I’m glad someone said this to his face and more of these CEOs need to hear it: If the conversation around tech seems negative now it’s because you, your products, and what they’re doing to our society fucking sucks. You stopped making optimistic technology and unleashed MechaHitler the job killer.
NIH leadership wants to strip worker's rights from the majority of scientists on its campuses.
But, the NIH does not get to say whether we are a union or not. We do, and we are no strangers to having to fight for our union. Time for a short history lesson. 1/n
Disclaimer: all views are my own.
One question (of many) I have is whether opinions change if researchers are asked if they think AI scientific reviews should be relied on for their colleagues' papers. Often, what researchers want differs depending on whether they are the author or the reviewer.
I'd posit that getting input on a manuscript in prep from a real human colleague could serve the same purpose, while also deepening interpersonal connections that have secondary positive effects. The value of this cannot be understated, but it does take more effort.
For example, the consensus that AI scientific reviews are most useful as part of the manuscript preparation process fits with my current feelings. This is a great way to identify blind spots to sort out before peer review, which may save time in review and reduce reviewer burden in the long run.
There will be lots to consider on this topic going forward and this provides a useful framework as well as some baseline impressions. As an editor, it's also good to know that my current opinions are in line with where authors are at the moment.
This is a really cool approach to both evaluating AI scientific reviews and determining how authors think and feel about having AI tools in the loop.
Props to @reviewcommons.org and @qedscience.bsky.social for making this effort!
link.springer.com/article/10.1...
In addition to the medical ethics violations, the leadership have stained otherwise important research programs. According to the files, CEO Davis was actively soliciting Epstein with presentations from research directors as late as 2017!
As an alumni of Mount Sinai, I am horrified by this reporting. Sadly, the culture of prioritizing money and prestige above all else was strong at Sinai. I’m less surprised than I would like to be. These doctors and CEO Kenneth Davis need to go. Now.
Holden Thorp’s slides from a recent talk.
The fundamental flaw with @holdenthorp.bsky.social’s argument is thinking that organizations and people can only do one thing at a time (even though he is really close to acknowledging it!)
Let’s break it down, bc I’m tired of this. 🧵
(cred: @ianlmorgan.bsky.social)
Headed to @keystonesymposia.bsky.social on Stem Cell Models in Embryology tomorrow! Looking forward to hearing what’s new at the interface of #stemcells and #DevBio and finding promising stories for @natcomms.nature.com. Want to chat? DM me or find me there!
Feel like this is a good lesson to keep generative AI far from your editorial process as a matter of policy. Ars is a great website. If it can happen there it can happen to anyone that allows these tools in
My quote of the day
Ignorance more frequently begets confidence than does knowledge: it is those who know little, and not those who know much, who so positively assert that this or that problem will never be solved by science.
Charles Darwin
Society for Developmental Biology logo Society for Developmental Biology 2026 Award Winners Edwin G. Conklin Medal Headshot of Lee Niswander Lee Niswander, University of Colorado Boulder Society for Developmental Biology Lifetime Achievement Award Headshot of Alexandra Joyner Alexandra Joyner, Memorial Sloan Kettering Cancer Center Viktor Hamburger Outstanding Educator Prize Headshot of Roberto Mayor Roberto Mayor, University College London Elizabeth D. Hay New Investigator Award Headshot of Jeffrey Farrell Jeffrey Farrell, National Institutes of Health Society for Developmental Biology Trainee Science Communication Award Headshot of Nicholas Desnoyer Nicholas Desnoyer, The Sainsbury Laboratory
Congrats to the 2026 SDB Award Winners!
Conklin Medal: Lee Niswander
SDB Lifetime Achievement Award: Alexandra Joyner
Hamburger Outstanding Educator Prize: Roberto Mayor
Hay New Investigator Award: Jeffrey Farrell
SDB Trainee SciComm Award: Nicholas Desnoyer
bit.ly/4afnjiC
The City’s Budget is our future. And you deserve to know how it works.
WE WON: Today, workers at the Environmental Defense Fund (EDF) voted to form a union, with 81% in favor! ✊
At over 460 eligible employees, this will be the largest environmental nonprofit union in the US.
newsguild.org/workers-at-e...
Half the battle with this guy is just reiterating government can and already does good things that are so ubiquitous they’ve become invisible. He’s able to remind you in a way that you understand that good government can exist outside of a 24/7 vengeance operation, and in NYC already does.
I’ve yet to see an AI-generated summary that does better than the author-written abstract.
In the end I think of an AI manuscript review like a scarecrow. It’ll fend off skittish birds but it’ll never have a brain.
So how could AI reviews be useful? I would be interested in seeing fields develop tailored prompts for authors to get AI reviews, which could then be included with #preprints.
Can a journal that uses it be considered a Peer Reviewed Journal? How would that affect indexing? Similar to authorship, and AI cannot be held accountable for its comments. This is an issue for research integrity.
First, what is Human+AI review? It’s not #PeerReview as normally defined. Much of the value added by a journal is soliciting high-quality reviews from relevant experts, and I don’t see the same value from this approach.
Some thoughts and questions about the Human+AI review pilot published by NEJM AI. There will be lots of efforts to integrate AI into #SciencePublishing and this is my current baseline.
ai.nejm.org/doi/full/10....
When you submit your manuscript, do you want 2/3 of the reviews you get to be AI-generated? Call me old fashioned, but I'd like to keep the "peer" in peer-review.
Accelerating Science with Human+AI Review | NEJM AI ai.nejm.org/doi/full/10....
👀
“If you are a 75-year-old professor reading this, now would be a good time to take emeritus status. It’s something that a lot of people could do that would contribute enormously. The job market for young people would be in much better shape if most people over 75 retired.”
What I always hear: “We can’t afford to pay our workers more than $15 an hour.”
What I never hear: “We can’t afford to pay our CEO millions of dollars a year.”
Funny how that works.
We probably shouldn’t have favourites amongst projects in the lab, but this one was mine. I’m so pleased we can finally share it:
Our work is now out in Nature Communications:
🔗 nature.com/articles/s41467-025-67692-7
Things fall through the cracks, but this is a long time for a paper to be in limbo. I’d suggest contacting the Editor in Chief for an update. Feel free to warn them that you may withdraw and go elsewhere!
Everything is getting stupider
Come join our team! Feel free to DM me with any questions.
