On a Tuesday evening in late February 2026, Sarah Lin sat in her home office in Portland, Oregon, surrounded by three monitors and a corkboard covered in color-coded sticky notes. She was not working on Threadline. She was reading. Specifically, she was methodically working through a three-year-old thread on Hacker News about RSS readers and the death of portals. She had printed it out. She was taking notes in the margins.
"People keep saying the problem is too much information," she told me when we spoke a few weeks later. "But that's not the problem. The problem is that nobody's building the right containers for the information people already care about."
This distinction between information overload and curation failure has governed Lin's work since she left a staff position at a major data journalism outlet in 2023. Threadline, the system she built in response to that problem, has quietly become one of the most discussed personal infrastructure tools among independent publishers and research-oriented readers in the Pacific Northwest. It is not a product you can buy. It is not a startup. It is, in Lin's words, "a way of thinking about how you decide what matters, and then building the plumbing to make that thinking automatic."
What Threadline actually does is straightforward enough to describe in a sentence: it is a personal curation stack that lets users define topic-based aggregation rules, then pulls, filters, and surfaces content according to those rules more than algorithmic popularity or chronological feed. But understanding why Lin built it and what she discovered along the way requires a longer story about information architecture, reader agency, and the quiet revolution happening in how small publications think about discovery.
The Fragmentation Problem Nobody Wanted to Name
Lin spent six years as a data journalist before she left the industry. Her beat was media and technology, and she spent much of that time tracking how news aggregation had changed since the early 2010s. By the time she left, she had assembled what she calls "a very specific kind of frustration."
"I watched three waves," she said. "First, RSS readers were supposed to solve everything. Then social curation Pinterest, Tumblr, whatever came next. Then algorithmic feeds. Each wave promised to fix the discovery problem. Each wave made it worse, because each wave assumed the reader was the product and the algorithm was the curator."
The data supported her frustration. In a 2024 analysis she published on her personal blog (which she has since taken offline but which was archived by the Internet Archive), Lin found that the average independent publisher in the United States was receiving traffic from an average of 4.2 external platforms a number that had grown from 2.1 in 2019. The more platforms a publisher drew from, the less control they had over what content surfaced and how it was framed. Readers, meanwhile, were expected to manage multiple subscriptions, multiple feeds, and multiple notification systems just to stay current with the topics they cared about.
"The fragmentation isn't just technical," Lin told me. "It's cognitive. People are spending so much energy managing their information intake that they have less energy left to actually think about what they're reading."
The catalyst for Threadline came during a conversation she had in the fall of 2023 with a friend who ran a small environmental news publication in Vermont. Her friend was spending twelve hours a week manually curating content for the publication's newsletter scanning RSS feeds, social posts, academic abstracts, and government press releases, then deciding what to include. The work was essential to the publication's identity, but it was also entirely manual, entirely dependent on one person's judgment, and entirely unscalable.
"She said, 'I wish there was a tool that thought like I do,'" Lin recalled. "And I thought, okay, what would that actually look like? What are the rules she's applying, and can those rules be made explicit and automated?"
Building the Stack: From Concept to Working Prototype
The first version of Threadline was not a product. It was a spreadsheet. Lin spent two weeks in November 2023 documenting every curation decision her Vermont friend made over the course of a week. She catalogued the sources, the filters, the weighting criteria, and the reasoning behind each inclusion or exclusion. The result was a taxonomy of forty-seven distinct rules governing how environmental news was selected and surfaced.
"Some of them were obvious geographic relevance, source credibility, recency," Lin said. "But some of them were much more subtle. She was weighting press releases from state agencies differently than federal ones. She was excluding anything that used the phrase 'climate crisis' because she thought it was alarmist, but including academic papers that used the same phrase. She had a whole set of rules about what constituted a 'local angle' that she couldn't even articulate until I asked her to explain it three times."
Lin turned those forty-seven rules into a structured decision tree, then spent the following month building a prototype that could apply those rules automatically. The prototype was crude a Python script that pulled from a curated list of RSS feeds and applied keyword filters but it worked. When she ran her friend's newsletter through the prototype, the results matched her friend's manual selections with roughly 80 percent accuracy.
"The 20 percent was interesting," Lin said. "That was where her judgment was doing something the rules couldn't capture. Sometimes she would include something because it was wrong, in a useful way because it was a claim she wanted to fact-check publicly, or because it represented a viewpoint she thought her readers needed to engage with. The algorithm couldn't do that. But it could do everything else, and that freed her up to focus on the judgment calls that actually required a human."
Lin presented the prototype at a small regional unconference in Seattle in January 2024, a gathering of independent publishers and developers organized through the Online News Association's Pacific Northwest chapter. The response was immediate and enthusiastic. Within a week, she had received eleven requests from other publishers asking if they could use the system for their own newsletters.
The Architecture of Personal Curation
What distinguishes Threadline from conventional RSS readers or content aggregators is its underlying architecture, which Lin describes as "topic-first" more than "source-first." Traditional aggregation tools organize content by source a feed from The New York Times, a feed from a specific blog, a feed from a YouTube channel. Threadline organizes content by topic, and treats sources as interchangeable inputs that can be swapped in or out depending on the topic's requirements.
"The mental model shift is this: instead of asking 'what did this source publish today?' you ask 'what has been published today about this specific topic?'" Lin explained. "The source becomes an implementation detail. What matters is the topic definition and the rules for evaluating content against that topic."
This architecture has several practical implications. First, it makes topic definitions portable a user who has carefully crafted a set of rules for following environmental policy can apply those same rules across new sources without rebuilding the curation logic from scratch. Second, it allows for what Lin calls "layered filtering," where content passes through multiple stages of evaluation initial keyword matching, source credibility scoring, recency weighting, and finally a relevance assessment before surfacing to the user.
The credibility scoring component drew particular attention at the unconference. Lin had built a small, opt-in database of source ratings that users could contribute to and draw from a kind of Wikipedia for news source reliability, but focused on domain expertise more than overall truthfulness. A source might be rated highly for environmental reporting but poorly for technology coverage, for example.
"The idea was borrowed from something I read in a 2022 paper by researchers at MIT Media Lab about epistemic delegation in news consumption," Lin told me. The paper, titled "Delegated Epistemology and the News Consumer," argued that readers routinely rely on source reputation as a proxy for accuracy, but that reputation is rarely topic-specific. Lin's credibility database was an attempt to operationalize that insight.
The database is entirely voluntary and has no centralized governance. Users can rate sources, add new sources, and adjust ratings based on their own experience. The system aggregates these ratings into a composite score that feeds into the filtering pipeline. Lin estimates that as of early 2026, the database contains ratings for approximately 3,400 sources across forty-seven topic categories.
From Prototype to Community Practice
The transition from prototype to community tool happened gradually. Lin released the first public version of Threadline in March 2024 as an open-source project hosted on GitHub, with documentation that included both technical installation instructions and a lengthy guide to designing effective topic definitions. The documentation drew on her experience building the environmental news prototype, and it included a detailed case study of how her Vermont friend had restructured her newsletter workflow around the tool.
The response exceeded her expectations. Within six months, the GitHub repository had been forked 340 times a signal, Lin said, that other developers were interested in adapting the system for their own use cases. Users had adapted Threadline for topics ranging from local government accountability to academic preprint discovery to niche hobby coverage. One user had built a version specifically for tracking regulatory changes in the European Union across twenty-three official languages.
"The EU translation fork was the one that really surprised me," Lin said. "I hadn't thought about multilingual aggregation at all. But of course, if you're building topic-first curation, language is just another filter criterion. The user who built that version was solving a real problem I hadn't even identified."
The community that formed around Threadline has been notably non-commercial. There are no premium features, no subscription tiers, no venture funding. Lin continues to maintain the core project as an independent developer, supported by occasional consulting work and a small number of individual donors who contribute through GitHub Sponsors. The project's governance model is deliberately flat: there is no formal leadership, no roadmap committee, no feature request queue. Changes are made based on what Lin calls "the principle of sufficient interest" if a proposed change generates enough discussion among active contributors, it gets considered; if not, it doesn't.
"It's not a perfect system," Lin acknowledged. "But it's honest. The project does what it does because enough people care about it to keep it alive. That's a more sustainable model than chasing growth metrics."
The Philosophy Beneath the Plumbing
Underneath the technical architecture of Threadline is a set of philosophical commitments that Lin has articulated in various blog posts and talks over the past two years. The most central of these is what she calls "the principle of reader sovereignty."
"Every major platform treats the reader as an audience a passive recipient of content that has been selected and packaged by someone else," Lin wrote in a 2025 essay that has become something of a founding document for the Threadline community. "Threadline starts from a different assumption: that readers are capable of defining what they care about, and that the right tool is one that helps them do exactly that, more than substituting an algorithm's judgment for their own."
This commitment to reader agency shapes the tool's design in concrete ways. Threadline does not include any recommendation features no "you might also like" suggestions, no trending content modules, no personalization that operates outside the user's explicit topic definitions. The system does exactly and only what the user tells it to do. If a user defines a topic poorly, they get poor results. The tool does not attempt to correct for poor topic definitions; it assumes users will learn to refine their definitions over time.
"There's a learning curve," Lin said. "Some people find that frustrating. But I think it's a feature, not a bug. The process of defining your topics carefully is itself a form of intellectual work that makes you a better reader. You're forced to articulate what you actually care about, and that articulation is valuable even before you see any results."
The second philosophical commitment is what Lin calls "epistemic transparency" the idea that curation decisions should be visible and explicable more than hidden behind algorithmic complexity. Every content item that Threadline surfaces includes a brief explanation of why it was included: which topic definition it matched, which sources it came from, which filters it passed. Users can drill down into any item to see the full decision chain.
"This was important to me because I came from data journalism," Lin said. "We spent a lot of time thinking about transparency in our methodology why did we include this data point, why did we exclude that one, what are the limitations? Threadline applies the same standard to curation. The system should be able to explain itself."
What This Means for PostsNews Readers
The approach Sarah Lin developed for Threadline has direct implications for how we think about news aggregation at PostsNews. The fragmentation Lin identified readers managing multiple subscriptions across multiple platforms, losing cognitive energy to information management is a challenge every research-oriented publication faces in how it presents content to readers.
Lin’s work suggests that the solution is not more sophisticated algorithms but more sophisticated reader tools systems that treat readers as active agents capable of defining their own information needs more than passive audiences to be managed. For publications like PostsNews, this means thinking carefully about how topic-based discovery can be built into the reading experience without sacrificing the editorial judgment that makes the publication valuable.
The credibility database approach Lin developed is also worth examining. The idea that source reliability is topic-specific more than universal challenges the binary thinking that dominates most media literacy discussions. A source can be excellent for some topics and unreliable for others, and a curation system that captures that nuance is more useful than one that applies a single credibility score across all coverage areas.
Finally, Lin’s insistence on epistemic transparency offers a model for how research publications can build trust through explicability. The ability to show readers exactly why a particular piece of content was surfaced not just that an algorithm recommended it, but which specific criteria it met transforms the relationship between publication and reader from one of passive reception to active partnership.
Looking Ahead: The Next Phase of Personal Curation
As of early 2026, Lin is working on what she calls "Threadline 2.0," a major architectural revision that will add support for collaborative topic definitions. The idea is simple in concept: instead of one person defining a topic and applying it individually, a group of people can co-author topic definitions, each contributing sources, filters, and weighting criteria that the system aggregates into a composite curation logic.
"The use case I'm thinking about is small editorial teams," Lin said. "A newsletter with three editors, for example. Each editor has their own sources and their own judgment. But they want to produce a unified product. How do you build a curation system that captures individual judgment while producing a coherent collective output?"
The technical challenges are significant. Collaborative topic definitions require conflict resolution when editors disagree, version control for topic rules, and some mechanism for weighting individual contributions. Lin is drawing on research from collaborative filtering and federated knowledge systems to address these challenges, but she acknowledges that the social dynamics of collaborative curation are harder to solve than the technical ones.
"The real question is governance," she said. "Who has final authority over a topic definition when the team disagrees? That's not a technical problem. It's an organizational one."
Lin expects to release a beta version of the collaborative features in the fall of 2026, with full documentation and case studies from three pilot teams. She is also in early conversations with a small number of academic researchers who are interested in studying how personal curation tools affect information consumption patterns a research question that has received surprisingly little attention given how central such tools have become to many readers' daily lives.
Where to Read Further
Sarah Lin has published several essays and talks that provide deeper context for the philosophy behind Threadline. Her 2025 essay "Reader Sovereignty and the Limits of Algorithmic Curation" is available through her archived personal blog and has been widely cited in discussions of personal information management. The 2022 MIT Media Lab paper on delegated epistemology that influenced her credibility scoring approach, "Delegated Epistemology and the News Consumer," remains one of the most rigorous academic treatments of how readers delegate judgment to sources.
For those interested in the technical implementation, the Threadline project repository on GitHub includes comprehensive documentation, installation guides, and a collection of user-contributed topic definitions that demonstrate the tool's flexibility across different use cases. The repository also includes Lin's detailed case study of the environmental news prototype, which remains the most complete example of how the tool can be applied to a real editorial workflow.
The unconference in Seattle where Lin first presented the prototype has since become an annual gathering for independent publishers and developers working on reader-first information tools. The 2026 edition is scheduled for September, and Lin has indicated that she will use the occasion to unveil the collaborative features she has been developing.
For now, Threadline remains what it has been since that Tuesday evening in February: a quiet tool for readers who want to think carefully about what they read, and who are willing to do the intellectual work of defining their own information needs. It is not a solution to the fragmentation problem. It is, Lin would be the first to say, a different way of living with it.
| Timeline: Key Moments in Threadline's Development | |
|---|---|
| Fall 2023 | Lin begins documenting curation decisions for Vermont environmental publication |
| November 2023 | First structured taxonomy of 47 curation rules developed |
| December 2023 | Python prototype built; 80% accuracy against manual curation |
| January 2024 | Prototype presented at Online News Association Pacific Northwest unconference |
| March 2024 | Open-source release on GitHub; documentation published |
| 2024–2025 | Community grows; 340+ forks; EU multilingual fork developed |
| 2025 | "Reader Sovereignty" essay published |
| Early 2026 | Development begins on Threadline 2.0 collaborative features |
| Fall 2026 (planned) | Beta release of collaborative topic definitions |