I want to tell you about something new today: we’re launching the @qedScience Insight Feed!🥂 I’m excited about it and want to tell you what it does and why we built it. Part of it is that I miss Science Twitter 😌 https://x.com/i/article/2110325357920256000
Missing Sci-Twitter and introducing QED's new "Insight Feed"
I miss Science Twitter. I know I’m not the only one missing it. Scientists tell me this at every scientific meeting I attend. We lost something significant when it died. Although I might be talking to the void, I’m writing this here on X because this is where I met many of the people I want to say it to. I miss what we had here, and this below is my attempt at recreating it. I think we can save some of it and maybe even improve it (Bluesky hasn’t filled that gap for me).
I made tons of friends on Sci-Twitter. It led to collaborations and invitations that would probably never have happened otherwise. I also had bad fights here, but that happened mostly after 2023, and at that point, Sci-Twitter was already dying (for a couple of reasons), and X became a very nasty place that only bots and trolls enjoy. It’s sad that people in tech, news, or fields other than science don’t know that Sci-Twitter is dead. They still think X is the place to be, and it kind of gaslights scientists missing old Twitter.
Back in the good days, you could put an idea out there and people would respond because they found it interesting. The community was huge. This was true also when it came to “publishing”. Throwing your preprint on Twitter had an impact. You didn’t have to wait (years) for a journal to decide it deserved attention. I still preprint all my work, I can’t imagine going back to the dark ages before BioRxiv, but without Twitter, preprints don’t get the same exposure.
Sci-Twitter felt like talking during the coffee break at a conference (but more creative, and funnier). You’d hear about a paper you would otherwise have missed, from someone who could explain why you should care. Often the thread was enough to understand the main idea. People posted amazing threads… with memes and videos and whatnot. Then you’d read the replies and discover that someone had a problem with the experiment, and another someone would explain why it’s actually OK. Sometimes the authors would join in, and you could follow the argument and decide for yourself. The scientific process at its finest.
That was incredibly useful. There’s too much to read, even in your own small field. And some of the things that could change how you think about your research come from fields you don’t follow at all. Having other scientists point them out made a huge difference, and you learned who to follow for their recommendations. The algorithm worked. That’s not the case anymore. It just doesn’t.
I think almost all former Twitter users I know miss some version of this. Twitter became so contaminated with politics and violence and trolls and bots that finding the science started to feel like punishment. Every time I open it, I feel like Dumbledore drinking the poison in an attempt to destroy the fake Horcrux. Add the spam, and I end up regretting it every time.
But don’t despair!!! This week we’re launching QED’s “Insight Feed”. A lot of what I want it to become comes from missing the Sci-Twitter of the old days. It’s not a social network. Not now, at least. It’s just the science without the bullshit.
The Insight Feed is just where we’re starting. We want to add ways for people to connect around the work and develop collaborations. I’d love to help rebuild the community we used to have. A feed alone won’t do that, but it’s a start.
The immediate goal is to help you keep up with science that’s relevant to you, without getting spammed, and without the hate. That sounds simple, but getting it right is hard. You can already ask an LLM to find papers for you. It will happily give you things to read. But you then have to work out whether they’re worth your time, and keep explaining what you’re interested in. It’s easy to end up with another stream of recommendations you stop looking at. We’ve put a lot of work into calibrating this.
The Insight Feed uses your online research profile and your activity in QED to understand what’s relevant to you. What you read and write helps it follow your interests as they develop. It can recommend preprints or pick up something in a conference abstract that you would probably never have seen. Published papers (papers in journals) are part of it, but science happens in other places too, and we look beyond journals for work that could matter to you.
The content goes through QED’s filtering for quality and relevance. We tell you why every recommendation is relevant (what’s in it for you). Sometimes it won’t be obvious from the title or abstract, but the paper will contain a method that you need. Or maybe there’s one relevant claim that’s hard to notice in a plant paper and you, like most, ignore the plant literature.
The feed keeps updating at your pace. We want you to open it because you expect to find something useful, without feeling that you’ve acquired another obligation to keep up. And we want to help you understand what you’re looking at. Is there enough evidence to take the conclusion seriously? What’s the next step to gain more confidence? Or what’s the next experiment that YOU could do to take this in an original direction? In the (near) future, we’ll also help you make this happen.
That’s why QED exists: the scientific community outsourced too much judgment to journals. We use where something was published as a proxy for whether we should trust it. With so much increasingly complicated work being published, it’s becoming harder to assess the evidence ourselves. Journals fail at handling the flood. AI could help enormously here. But if it just produces more content for us to sort through, we’ll hit a wall.
QED puts tools for critical evaluation in scientists’ hands. When something in the feed catches your attention, you can dip into the work and interrogate the evidence. QED’s review tools can help you examine a weakness that might affect the conclusion. You can question its assessment too, and as always, your feedback matters. What you do in QED is private and secure (more about privacy and security on our website, qedscience.com).
There are lots of big promises being made about AI in science (and lots of worries). Some of the promises are exaggerated (we’re not going to cure all diseases any time soon). But I hope we can all agree (no way…) that AI that helps you find the new science you need to know and connect with other scientists is a good AI to have. I think that’s true also for AI tools that help you assess science properly.
So try QED’s Insight Feed now! qedscience.com