Stephen Ornes

Text on lefthand side reads 2025 AAAS Kavli Science Journalism Award Fall Lecture Series / AI and Science Journalism: / An Uneasy Future / Featuring Stephen Ornes / September 30, 2025 / 11:00 A.M. ET. Headshot of Stephen Ornes smiling on righthand side. AAAS, Kavli Foundation, and Grady College logos along the bottom of image.

Three-time AAAS Kavli Science Journalism Award winner @stephenornes.bsky.social will kick off this year's AAAS Kavli Fall Lecture Series at @ugagrady.bsky.social
next week. Ornes will discuss the history of AI, why it has become so seductive to cover, and the demands for getting it right.

Thanks!

The AI Was Fed Sloppy Code. It Turned Into Something Evil. | Quanta Magazine The new science of “emergent misalignment” explores how PG-13 training data — insecure code, superstitious numbers or even extreme-sports advice — can open the door to AI’s dark side.

In a recent study, researchers introduced a small data set of easily hackable code to a collection of chatbots. Shortly after, the AI began to offer helpful tips for how to get away with murder. www.quantamagazine.org/the-ai-was-f...

Scientific societies say they'll step up after Trump puts key climate report in doubt Two major scientific societies say they will try to fill the void from the Trump administration’s dismissal of scientists writing a cornerstone federal report on what climate change is doing to the Un...

The Trump administration has dismissed the scientists behind the U.S. climate report.

Now scientific societies are stepping in to finish the job—because the climate crisis doesn’t pause for politics.

Science will not be silenced.

apnews.com/article/clim...

#ClimateScience #StandUpForScience

Hegseth, Waltz, Gabbard: Private Data and Passwords of Senior U.S. Security Officials Found Online Donald Trump's most important security advisers used Signal to discuss an imminent military strike. Now, reporting by DER SPIEGEL has found that the contact data of some of those officials, including ...

In other news, Der Spiegel
found private phone numbers, email addresses and some passwords for Waltz, Gabbard and Hegseth online

www.spiegel.de/internationa...

Quantum Speedup Found for Huge Class of Hard Problems | Quanta Magazine It’s been difficult to find important questions that quantum computers can answer faster than classical machines, but a new algorithm appears to do it for some critical optimization tasks.

It’s been difficult to find important questions that quantum computers can answer faster than classical machines, but a new algorithm appears to do it for some critical optimization tasks. @stephenornes.bsky.social reports: www.quantamagazine.org/quantum-spee...

Why Do Researchers Care About Small Language Models? | Quanta Magazine Larger models can pull off a wider variety of feats, but the reduced footprint of smaller models makes them attractive tools.

Small language models (SLMs) can use only a few billion parameters — a mere fraction of their LLM counterparts.

buff.ly/i0Jjuq4

John, who died in 2022, had more to say and more to do. The world is richer for his work, though, and Pi Day offers an appropriate day to raise a glass to him, at least 3.14... times.

Math artist John Sims with science writer Stephen Ornes and Ella Toy, owner of Alma Sue's Quilt Shop in Sarasota, Florida.

This is John with Ella Toy and me, in Alma Sue's quilt shop. In this shop, together with a crew of Amish snowbirds, John pieced together his enormous, Gee's Bend-inspired quilts. And here's his photo of the quilters in action.

YouTube video by John Sims The Pi Day Anthem featuring John Sims and Vi Hart

Happy Pi Day to all who celebrate! Math and political artist John Sims found inspiration in everyone's favorite irrational number to produce quilts, prints, dresses, spoken word poetry, and more. Together with Vi Hart, he also produced this Pi Day Anthem: www.youtube.com/watch?v=NsKH...

Why Do Researchers Care About Small Language Models? | Quanta Magazine Larger models can pull off a wider variety of feats, but the reduced footprint of smaller models makes them attractive tools.

Larger models can pull off greater feats, but the accessibility and efficiency of smaller models make them attractive tools. @stephenornes.bsky.social reports: www.quantamagazine.org/why-do-resea...

A screenshot of an email notifying Spotsylvania parents that funding for a program for youth with disabilities has been canceled

the richest man in the world has decided that your kids don't deserve special education programs

I've met with many cabinet nominees across many admins, and my meeting with RFK Jr. was by far the most troubling.

He rattled off debunked conspiracy theories & lied to my face—denying his own statements he'd made on record. We cannot confirm an anti-vaxxer as Secretary of Health.

Big congrats!

Senate panel advances RFK Jr.'s nomination to be health secretary The Senate Finance Committee voted along party lines to send Kennedy’s bid to lead the Department of Health and Human Services to the full chamber.

Senator Bill Cassidy just voted to advance illness, disability, and death for the American people.

Disgraceful abandonment of your obligations as a doctor and an elected official. The preventable deaths this man will cause are on your hands.
www.nbcnews.com/politics/con...

Thread. Nothing good.

Screenshot of a poster titled Discovering Our Journalism January 8 to February 17 2025

"Discovering Our Journalism: Empowering First-Gen Undergraduates and Early Career Science Writers Through Mentorship" is a 2024 NASW #SciWriGrants recipients—and they're currently seeking mentee signups!

Learn How to Apply: sites.google.com/uri...
About the Grant: www.nasw.org/article...

💚🧪📝

It's time to apply for a Knight Science Journalism Program fellowship at MIT! We have 10 openings for a 2025-26 academic year fellowship and one opening for a one semester fellowship for journalists from Africa and the Middle East. Learn more here.
ksj.mit.edu/news/2024/11...

Debate May Help AI Models Converge on Truth | Quanta Magazine Letting AI systems argue with each other may help expose when a large language model has made mistakes.

A little healthy debate can go a long way: Computer scientists are using moderated argument as a key strategy to strengthen large language model accuracy.

www.quantamagazine.org/debate-may-h...