iiinova — where your digital self keeps evolving

QED Science @qedscience.bsky.social · 07/09/2026

New @science.org feature: About a quarter of AI’s criticisms are missed by human reviewers. Does that mean we should trust AI to replace human judgement? Only if it’s fit for purpose. Read how Science gets real about the current state of AI in peer review. tinyurl.com/5n6bewkv

tinyurl.com

Can AI help solve the peer-review crisis? Here are its promises and pitfalls

Amid a shortage of human reviewers, tests indicate AI can assist—but is not a savior

000

QED Science @qedscience.bsky.social · 31/08/2026

He ran Blogger for half a billion readers a month at Google. Now Guy Ronen, QED's VP of Product, is tackling something harder: "We want to help scientists validate research and push scientific discovery forward," he told Pulse 2.0. → tinyurl.com/34suy8dh (Insights Feed launches soon 👀)

QED Science @qedscience.bsky.social · 28/08/2026

We asked webinar attendees which AI tools they use for grant review.* 52% ChatGPT · 40% Claude · 33% QED Science · 17% Gemini. Researchers aren't picking one tool. They're building their own mix around the grant process. Full summary → tinyurl.com/3aa78jf5 *Multiple answers allowed.

311

QED Science @qedscience.bsky.social · 27/08/2026

Would your proposal survive a reviewer's first read? The takeaway from last week's webinar (and more) were turned into a practical checklist: step-by-step pre-submission review and specific guidance for 2026–2027 ERC applicants. 📥 Download → bit.ly/4xYf0Sm

QED Science @qedscience.bsky.social · 25/08/2026

"If you write your risk section as a straight line, it looks naive." Dr. Angela Andersen, Co-Founder, Life Science Editors. Every reviewer knows science is messy. Your proposal should show you know it too. Full recording here: bit.ly/4giP5Pu

QED Science @qedscience.bsky.social · 20/08/2026

Today's webinar on grant review wrapped with a packed Q&A - the biggest themes: responsible AI use and keeping scientific judgment with the researcher. Recording + full Q&A available next week. Thank you @angelaandersen.bsky.social and Prof. Barak Rotblat 🎓 🔔 Follow to get notified.

QED Science @qedscience.bsky.social · 18/08/2026

900+ researchers have already registered for our webinar. Join us this Thursday. Learn how to find weaknesses in your grant, strengthen feasibility and use AI as a critical reader without treating it as the scientific authority. 📅 Aug 20 ⏰ 11 AM ET 🔗 Register here: tinyurl.com/ywn6evs8

170

QED Science @qedscience.bsky.social · 16/08/2026

What if the biggest problem in peer review is what never makes it into the paper? Our CEO Niv Mastboim on failed experiments, missing controls, and incomplete evidence, and why scientific review needs a fuller picture. 🎥👇

040

QED Science @qedscience.bsky.social · 16/08/2026

What if the biggest problem in peer review is what never makes it to the paper? Our CEO Niv Mastboim on failed experiments, missing controls, and incomplete evidence, and why scientific review needs a fuller picture. 🎥👇

QED Science @qedscience.bsky.social · 12/08/2026

Happy to see the results of our cooperation with @elife.bsky.social. In this pilot, eLife reviewers used QED’s claim-mapping technology and rated 80% of the comment-to-claim assignments as accurate. Clear evidence that structured AI and expert judgment can work together.

QED Science @qedscience.bsky.social · 11/08/2026

Also, @barakrotblat.bsky.social wrote a piece around this exact idea: the claim as the basic unit of science, and whether the evidence really supports it. Feels very close to what you’re pointing at here.

QED Science @qedscience.bsky.social · 11/08/2026

Very much the same motivation: identify the claims, map them to the evidence, and ask whether they actually hold up. Not category theory, but definitely less rhetoric, more structure 🙂 Nature covered our approach here: bit.ly/4wj2Xhk

bit.ly

This AI tool claims to pick the top 1% of preprints. Should researchers trust it?

QED Science says that its metrics reduce bias by assessing papers solely on the basis of their originality and validity.

111

QED Science @qedscience.bsky.social · 10/08/2026

Check out this recent interview of our CEO, Niv Mastboim on Nature. Special thanks to @miryamnaddaf.bsky.social for a great article.

000

QED Science @qedscience.bsky.social · 10/08/2026

Nature spoke with our co-founder and CEO, Niv Mastboim about how QED Science evaluates scientific work based on originality and validity, independently of author identity, institution, or publication venue. bit.ly/4wj2Xhk

041

QED Science @qedscience.bsky.social · 06/08/2026

You've read your grant proposal so many times you can't see it anymore. That means you can’t see the unclear framing, logic gaps, or structural weaknesses that are exactly why strong science gets rejected. Join us on Aug 20 to learn how to find these weaknesses before a panel does. bit.ly/4gbtTKZ

bit.ly

Welcome! You are invited to join a webinar: How to Find Weaknesses in Your Grant Before Reviewers Do. After registering, you will receive a confirmation email about joining the webinar.

Submitting an ERC Advanced Grant by August 27? Get a critical second pass on your proposal while there is still time to strengthen it. In this practical live webinar, Dr. Angela Andersen, Prof. Barak...

000

QED Science @qedscience.bsky.social · 06/08/2026

Looking forward to hosting you @Angela Andersen on our August 20 webinar.

QED Science @qedscience.bsky.social · 04/08/2026

The closer you are to your grant proposal, the harder it can be to see what a reviewer will. Join our free webinar on August 20 to learn how to spot gaps in logic, unclear framing and unsupported claims before submission. 11 AM ET | 5 PM CEST Register: bit.ly/4gbtTKZ @angelaandersen.bsky.social

032

QED Science @qedscience.bsky.social · 19/07/2026

Congrats on the 99 score! This is what good science looks like!🏆

QED Science @qedscience.bsky.social · 15/07/2026

Peer review was built for a world with fewer papers and more time. We're building for the world science has become. Grateful to The AIJournal for naming QED Science the top platform for 2026, and for recognizing that rigor and scale do not have to work against each other. bit.ly/4gzNANi

032

QED Science @qedscience.bsky.social · 14/07/2026

More than half of this year's top-scoring life science papers came from international teams. Even accounting for team size: international collaboration = 45% higher odds of hitting the top 1% (OR 1.45, p<0.001). Great science crosses borders. Read the research: bit.ly/4gxTY7K

QED Science @qedscience.bsky.social · 13/07/2026

@angelaandersen.bsky.social - Watch out for our blog today. It answers some of your questions.👀

011

Lucky for you, If you upload a paper to the QED platform, you automatically get tailored recommendations for preprints based on your inferred interests! 😀

000

We also tried a hyper aggressive anonimization strategy which risks removing scientific content but gender disparity remained. There could be many possible reasons for this phenomenon, but we do not believe the QED score is itself biased.

101

This is an important concern. The anonimization we did handled the reference section as well. We indeed saw a minor gender disparity in the score which becomes more significant at the very top. We however found no evidence of gender information leaking into the evaluation.

201

We try to understand if the issue undermines the science, is the science interesting and robust does the issue invalidate it? Our evaluation thinks this is an interesting paper worth paying attention to, and so do we.

000

QED is designed to evaluate the science itself, not the polish of the manuscript. Issues in figures should be considered seriously, but they don't automatically determine the overall scientific evaluation.

100

000

That only works if the system stays transparent and open to scrutiny. That’s why manuscripts are anonymized before scoring, and the evaluation focuses on claims and evidence, not author names, institutions, journal brands, or citations.

000

Fair questions, and we’re following the reactions closely too. Scientists need better ways to navigate the flood of preprints. But no one needs another prestige system for its own sake. For us, QED Score is a first-pass signal to help screen papers and decide what may be worth reading closely.

100

For us, QED Score is designed as a first-pass signal to help screen and navigate a growing volume of papers. Not a final word. A way to help more people decide what is worth reading closely. That’s also why we published the methodology, validation approach, and limitations.

000

Thanks for the critique. We agree the risk is real: any quality signal can be misused if it becomes a shortcut or gatekeeping tool. That’s why the question isn’t just whether AI enters science. It already is. The question is how to make it transparent, validated, limited in scope, and accountable.

200

This is for people who do want another way into 57,000+ preprints. It’s not here to replace reading, judgment, or debate. It’s a signal to help more good science get found.

100

There are many kinds of important research, from foundational studies to methods, resources, and incremental advances, that are essential to scientific progress. Not being in The 1% says nothing about whether a paper is valuable.

020

We agree that being in The 1% isn't and shouldn't be the goal of science. The 1% was intended to highlight one very small subset of manuscripts, not define what good science looks like.

400

Glad it’s useful. The goal was always to spotlight the science itself. That’s why we anonymize manuscripts before scoring: no names, no institutions, no prestige signals. Just the science.

100

Scores should never replace reading, judgment, or debate. But they can help more people find the science worth reading in the first place.

Totally hear the concern. For us, this is not about glam. It’s about gems. The goal is not to create more bling around science, but to help surface strong work that can get missed when prestige is the main signal.

QED Science @qedscience.bsky.social · 25/06/2026

Thank you so much.🙏 That’s exactly what we hoped The 1% could do help people discover incredible biology with the first signal coming from the science itself, not journals or institutions. Any curation system has to be shaped carefully. QED Score looks at claims and evidence, not prestige signals.

100

Most breakthrough science sits in preprints, unread. 57,455 last year alone. 574 exceptional. We found them. the-one-percent.qedscience.com

the-one-percent.qedscience.com

QED — The most compelling preprints of 2025