Filtering the Information Network on X

The Neighborhood Map & Virtual Feeds

Abstract

Based off an idea of an ‘algorithm store’ from Jack Dorsey, the ‘Neighborhood Map & Virtual Feeds’ is a simpler visual-first solution that achieves the same effects with easier accessibility. By applying the social preferences of the human brain, we achieve systemic balance with local information networks. The core benefits are disrupting propaganda targeting and improving the signal-to-noise ratio across the network.

Understanding the Human Problem

X can disrupt propaganda and refine conversation by applying the science of community. Our brains are striving for trusted local discussion within neighborhood boundaries. For example, Times Square isn’t the best place for finance conversations - that’s more Wall Street - and the best Jazz is in West Village. But now all we have is one choice: the Stadium.

It creates a specific type of exchange where the biggest idea army wins, positions are reduced to (false) dichotomies, and mockery/noise is the most effective tactic for defeating the opponent. This is limiting. We must spread out the conversation to: hear nuance, make it harder to target people/groups, and achieve a higher signal to noise ratio.

I think the greatest idea — which we need — is an algorithm store, where you choose how you see all the conversations. -Jack Dorsey

We do need more options for how we see the conversation. The problem with an algorithm store is its limited accessibility - its too complex to understand/modify for the average user. For such a human problem, the solution needs to be a simple and natural idea with a visual UI for higher adoption so to improve the network-wide conversation.

Constraints of Current Solutions

We already have different content areas of X: crypto, sports, AI, etc. but they interact like sections within the Stadium rather than communities with clear boundaries; because sections have scaled too large to keep trust & reputation intact and its more profitable for engagement to exploit the masses rather than maintain a reputation with quality work.

Community notes (should be called Public Notes) only offers a marginal cost to producing bad content; but it is a step in the right direction. Communities (should be called Forums) achieve this non-systematically but limit the rapid spread of information - X’s main value proposition. What’s missing is niche networks that consistently exchange with each other.

A Visual-First Solution

Any significant solution for X will disrupt how we all play the game; and that’s why they have rarely been tried. Incentive structures will change, information flows will reroute, and the type of content (popular or unpopular) will calibrate to the new game. The Algorithm Store would do this - as would my proposed solution. I will outline its systemic effects as benefits and disadvantages, as well as some open questions.

The Neighborhood Map & Virtual Feeds

The Neighborhood Map is a visual-first network graph of neighborhoods (groups of connected accounts). Users can ‘travel’ to different neighborhoods with a Virtual Feed and lets them customize their view easily and in new ways - like Google Maps or Zillow filter features.

Imagine bohemian “high activity areas” (a new way of viewing what’s trending) to explore with just a click. A combination of filters like hashtag, location, account size, etc. will reveal a new visual map, which when clicked presents a Virtual Feed as if you were following those accounts. Currently, it’s not just difficult but penalized if you unfollow/follow many different accounts.

Prominent neighborhoods (hollywood celebrities, sportsball players, US politicians) can be named by manual input to give the visitor a better sense of direction of where the information is flowing and how to get from one place to another. Small but niche neighborhoods (where nuance thrives) will be more easily discoverable through connections between niche accounts. Neighborhoods can reshape over time, so a historical view is imperative.

For more advanced users, they can use different measurements of centrality.

In order to protect local neighborhoods from being overwhelmed by foreign opinions, there needs to be a distinction between Residents (true followers) and Visitors (virtual followers); first labeling them and then prioritizing Resident replies over Visitor replies - which are submitted when viewing through a Virtual Feed.

Benefits

The game on most social media, including X, is to create the most widely applicable (or popular) content as possible to accumulate social currencies — likes, comments, shares, followers, etc. To get started on this journey, common advice says to start with a niche, but here is where it changes.

A niche of The Stadium is different than a niche of a Neighborhood; and one much more difficult to discern as sections in The Stadium are relatively large and attention competition is high. But in a Neighborhood niches are not just in lower competition, but there are more of them overall and they are increasingly specific. Now you start in a Neighborhood, then eventually graduate to the Stadium.

Ultimately, the effect is higher signal with lower noise. This solution boosts the discoverability of the uncommon and unpopular (where all great ideas originate) and alleviates the noise of “fluff accounts” (ie. content repackagers, thread makers, and ideological propagandists) as these are late-stage symptoms of a social system that is failing to find new frontiers to discover. Fluffers will no longer have the same effect when people seek to explore the backstreets and alleyways of this new map; because the amount of people in the general feed would decrease (increasing competition) and they will have to recalibrate to new ways of appealing to the attention economy by going more niche and specific.

Propagandists will also be disrupted as targeting methods will have changed. Decentralizing the network makes their messaging increasingly inefficient and they’ll have to expand their number of operators for wider presence in more niche neighborhoods. But in neighborhoods where more people know each other, visitors and other foreign actors are put at a disadvantage. Trust will be sticky to those with reputation and that will be hard to generate if your only act is exploiting the community for engagement farming. Local (within degrees of separation) action will be effective again.

Refining conversation isn’t about lessening the conversation, it’s about retaining signal and eliminating noise. It’s about keeping the discoverability of new ideas - especially the uncommon and unpopular - without immediately pitting them against the mainstream status quo. It’s about limiting the amount of foolish insolent talk, which requires an order of magnitude to offload as it does to produce, by having more severe penalties for producing it (local reputation leading to ostracization). This is how human network communication has always worked, and we need to return for balance again.

Disadvantages

Great improvements disrupt. Here’s how the former system will be disrupted and what we can do to balance the offsets. An F-117 is not aerodynamic, but with enough small adjustments it can achieve greater functionality than aged aerodynamic aircraft. Great innovations do not follow the rules of the previous system.

This solution will disincentivize users from actually following accounts as they can use the virtual feed to frequent content areas. Virtual feed users can “save” a feed to visit instead of collecting too many accounts for the algorithm to maybe show later on. If accounts get less followers that turn into engagement, they will have to find neighborhoods to thrive in, changing the type of niche content they produce.

This solution will also change how ad targeting works by focusing on groups of users rather than specific users. It may temporarily disrupt revenue, but if X wants to become more self-sufficient on user-driven revenue, it’s a great opportunity to do so. Like Grok, users pay a subscription to access features that offer an order of magnitude greater utility. People pay for an advantage; especially in an information war. Therefore, make this solution a premium-only feature to shift the vulnerable revenue from advertisers to more secure revenue paid by users.

A cost benefit analysis can look at the average revenue a user brings in from ad revenue, against a monthly subscription. But more importantly is to put a price on increased network efficiency - drawing more users to the platform and increasing the average amount of time spent online. These are some of the many open questions that a deep discussion with the team can discern answers and next steps to.

From Here to Implementation

If you’re reading this and work for X, then let’s have a talk. I’m happy to expand some elements and/or discuss systemic effects. I can show you mock up UI designs about how this could look, but I would also need to get more information from you. I’m based in Austin, TX. Please contact me via X: https://x.com/alexandercurves

If you can implement this yourself, then go on and get to it. But the only thing you cannot do is nothing. X is dying from its users following the current system’s function. It’s needs change. The future is either bright and full of gold, or a whimpering candlelit flame that fades in the darkness. You must try something; and I recommend this. Godspeed.