Sara Beykpour helped build the feed. Now she's building its cure.

For most of the last fifteen years, the news arrived the same way: as an endless column of posts, refreshing faster than anyone could read it. Sara Beykpour helped design that experience. She spent more than a decade at Twitter, working on the products that taught a generation to scroll - Vine, Periscope, and eventually Twitter Blue. Few people have shaped how modern feeds feel as directly as she has.

So there is something worth noticing in the fact that she left, and that the company she built afterward is, in many ways, an argument against the thing she helped popularize. Particle, the AI news app she co-founded in 2023, is not another infinite timeline. It is a deliberate attempt to make the news finite again - to give a reader the day's important stories, summarized, with a clear view of how each one is being covered across the political spectrum.

A decade inside the feed

Beykpour studied software engineering and cognitive science at the University of Waterloo, a pairing that reads, in hindsight, like a job description for building consumer software. One half is how to make the machine work. The other is how people actually think. She joined Twitter in the early 2010s and stayed through the era when the company was defining what social video and real-time media looked like.

Her fingerprints are on a run of products that each marked a moment. Vine compressed video into six looping seconds and launched a generation of creators. Periscope made live streaming from a phone feel ordinary. Twitter Blue, which she led as a senior director of product management, was the platform's first serious swing at subscriptions and a reworked advertiser experience. Three products, three different eras of the social internet.

The through-line across all of it was proximity to the news. Twitter was where breaking events landed first, where a plane on the Hudson or an election night played out in real time. Beykpour and her eventual co-founder, Marcel Molina - a senior staff engineer who worked at both Twitter and Tesla - spent years at that intersection of media and product. When they left to start something of their own, the subject matter was almost predetermined.

The two-month pivot

The founding story has a detail that says a lot about how Beykpour works. Particle started in February 2023 as a social product. Within about two months, the founders killed that idea and pivoted to news. That is not a small decision - it means throwing out weeks of work and the original pitch - but it's the kind of move that separates product people who are attached to features from ones who are attached to problems.

By April 2023 the company had raised a $4.4 million seed round, backed by Kindred Ventures, Adverb Ventures, and angels including Twitter and Medium co-founder Ev Williams and Behance founder Scott Belsky. A little over a year later, in June 2024, Lightspeed Venture Partners led a $10.9 million Series A, with Axel Springer - the publisher behind Business Insider and Politico - joining the round. Total funding sits at roughly $15.3 million.

Same story, every side

What Particle actually does is easiest to understand through the feature Beykpour keeps returning to. Rather than serving a single article, the app clusters coverage of a story from many outlets, uses AI to summarize it, and then shows the reader how that story is being framed from different points on the political spectrum. A reader can see, at a glance, that the left-leaning version and the right-leaning version of the same event share the facts but not the emphasis.

Beykpour describes the design goal as sustainability, not just for readers but for the people producing the journalism. Where many AI news tools scrape publishers for raw material, Particle's early move was to sign Reuters as a partner and experiment with monetization that sends value back to newsrooms. It's a slower path, and a deliberate one.

Her stated mission for the company is disarmingly broad: to help everyone understand more about the world faster, through AI, starting with the news. The "starting with" is doing a lot of work in that sentence. News is the wedge, not necessarily the whole ambition.

Building at AI speed

Beykpour has been candid about how much AI has changed the mechanics of building software. At Twitter, a feature could take a quarter. Now, she has said, the same kind of work collapses from months into hours, and small teams can achieve what used to require large ones. Particle runs lean - a team of around 19 people - and that is partly the point. The tooling that lets a tiny company take on the entire news industry is the same tooling the product is built on.

What she's actually betting on

There's a tidy irony to Beykpour's career, and she doesn't seem to run from it. The infinite feed was one of the defining product ideas of her professional life. Particle is a wager that people have had enough of it - that the next valuable thing in news is not more, faster, but less, clearer. Coming from someone who helped invent the former, the argument carries more weight than it would from an outsider.

Whether a 19-person company can reshape how the world reads the news is an open question. The category is famously brutal, and plenty of well-funded attempts have come and gone. But Beykpour has a track record of shipping products people end up living inside, a co-founder she has trusted for years, and a thesis specific enough to be testable. For anyone watching where AI meets journalism, Particle is one of the clearest experiments to keep an eye on.

For now, the pitch stays simple, and she repeats it like a mantra: understand more of the world, faster. It started, of all places, on a feed she helped build - and it points, now, at something quieter.