It is, you might say, a complicated moment for news online. There are the efforts to erode the First Amendment, the dominant platforms that aren’t sending traffic like they used to, the complexities of an ever-changing ad business, and on and on the list goes. Maybe most of all, there’s the rise of AI, and platforms that ingest an internet’s worth of news, abstract it away into a mush of semi-true information, and then serve it up to anyone who asks their chatbot what’s new. Into that fray comes Particle, a long-in-the-works new platform from a couple of former Twitter product leaders that is designed to help people find and make sense of the news a little more easily. With a lot of AI.

Particle’s plan is to use AI to do two particularly useful things. First, it organizes lots of articles and coverage into collections the platform calls “Stories,” so you can get lots of information and perspective on whatever you’re reading about. Some stories in Particle include more than 100 news articles, plus X posts, a section of salient quotes on the subject, and more. It’s a lot of stuff.

Particle also uses AI to summarize all those articles, right at the top of each story page. By default it offers a bulleted list of information, like you might expect from ChatGPT, but you can tweak the output in lots of ways through what the app calls “summary styles.” You can select “Opposite Sides” to get a readout of roughly the two viewpoints on Trump’s proposed new cabinet, say, or pick “Explain Like I’m 5” to have the latest Gaza developments explained in the simplest possible terms. Particle even has the app rewrite the headline for you in various ways, to make it simpler or funnier (or, in my experience, mostly just more confusing.) You can also just directly ask a question, and Particle’s AI bot will try and answer.

This kind of AI organization and summarization is everywhere in the app. When you first download Particle, the app has you Tinder-swipe your way through some headlines, signaling what you’d like to see more and less of in your feed. You can also follow specific publications or journalists, and see their stuff more prominently. Particle attempts to suss out the political leanings of each article and publisher, both to call out one-sided coverage and to attempt to find a balance.

Particle is a nice-looking and extremely information-dense app, and in my experience as a beta tester it has been a pretty useful way to get a quick overview of big issues. It’s also full of the same ideas that so many other companies have tried and failed. Apps like Circa couldn’t manage to build an audience and business out of the same kind of broadly useful summarization and aggregation. Discors had some neat ideas about structure that also didn’t amount to much. Snapchat and Facebook both once had news-aggregation dreams, and have left those behind. There just isn’t much evidence that an app like this can work.

Particle knows all this, of course. Sara Beykpour, the company’s CEO, spent more than a decade at Twitter and understands well the complexities of information sharing on the internet. Her co-founder Marcel Molina also spent years working on some of Twitter’s timeline and other services. They’re both convinced that the advent of AI makes it possible to do these things better, and at bigger scale. Particle says it has found ways to vastly reduce AI hallucinations and inaccuracies — some of which involve human editorial oversight — and has made deals with Reuters, Time, Fortune, and others to share their content. Getting those deals done, and getting them right, will be key for Particle’s staying power.

In the wake of a heated election, and on an internet where information and misinformation are both plentiful and nearly indistinguishable, Particle believes it has found a way to cut through. And is hoping that’s what people actually want.