Writing Monks allows a group of people to create a shared Twitter account. Shared Twitter accounts have the same advantages as media companies: multiple people can create better content than one person, which is why newspapers have multiple journalists. It also has many of the same needs as media companies: it needs a way to incentivize content creators and a way to select the best content to publish. Writing Monks addresses these two needs in a new and better way.
In this article, I will describe in detail the problem that Writing Monks (WM) is trying to solve. What the current solutions are and why Writing Monks’ solution is better.
Determining which content is good and which is not is really difficult. This is especially true for long-form content. Even the Harry Potter series, which is brilliant, was rejected by 12 different publishers before it was actually published. The traditional publishing process, in which the author first sends a summary (about 2 pages long) to a literary agent, and if the agent likes it, they ask the author to provide a partial manuscript (about 30 pages long), and then finally the full manuscript, is too lengthy and expensive to scale to our current content platforms (such as Twitter, Reddit, Medium, newsletters, etc.).
Current online platforms decide which content to show users based on two factors:
- The number of likes/upvotes/downvotes the content receives.
- The people the user follows.
In summary, the amount of attention a piece of content receives depends partly on its quality and partly on the author’s following.
As an online content creator, the best thing you can do is to increase your influence and find your followers. This requires you to post frequently. The best blog post about shrimp farming or the best opinion on some current political event is unlikely to create the next Harry Potter. Good content is usually produced by people who have been accumulating insights for years and then distill them into an article every once in a while.
On platforms like Hacker News, users do not follow specific authors, so the platform’s goal is to highlight only the best content, regardless of the author. They achieve this by using upvotes and downvotes. This system has been very successful but has its flaws. Users only want to read the most popular posts, so there is a high risk in reading a brand new post (with no upvotes or downvotes). This makes the first vote after a post is published very important, and because of this, it is easy to manipulate the front page display through early voting fraud.
First, we need to define what constitutes “good” Twitter content. It can be superficially said that the post with the most likes is the best. However, for a large and diverse enough audience, the content that performs the best always revolves around cats and dogs – cute and adorable, but not very nutritious.
What Writing Monks wants is to find the best content for a specific topic. On Writing Monks, each shared Twitter account (also known as a publication) has guidelines that specify what topics can be posted and what types of content are prohibited. The team that owns the publication will select a group of moderators to enforce these guidelines. So far, this is similar to what happens on Reddit or Hacker News (HN). However, compared to Reddit and HN, Writing Monks’ approach can scale the behavior of moderators and incentivize users to find the best content as early as possible.
Whenever a team member suggests publishing a piece of content, Writing Monks creates a prediction market to predict “how many likes this article might get 24 hours after it is posted on the shared Twitter account.” The market is Writing Monks’ best tool for predicting the future. If you want to know which company will have more cash flow in the next 20 years, look at the stock market; if you want to know who will win the upcoming boxing match, look at sports betting; if you want to see well-performing content on Twitter, look at the accounts managed by Writing Monks (laughs).
Any team member can place bets in the prediction market. If a bet is large enough, it will change the market’s prediction. The team can maximize profits by shifting the market from inaccurate estimates to accurate estimates. Once the market estimate is correct, no one can make it “more correct,” so no one can profit anymore. This is how we incentivize users to find good content as early as possible.
Predictive markets also allow moderation to be scaled. Moderators no longer need to review every post. They can simply look at the predictions and only focus on the posts that are expected to receive the most likes. Each post has three days to be published. If they are not published, each bettor will get their money (tokens) back. Incentivizing bettors not to bet on posts that violate the publication’s guidelines, as the moderators will mark these posts and the bettors’ money will be locked up for three days.
Vitalik Buterin proposed a similar idea in this
