service for those who don't want to live under Elon Musk's thumb. But the financial, logistical, personnel, and "network effect" challenges are substantial — perhaps insurmountable. Even if a challenger overcame all of those issues, it would still have the same problem that Twitter faces today: Wherever it lands on its policies for acceptable content, some factions will find them unacceptable, and it will face allegations of bias, corruption, and spreading misinformation.

The South Seattle Emerald is committed to holding space for a variety of viewpoints within our community, with the understanding that differing perspectives do not negate mutual respect amongst community members.

The opinions, beliefs, and viewpoints expressed by the contributors on this website do not necessarily reflect the opinions, beliefs, and viewpoints of the Emerald or official policies of the Emerald.

Featured Image: Photo by Sattalat Phukkum/Shutterstock.com