The reading room · Digna Legi
And You Will Know Us by the Company We Keep
/100
80–100: high value. 70–79: worth the time. Below 70: below the usual publication threshold.
Evidence-reviewed score based on available publisher text. The evidence is sampled with gaps, so the full continuity and completeness of the essay cannot be fully assessed.
Scores reflect one reader’s profile, not an objective quality rating. Best is a separate personal selection.
How scoring works →This brief · about 3 min with detail
Why read this
Wei argues that feed quality follows from graph architecture: social graphs create different failures than interest graphs.
AI brief · Checked against source text
The main idea
The essay argues that Western social apps made a consequential architectural bet by tying the content feed to the social graph. That choice creates cold-start pressure to make users follow people quickly, then later produces mismatched feeds, context collapse, and hard-to-reverse social dynamics because people rarely prune their networks. Wei’s central distinction is between a social graph, built from people and relationships, and an interest graph, built from what a user actually wants to consume.
Go a little deeper
Graph choices become one-way decisions
Wei’s strongest mechanism is reversibility. A product can change buttons, ranking, or formats, but once users have accumulated followers and follows, the graph becomes socially sticky. Unfollowing is awkward, public enough to matter, and emotionally loaded; follower growth also changes what people dare to post. That makes early growth tactics more like architecture than marketing.
Interest is poorly approximated by acquaintance
The essay separates liking a person from wanting all their output. A user may want one topic from an account and reject another, but account-based feeds bundle them together. At scale, this turns ordinary variation into feed noise, and the app often responds downstream with ranking algorithms rather than upstream with a better way to source content.
Context collapse changes production, not just consumption
Graph design affects both sides of the feed. When Facebook expanded from a bounded school network to everyone in a person’s life, the audience became harder to imagine and harder to address. The problem was not merely that users saw unwanted posts; it was that creators faced an implied audience spanning incompatible social contexts.
Subgroups are a product primitive
Wei connects social health to whether an app lets people form appropriately sized groups. Monolithic feeds simulate a public commons, so users behave as if anyone could judge them. Messaging apps and groups, by contrast, let users sort interaction by context. The Slack example extends the point beyond consumer apps: changing communication topology changes institutional politics.
A case from the article
Pinterest’s early skew became self-reinforcing
Wei uses Pinterest to show path dependence. Because the feed drew from each user’s social graph, men joining early saw many pins from women, who were strong early adopters. That made the product appear female-centric, discouraging some male users and reinforcing the skew. His point is not about Pinterest’s category, but about how feed sourcing can convert an initial imbalance into identity.
How the case is made
The case is made through product observation, historical comparison, and named examples from social apps including Facebook, Twitter, Instagram, TikTok, Pinterest, WeChat, and Slack.
Where the idea has limits
The essay is strongest as a design diagnosis, not as a settled taxonomy: Wei explicitly says graph design has “more open questions than answers,” especially because scaled social networks are hard to prototype.
A question to take away · from Digna Legi
Where does your product force one audience, identity, or graph to do work that should be split across contexts?
What the original adds
The original adds many product-specific wrinkles, including contact-book onboarding, suggested-follow subsidies, algorithmic feed patches, topic pickers, and the idea of using graph design offensively to encode unique intelligence.
About this brief
AI-written, then separately checked for source support, useful detail and clarity. The author’s claims and our editorial question are kept separate. The original remains the author’s work. How we select and summarise →
How was this brief?
Rate this summary, separately from the author’s article.
Optional. Saved in this browser; shared only if you allow analytics.
How was the original article?
Rate the author’s original after reading it.
Optional. Saved in this browser; shared only if you allow analytics.
Digna legi. Worth reading.